Changes in / [24d6572:62d62db]
- Files:
-
- 2 deleted
- 52 edited
Legend:
- Unmodified
- Added
- Removed
-
doc/papers/llheap/Paper.tex
r24d6572 r62d62db 252 252 Dynamic code/data memory is managed by the dynamic loader for libraries loaded at runtime, which is complex especially in a multi-threaded program~\cite{Huang06}. 253 253 However, changes to the dynamic code/data space are typically infrequent, many occurring at program startup, and are largely outside of a program's control. 254 Stack memory is managed by the program call/return-mechanism using a simpleLIFO technique, which works well for sequential programs.255 For stackful coroutines and user threads, a new stack is commonly created in dynamic-allocation memory.254 Stack memory is managed by the program call/return-mechanism using a LIFO technique, which works well for sequential programs. 255 For stackful coroutines and user threads, a new stack is commonly created in the dynamic-allocation memory. 256 256 This work focuses solely on management of the dynamic-allocation memory. 257 257 … … 293 293 \begin{enumerate}[leftmargin=*,itemsep=0pt] 294 294 \item 295 Implementation of a new stand-alone concurrent low-latency memory-allocator ($\approx$1,200 lines of code) for C/\CC programs using kernel threads (1:1 threading), and specialized versions of the allocator for the programming languages \uC and \CFAusing user-level threads running on multiple kernel threads (M:N threading).296 297 \item 298 Extend the standard C heap functionality by preserving with each allocation: its request size plus the amount allocated, whether an allocation is zero fill , andallocation alignment.295 Implementation of a new stand-alone concurrent low-latency memory-allocator ($\approx$1,200 lines of code) for C/\CC programs using kernel threads (1:1 threading), and specialized versions of the allocator for the programming languages \uC~\cite{uC++} and \CFA~\cite{Moss18,Delisle21} using user-level threads running on multiple kernel threads (M:N threading). 296 297 \item 298 Extend the standard C heap functionality by preserving with each allocation: its request size plus the amount allocated, whether an allocation is zero fill and/or allocation alignment. 299 299 300 300 \item … … 365 365 366 366 The following discussion is a quick overview of the moving-pieces that affect the design of a memory allocator and its performance. 367 It is assumed that dynamic allocates and deallocates acquire storage for a program variable, referred to asan \newterm{object}, through calls such as @malloc@ and @free@ in C, and @new@ and @delete@ in \CC.367 Dynamic acquires and releases obtain storage for a program variable, called an \newterm{object}, through calls such as @malloc@ and @free@ in C, and @new@ and @delete@ in \CC. 368 368 Space for each allocated object comes from the dynamic-allocation zone. 369 369 … … 378 378 379 379 Figure~\ref{f:AllocatorComponents} shows the two important data components for a memory allocator, management and storage, collectively called the \newterm{heap}. 380 The \newterm{management data} is a data structure located at a known memory address and contains all information necessary to manage the storage data. 381 The management data starts with fixed-sized information in the static-data memory that references components in the dynamic-allocation memory. 380 The \newterm{management data} is a data structure located at a known memory address and contains fixed-sized information in the static-data memory that references components in the dynamic-allocation memory. 382 381 For multi-threaded programs, additional management data may exist in \newterm{thread-local storage} (TLS) for each kernel thread executing the program. 383 382 The \newterm{storage data} is composed of allocated and freed objects, and \newterm{reserved memory}. … … 385 384 \ie only the program knows the location of allocated storage not the memory allocator. 386 385 Freed objects (white) represent memory deallocated by the program, which are linked into one or more lists facilitating easy location of new allocations. 387 Reserved memory (dark grey) is one or more blocks of memory obtained from the operating systembut not yet allocated to the program;388 if there are multiple reserved blocks, they are also chained together , usually internally.386 Reserved memory (dark grey) is one or more blocks of memory obtained from the \newterm{operating system} (OS) but not yet allocated to the program; 387 if there are multiple reserved blocks, they are also chained together. 389 388 390 389 \begin{figure} … … 395 394 \end{figure} 396 395 397 In m ost allocator designs, allocated objects have management data embedded within them.396 In many allocator designs, allocated objects and reserved blocks have management data embedded within them (see also Section~\ref{s:ObjectContainers}). 398 397 Figure~\ref{f:AllocatedObject} shows an allocated object with a header, trailer, and optional spacing around the object. 399 398 The header contains information about the object, \eg size, type, etc. … … 404 403 When padding and spacing are necessary, neither can be used to satisfy a future allocation request while the current allocation exists. 405 404 406 A free object alsocontains management data, \eg size, pointers, etc.405 A free object often contains management data, \eg size, pointers, etc. 407 406 Often the free list is chained internally so it does not consume additional storage, \ie the link fields are placed at known locations in the unused memory blocks. 408 407 For internal chaining, the amount of management data for a free node defines the minimum allocation size, \eg if 16 bytes are needed for a free-list node, allocation requests less than 16 bytes are rounded up. 409 The information in an allocated or freed object is overwritten when it transitions from allocated to freed and vice-versa by new management information and/or program data.408 The information in an allocated or freed object is overwritten when it transitions from allocated to freed and vice-versa by new program data and/or management information. 410 409 411 410 \begin{figure} … … 428 427 \label{s:Fragmentation} 429 428 430 Fragmentation is memory requested from the operating systembut not used by the program;429 Fragmentation is memory requested from the OS but not used by the program; 431 430 hence, allocated objects are not fragmentation. 432 431 Figure~\ref{f:InternalExternalFragmentation} shows fragmentation is divided into two forms: internal or external. … … 443 442 An allocator should strive to keep internal management information to a minimum. 444 443 445 \newterm{External fragmentation} is all memory space reserved from the operating systembut not allocated to the program~\cite{Wilson95,Lim98,Siebert00}, which includes all external management data, freed objects, and reserved memory.444 \newterm{External fragmentation} is all memory space reserved from the OS but not allocated to the program~\cite{Wilson95,Lim98,Siebert00}, which includes all external management data, freed objects, and reserved memory. 446 445 This memory is problematic in two ways: heap blowup and highly fragmented memory. 447 446 \newterm{Heap blowup} occurs when freed memory cannot be reused for future allocations leading to potentially unbounded external fragmentation growth~\cite{Berger00}. 448 Memory can become \newterm{highly fragmented} after multiple allocations and deallocations of objects, resulting in a checkerboard of adjacent allocated and free areas, where the free blocks have become very small.447 Memory can become \newterm{highly fragmented} after multiple allocations and deallocations of objects, resulting in a checkerboard of adjacent allocated and free areas, where the free blocks have become to small to service requests. 449 448 % Figure~\ref{f:MemoryFragmentation} shows an example of how a small block of memory fragments as objects are allocated and deallocated over time. 450 449 Heap blowup can occur due to allocator policies that are too restrictive in reusing freed memory (the allocated size cannot use a larger free block) and/or no coalescing of free storage. … … 452 451 % Memory is highly fragmented when most free blocks are unusable because of their sizes. 453 452 % For example, Figure~\ref{f:Contiguous} and Figure~\ref{f:HighlyFragmented} have the same quantity of external fragmentation, but Figure~\ref{f:HighlyFragmented} is highly fragmented. 454 % If there is a request to allocate a large object, Figure~\ref{f:Contiguous} is more likely to be able to satisfy it with existing free memory, while Figure~\ref{f:HighlyFragmented} likely has to request more memory from the operating system.453 % If there is a request to allocate a large object, Figure~\ref{f:Contiguous} is more likely to be able to satisfy it with existing free memory, while Figure~\ref{f:HighlyFragmented} likely has to request more memory from the OS. 455 454 456 455 % \begin{figure} … … 475 474 The first approach is a \newterm{sequential-fit algorithm} with one list of free objects that is searched for a block large enough to fit a requested object size. 476 475 Different search policies determine the free object selected, \eg the first free object large enough or closest to the requested size. 477 Any storage larger than the request can become spacing after the object or besplit into a smaller free object.476 Any storage larger than the request can become spacing after the object or split into a smaller free object. 478 477 % The cost of the search depends on the shape and quality of the free list, \eg a linear versus a binary-tree free-list, a sorted versus unsorted free-list. 479 478 … … 489 488 490 489 The third approach is \newterm{splitting} and \newterm{coalescing algorithms}. 491 When an object is allocated, if there are no free objects of the requested size, a larger free object may be split into two smaller objects to satisfy the allocation request without obtaining more memory from the operating system.492 For example, in the \newterm{buddy system}, a block of free memory is split into two equal chunks, one of those chunks is again split into two equal chunks, and so on until a block just large enough to fit the requested object is created.493 When an object is deallocated it is coalesced with the objects immediately before and after it in memory, if they are free, turning them into one larger object.490 When an object is allocated, if there are no free objects of the requested size, a larger free object is split into two smaller objects to satisfy the allocation request rather than obtaining more memory from the OS. 491 For example, in the \newterm{buddy system}, a block of free memory is split into equal chunks, one of those chunks is again split, and so on until a minimal block is created that fits the requested object. 492 When an object is deallocated, it is coalesced with the objects immediately before and after it in memory, if they are free, turning them into one larger block. 494 493 Coalescing can be done eagerly at each deallocation or lazily when an allocation cannot be fulfilled. 495 In all cases, coalescing increases allocation latency, hence some allocations can cause unbounded delays during coalescing.494 In all cases, coalescing increases allocation latency, hence some allocations can cause unbounded delays. 496 495 While coalescing does not reduce external fragmentation, the coalesced blocks improve fragmentation quality so future allocations are less likely to cause heap blowup. 497 496 % Splitting and coalescing can be used with other algorithms to avoid highly fragmented memory. … … 501 500 \label{s:Locality} 502 501 503 The principle of locality recognizes that programs tend to reference a small set of data, called a working set, for a certain period of time, where a working set iscomposed of temporal and spatial accesses~\cite{Denning05}.502 The principle of locality recognizes that programs tend to reference a small set of data, called a \newterm{working set}, for a certain period of time, composed of temporal and spatial accesses~\cite{Denning05}. 504 503 % Temporal clustering implies a group of objects are accessed repeatedly within a short time period, while spatial clustering implies a group of objects physically close together (nearby addresses) are accessed repeatedly within a short time period. 505 504 % Temporal locality commonly occurs during an iterative computation with a fixed set of disjoint variables, while spatial locality commonly occurs when traversing an array. 506 Hardware takes advantage of t emporal and spatial localitythrough multiple levels of caching, \ie memory hierarchy.505 Hardware takes advantage of the working set through multiple levels of caching, \ie memory hierarchy. 507 506 % When an object is accessed, the memory physically located around the object is also cached with the expectation that the current and nearby objects will be referenced within a short period of time. 508 For example, entire cache lines are transferred between memory and cache and entire virtual-memory pages are transferred between disk and memory.507 For example, entire cache lines are transferred between cache and memory, and entire virtual-memory pages are transferred between memory and disk. 509 508 % A program exhibiting good locality has better performance due to fewer cache misses and page faults\footnote{With the advent of large RAM memory, paging is becoming less of an issue in modern programming.}. 510 509 … … 532 531 \label{s:MutualExclusion} 533 532 534 \newterm{Mutual exclusion} provides sequential access to the shared 533 \newterm{Mutual exclusion} provides sequential access to the shared-management data of the heap. 535 534 There are two performance issues for mutual exclusion. 536 535 First is the overhead necessary to perform (at least) a hardware atomic operation every time a shared resource is accessed. 537 536 Second is when multiple threads contend for a shared resource simultaneously, and hence, some threads must wait until the resource is released. 538 537 Contention can be reduced in a number of ways: 539 1) Using multiple fine-grained locks versus a single lock , spreadingthe contention across a number of locks.538 1) Using multiple fine-grained locks versus a single lock to spread the contention across a number of locks. 540 539 2) Using trylock and generating new storage if the lock is busy, yielding a classic space versus time tradeoff. 541 540 3) Using one of the many lock-free approaches for reducing contention on basic data-structure operations~\cite{Oyama99}. … … 551 550 a memory allocator can only affect the latter two. 552 551 553 Assume two objects, object$_1$ and object$_2$, share a cache line.554 \newterm{Program-induced false-sharing} occurs when thread$_1$ passes a reference to object$_2$ to thread$_2$, and then threads$_1$ modifies object$_1$ while thread$_2$ modifies object$_2$.552 Specifically, assume two objects, O$_1$ and O$_2$, share a cache line, with threads, T$_1$ and T$_2$. 553 \newterm{Program-induced false-sharing} occurs when T$_1$ passes a reference to O$_2$ to T$_2$, and then T$_1$ modifies O$_1$ while T$_2$ modifies O$_2$. 555 554 % Figure~\ref{f:ProgramInducedFalseSharing} shows when Thread$_1$ passes Object$_2$ to Thread$_2$, a false-sharing situation forms when Thread$_1$ modifies Object$_1$ and Thread$_2$ modifies Object$_2$. 556 555 % Changes to Object$_1$ invalidate CPU$_2$'s cache line, and changes to Object$_2$ invalidate CPU$_1$'s cache line. … … 574 573 % \label{f:FalseSharing} 575 574 % \end{figure} 576 \newterm{Allocator-induced active false-sharing}\label{s:AllocatorInducedActiveFalseSharing} occurs when object$_1$ and object$_2$ are heap allocated and their references are passed to thread$_1$ and thread$_2$, which modify the objects.575 \newterm{Allocator-induced active false-sharing}\label{s:AllocatorInducedActiveFalseSharing} occurs when O$_1$ and O$_2$ are heap allocated and their references are passed to T$_1$ and T$_2$, which modify the objects. 577 576 % For example, in Figure~\ref{f:AllocatorInducedActiveFalseSharing}, each thread allocates an object and loads a cache-line of memory into its associated cache. 578 577 % Again, changes to Object$_1$ invalidate CPU$_2$'s cache line, and changes to Object$_2$ invalidate CPU$_1$'s cache line. … … 580 579 % is another form of allocator-induced false-sharing caused by program-induced false-sharing. 581 580 % When an object in a program-induced false-sharing situation is deallocated, a future allocation of that object may cause passive false-sharing. 582 when thread$_1$ passes object$_2$ to thread$_2$, and thread$_2$ subsequently deallocates object$_2$, and then object$_2$ is reallocated to thread$_2$ while thread$_1$ is still using object$_1$.581 when T$_1$ passes O$_2$ to T$_2$, and T$_2$ subsequently deallocates O$_2$, and then O$_2$ is reallocated to T$_2$ while T$_1$ is still using O$_1$. 583 582 584 583 … … 593 592 \label{s:MultiThreadedMemoryAllocatorFeatures} 594 593 595 The following features are used in the construction of multi-threaded memory-allocators: 596 \begin{enumerate}[itemsep=0pt] 597 \item multiple heaps: with or without a global heap, or with or without heap ownership. 598 \item object containers: with or without ownership, fixed or variable sized, global or local free-lists. 599 \item hybrid private/public heap 600 \item allocation buffer 601 \item lock-free operations 602 \end{enumerate} 594 The following features are used in the construction of multi-threaded memory-allocators: multiple heaps, user-level threading, ownership, object containers, allocation buffer, lock-free operations. 603 595 The first feature, multiple heaps, pertains to different kinds of heaps. 604 596 The second feature, object containers, pertains to the organization of objects within the storage area. … … 606 598 607 599 608 \subs ection{Multiple Heaps}600 \subsubsection{Multiple Heaps} 609 601 \label{s:MultipleHeaps} 610 602 611 603 A multi-threaded allocator has potentially multiple threads and heaps. 612 604 The multiple threads cause complexity, and multiple heaps are a mechanism for dealing with the complexity. 613 The spectrum ranges from multiple threads using a single heap, denoted as T:1 (see Figure~\ref{f:SingleHeap}), to multiple threads sharing multiple heaps, denoted as T:H (see Figure~\ref{f:SharedHeaps}), to one thread per heap, denoted as 1:1 (see Figure~\ref{f:PerThreadHeap}), which is almost back to a single-threaded allocator.605 The spectrum ranges from multiple threads using a single heap, denoted as T:1, to multiple threads sharing multiple heaps, denoted as T:H, to one thread per heap, denoted as 1:1, which is almost back to a single-threaded allocator. 614 606 615 607 \begin{figure} … … 635 627 \end{figure} 636 628 637 \paragraph{T:1 model } where all threads allocate and deallocate objects from one heap.638 Memory is obtained from the freed objects, or reserved memory in the heap, or from the operating system (OS);639 the heap may also return freed memory to the operating system.629 \paragraph{T:1 model (see Figure~\ref{f:SingleHeap})} where all threads allocate and deallocate objects from one heap. 630 Memory is obtained from the freed objects, or reserved memory in the heap, or from the OS; 631 the heap may also return freed memory to the OS. 640 632 The arrows indicate the direction memory conceptually moves for each kind of operation: allocation moves memory along the path from the heap/operating-system to the user application, while deallocation moves memory along the path from the application back to the heap/operating-system. 641 633 To safely handle concurrency, a single lock may be used for all heap operations or fine-grained locking for different operations. 642 634 Regardless, a single heap may be a significant source of contention for programs with a large amount of memory allocation. 643 635 644 \paragraph{T:H model } where each thread allocates storage from several heaps depending on certain criteria, with the goal of reducing contention by spreading allocations/deallocations across the heaps.636 \paragraph{T:H model (see Figure~\ref{f:SharedHeaps})} where each thread allocates storage from several heaps depending on certain criteria, with the goal of reducing contention by spreading allocations/deallocations across the heaps. 645 637 The decision on when to create a new heap and which heap a thread allocates from depends on the allocator design. 646 638 To determine which heap to access, each thread must point to its associated heap in some way. … … 673 665 An alternative implementation is for all heaps to share one reserved memory, which requires a separate lock for the reserved storage to ensure mutual exclusion when acquiring new memory. 674 666 Because multiple threads can allocate/free/reallocate adjacent storage, all forms of false sharing may occur. 675 Other storage-management options are to use @mmap@ to set aside (large) areas of virtual memory for each heap and suballocate each heap's storage within that area, pushing part of the storage management complexity back to the operating system.667 Other storage-management options are to use @mmap@ to set aside (large) areas of virtual memory for each heap and suballocate each heap's storage within that area, pushing part of the storage management complexity back to the OS. 676 668 677 669 % \begin{figure} … … 684 676 Multiple heaps increase external fragmentation as the ratio of heaps to threads increases, which can lead to heap blowup. 685 677 The external fragmentation experienced by a program with a single heap is now multiplied by the number of heaps, since each heap manages its own free storage and allocates its own reserved memory. 686 Additionally, objects freed by one heap cannot be reused by other threads without increasing the cost of the memory operations, except indirectly by returning free memory to the operating system, which can be expensive.687 Depending on how the operating system provides dynamic storage to an application, returning storagemay be difficult or impossible, \eg the contiguous @sbrk@ area in Unix.688 In the worst case, a program in which objects are allocated from one heap but deallocated to another heap means these freed objects are never reused.678 Additionally, objects freed by one heap cannot be reused by other threads without increasing the cost of the memory operations, except indirectly by returning free memory to the OS (see Section~\ref{s:Ownership}). 679 Returning storage to the OS may be difficult or impossible, \eg the contiguous @sbrk@ area in Unix. 680 % In the worst case, a program in which objects are allocated from one heap but deallocated to another heap means these freed objects are never reused. 689 681 690 682 Adding a \newterm{global heap} (G) attempts to reduce the cost of obtaining/returning memory among heaps (sharing) by buffering storage within the application address-space. 691 Now, each heap obtains and returns storage to/from the global heap rather than the operating system.683 Now, each heap obtains and returns storage to/from the global heap rather than the OS. 692 684 Storage is obtained from the global heap only when a heap allocation cannot be fulfilled, and returned to the global heap when a heap's free memory exceeds some threshold. 693 Similarly, the global heap buffers this memory, obtaining and returning storage to/from the operating systemas necessary.685 Similarly, the global heap buffers this memory, obtaining and returning storage to/from the OS as necessary. 694 686 The global heap does not have its own thread and makes no internal allocation requests; 695 687 instead, it uses the application thread, which called one of the multiple heaps and then the global heap, to perform operations. 696 688 Hence, the worst-case cost of a memory operation includes all these steps. 697 With respect to heap blowup, the global heap provides an indirect mechanism to move free memory among heaps, which usually has a much lower cost than interacting with the operating system to achieve the same goal and is independent of the mechanism used by the operating system to present dynamic memory to an address space. 698 689 With respect to heap blowup, the global heap provides an indirect mechanism to move free memory among heaps, which usually has a much lower cost than interacting with the OS to achieve the same goal and is independent of the mechanism used by the OS to present dynamic memory to an address space. 699 690 However, since any thread may indirectly perform a memory operation on the global heap, it is a shared resource that requires locking. 700 691 A single lock can be used to protect the global heap or fine-grained locking can be used to reduce contention. 701 692 In general, the cost is minimal since the majority of memory operations are completed without the use of the global heap. 702 693 703 704 \paragraph{1:1 model (thread heaps)} where each thread has its own heap eliminating most contention and locking because threads seldom access another thread's heap (see ownership in Section~\ref{s:Ownership}). 694 \paragraph{1:1 model (see Figure~\ref{f:PerThreadHeap})} where each thread has its own heap eliminating most contention and locking because threads seldom access another thread's heap (see Section~\ref{s:Ownership}). 705 695 An additional benefit of thread heaps is improved locality due to better memory layout. 706 696 As each thread only allocates from its heap, all objects are consolidated in the storage area for that heap, better utilizing each CPUs cache and accessing fewer pages. … … 708 698 Thread heaps can also eliminate allocator-induced active false-sharing, if memory is acquired so it does not overlap at crucial boundaries with memory for another thread's heap. 709 699 For example, assume page boundaries coincide with cache line boundaries, if a thread heap always acquires pages of memory then no two threads share a page or cache line unless pointers are passed among them. 710 Hence, allocator-induced active false-sharing cannot occur because the memory for thread heaps never overlaps.700 % Hence, allocator-induced active false-sharing cannot occur because the memory for thread heaps never overlaps. 711 701 712 702 When a thread terminates, there are two options for handling its thread heap. … … 720 710 721 711 It is possible to use any of the heap models with user-level (M:N) threading. 722 However, an important goal of user-level threading is for fast operations (creation/termination/context-switching) by not interacting with the operating system, which allows the ability to create large numbers of high-performance interacting threads ($>$ 10,000).712 However, an important goal of user-level threading is for fast operations (creation/termination/context-switching) by not interacting with the OS, which allows the ability to create large numbers of high-performance interacting threads ($>$ 10,000). 723 713 It is difficult to retain this goal, if the user-threading model is directly involved with the heap model. 724 714 Figure~\ref{f:UserLevelKernelHeaps} shows that virtually all user-level threading systems use whatever kernel-level heap-model is provided by the language runtime. … … 732 722 \end{figure} 733 723 734 Adopting this modelresults in a subtle problem with shared heaps.735 With kernel threading, an operation that isstarted by a kernel thread is always completed by that thread.736 For example, if a kernel thread starts an allocation/deallocation on a shared heap, it always completes that operation with that heap even if preempted, \ie any locking correctness associated with the shared heap is preserved across preemption.724 Adopting user threading results in a subtle problem with shared heaps. 725 With kernel threading, an operation started by a kernel thread is always completed by that thread. 726 For example, if a kernel thread starts an allocation/deallocation on a shared heap, it always completes that operation with that heap, even if preempted, \ie any locking correctness associated with the shared heap is preserved across preemption. 737 727 However, this correctness property is not preserved for user-level threading. 738 728 A user thread can start an allocation/deallocation on one kernel thread, be preempted (time slice), and continue running on a different kernel thread to complete the operation~\cite{Dice02}. 739 729 When the user thread continues on the new kernel thread, it may have pointers into the previous kernel-thread's heap and hold locks associated with it. 740 730 To get the same kernel-thread safety, time slicing must be disabled/\-enabled around these operations, so the user thread cannot jump to another kernel thread. 741 However, eagerly disabling/enabling time-slicing on the allocation/deallocation fast path is expensive, because preemption does not happen that frequently.731 However, eagerly disabling/enabling time-slicing on the allocation/deallocation fast path is expensive, because preemption is infrequent (milliseconds). 742 732 Instead, techniques exist to lazily detect this case in the interrupt handler, abort the preemption, and return to the operation so it can complete atomically. 743 Occasional ly ignoring a preemption should be benign, but a persistent lack of preemption can result in both short and long termstarvation;744 techniques like roll forward can be used to force an eventual preemption.733 Occasional ignoring of a preemption should be benign, but a persistent lack of preemption can result in starvation; 734 techniques like rolling forward the preemption to the next context switch can be used. 745 735 746 736 … … 800 790 % For example, in Figure~\ref{f:AllocatorInducedPassiveFalseSharing}, Object$_2$ may be deallocated to Thread$_2$'s heap initially. 801 791 % If Thread$_2$ reallocates Object$_2$ before it is returned to its owner heap, then passive false-sharing may occur. 792 793 For thread heaps with ownership, it is possible to combine these approaches into a hybrid approach with both private and public heaps.% (see~Figure~\ref{f:HybridPrivatePublicHeap}). 794 The main goal of the hybrid approach is to eliminate locking on thread-local allocation/deallocation, while providing ownership to prevent heap blowup. 795 In the hybrid approach, a thread first allocates from its private heap and second from its public heap if no free memory exists in the private heap. 796 Similarly, a thread first deallocates an object to its private heap, and second to the public heap. 797 Both private and public heaps can allocate/deallocate to/from the global heap if there is no free memory or excess free memory, although an implementation may choose to funnel all interaction with the global heap through one of the heaps. 798 % Note, deallocation from the private to the public (dashed line) is unlikely because there is no obvious advantages unless the public heap provides the only interface to the global heap. 799 Finally, when a thread frees an object it does not own, the object is either freed immediately to its owner's public heap or put in the freeing thread's private heap for delayed ownership, which does allows the freeing thread to temporarily reuse an object before returning it to its owner or batch objects for an owner heap into a single return. 800 801 % \begin{figure} 802 % \centering 803 % \input{PrivatePublicHeaps.pstex_t} 804 % \caption{Hybrid Private/Public Heap for Per-thread Heaps} 805 % \label{f:HybridPrivatePublicHeap} 806 % \vspace{10pt} 807 % \input{RemoteFreeList.pstex_t} 808 % \caption{Remote Free-List} 809 % \label{f:RemoteFreeList} 810 % \end{figure} 811 812 % As mentioned, an implementation may have only one heap interact with the global heap, so the other heap can be simplified. 813 % For example, if only the private heap interacts with the global heap, the public heap can be reduced to a lock-protected free-list of objects deallocated by other threads due to ownership, called a \newterm{remote free-list}. 814 % To avoid heap blowup, the private heap allocates from the remote free-list when it reaches some threshold or it has no free storage. 815 % Since the remote free-list is occasionally cleared during an allocation, this adds to that cost. 816 % Clearing the remote free-list is $O(1)$ if the list can simply be added to the end of the private-heap's free-list, or $O(N)$ if some action must be performed for each freed object. 817 818 % If only the public heap interacts with other threads and the global heap, the private heap can handle thread-local allocations and deallocations without locking. 819 % In this scenario, the private heap must deallocate storage after reaching a certain threshold to the public heap (and then eventually to the global heap from the public heap) or heap blowup can occur. 820 % If the public heap does the major management, the private heap can be simplified to provide high-performance thread-local allocations and deallocations. 821 822 % The main disadvantage of each thread having both a private and public heap is the complexity of managing two heaps and their interactions in an allocator. 823 % Interestingly, heap implementations often focus on either a private or public heap, giving the impression a single versus a hybrid approach is being used. 824 % In many case, the hybrid approach is actually being used, but the simpler heap is just folded into the complex heap, even though the operations logically belong in separate heaps. 825 % For example, a remote free-list is actually a simple public-heap, but may be implemented as an integral component of the complex private-heap in an allocator, masking the presence of a hybrid approach. 802 826 803 827 … … 817 841 818 842 819 \subs ection{Object Containers}843 \subsubsection{Object Containers} 820 844 \label{s:ObjectContainers} 821 845 … … 827 851 \eg an object is accessed by the program after it is allocated, while the header is accessed by the allocator after it is free. 828 852 829 The alternative factors common header data to a separate location in memory and organizes associated free storage into blocks called \newterm{object containers} (\newterm{superblocks} in~\cite{Berger00}), as in Figure~\ref{f:ObjectContainer}.853 An alternative approach factors common header data to a separate location in memory and organizes associated free storage into blocks called \newterm{object containers} (\newterm{superblocks}~\cite{Berger00}), as in Figure~\ref{f:ObjectContainer}. 830 854 The header for the container holds information necessary for all objects in the container; 831 855 a trailer may also be used at the end of the container. … … 862 886 863 887 864 \ subsubsection{Container Ownership}888 \paragraph{Container Ownership} 865 889 \label{s:ContainerOwnership} 866 890 … … 894 918 895 919 Additional restrictions may be applied to the movement of containers to prevent active false-sharing. 896 For example, if a container changes ownership through the global heap, then when a thread allocates an object from the newly acquired container itis actively false-sharing even though no objects are passed among threads.920 For example, if a container changes ownership through the global heap, then a thread allocating from the newly acquired container is actively false-sharing even though no objects are passed among threads. 897 921 Note, once the thread frees the object, no more false sharing can occur until the container changes ownership again. 898 922 To prevent this form of false sharing, container movement may be restricted to when all objects in the container are free. 899 One implementation approach that increases the freedom to return a free container to the operating system involves allocating containers using a call like @mmap@, which allows memory at an arbitrary address to be returned versus only storage at the end of the contiguous @sbrk@ area, again pushing storage management complexity back to the operating system.923 One implementation approach that increases the freedom to return a free container to the OS involves allocating containers using a call like @mmap@, which allows memory at an arbitrary address to be returned versus only storage at the end of the contiguous @sbrk@ area, again pushing storage management complexity back to the OS. 900 924 901 925 % \begin{figure} … … 930 954 931 955 932 \ subsubsection{Container Size}956 \paragraph{Container Size} 933 957 \label{s:ContainerSize} 934 958 … … 941 965 However, with more objects in a container, there may be more objects that are unallocated, increasing external fragmentation. 942 966 With smaller containers, not only are there more containers, but a second new problem arises where objects are larger than the container. 943 In general, large objects, \eg greater than 64\,KB, are allocated directly from the operating system and are returned immediately to the operating systemto reduce long-term external fragmentation.967 In general, large objects, \eg greater than 64\,KB, are allocated directly from the OS and are returned immediately to the OS to reduce long-term external fragmentation. 944 968 If the container size is small, \eg 1\,KB, then a 1.5\,KB object is treated as a large object, which is likely to be inappropriate. 945 969 Ideally, it is best to use smaller containers for smaller objects, and larger containers for medium objects, which leads to the issue of locating the container header. … … 970 994 971 995 972 \ subsubsection{Container Free-Lists}996 \paragraph{Container Free-Lists} 973 997 \label{s:containersfreelists} 974 998 … … 1005 1029 1006 1030 1007 \subsubsection{Hybrid Private/Public Heap} 1008 \label{s:HybridPrivatePublicHeap} 1009 1010 Section~\ref{s:Ownership} discusses advantages and disadvantages of public heaps (T:H model and with ownership) and private heaps (thread heaps with ownership). 1011 For thread heaps with ownership, it is possible to combine these approaches into a hybrid approach with both private and public heaps (see~Figure~\ref{f:HybridPrivatePublicHeap}). 1012 The main goal of the hybrid approach is to eliminate locking on thread-local allocation/deallocation, while providing ownership to prevent heap blowup. 1013 In the hybrid approach, a thread first allocates from its private heap and second from its public heap if no free memory exists in the private heap. 1014 Similarly, a thread first deallocates an object to its private heap, and second to the public heap. 1015 Both private and public heaps can allocate/deallocate to/from the global heap if there is no free memory or excess free memory, although an implementation may choose to funnel all interaction with the global heap through one of the heaps. 1016 Note, deallocation from the private to the public (dashed line) is unlikely because there is no obvious advantages unless the public heap provides the only interface to the global heap. 1017 Finally, when a thread frees an object it does not own, the object is either freed immediately to its owner's public heap or put in the freeing thread's private heap for delayed ownership, which allows the freeing thread to temporarily reuse an object before returning it to its owner or batch objects for an owner heap into a single return. 1018 1019 \begin{figure} 1020 \centering 1021 \input{PrivatePublicHeaps.pstex_t} 1022 \caption{Hybrid Private/Public Heap for Per-thread Heaps} 1023 \label{f:HybridPrivatePublicHeap} 1024 % \vspace{10pt} 1025 % \input{RemoteFreeList.pstex_t} 1026 % \caption{Remote Free-List} 1027 % \label{f:RemoteFreeList} 1028 \end{figure} 1029 1030 As mentioned, an implementation may have only one heap interact with the global heap, so the other heap can be simplified. 1031 For example, if only the private heap interacts with the global heap, the public heap can be reduced to a lock-protected free-list of objects deallocated by other threads due to ownership, called a \newterm{remote free-list}. 1032 To avoid heap blowup, the private heap allocates from the remote free-list when it reaches some threshold or it has no free storage. 1033 Since the remote free-list is occasionally cleared during an allocation, this adds to that cost. 1034 Clearing the remote free-list is $O(1)$ if the list can simply be added to the end of the private-heap's free-list, or $O(N)$ if some action must be performed for each freed object. 1035 1036 If only the public heap interacts with other threads and the global heap, the private heap can handle thread-local allocations and deallocations without locking. 1037 In this scenario, the private heap must deallocate storage after reaching a certain threshold to the public heap (and then eventually to the global heap from the public heap) or heap blowup can occur. 1038 If the public heap does the major management, the private heap can be simplified to provide high-performance thread-local allocations and deallocations. 1039 1040 The main disadvantage of each thread having both a private and public heap is the complexity of managing two heaps and their interactions in an allocator. 1041 Interestingly, heap implementations often focus on either a private or public heap, giving the impression a single versus a hybrid approach is being used. 1042 In many case, the hybrid approach is actually being used, but the simpler heap is just folded into the complex heap, even though the operations logically belong in separate heaps. 1043 For example, a remote free-list is actually a simple public-heap, but may be implemented as an integral component of the complex private-heap in an allocator, masking the presence of a hybrid approach. 1044 1045 1046 \subsection{Allocation Buffer} 1031 \subsubsection{Allocation Buffer} 1047 1032 \label{s:AllocationBuffer} 1048 1033 1049 1034 An allocation buffer is reserved memory (see Section~\ref{s:AllocatorComponents}) not yet allocated to the program, and is used for allocating objects when the free list is empty. 1050 1035 That is, rather than requesting new storage for a single object, an entire buffer is requested from which multiple objects are allocated later. 1051 Any heap may use an allocation buffer, resulting in allocation from the buffer before requesting objects (containers) from the global heap or operating system, respectively.1036 Any heap may use an allocation buffer, resulting in allocation from the buffer before requesting objects (containers) from the global heap or OS, respectively. 1052 1037 The allocation buffer reduces contention and the number of global/operating-system calls. 1053 1038 For coalescing, a buffer is split into smaller objects by allocations, and recomposed into larger buffer areas during deallocations. … … 1062 1047 1063 1048 Allocation buffers may increase external fragmentation, since some memory in the allocation buffer may never be allocated. 1064 A smaller allocation buffer reduces the amount of external fragmentation, but increases the number of calls to the global heap or operating system.1049 A smaller allocation buffer reduces the amount of external fragmentation, but increases the number of calls to the global heap or OS. 1065 1050 The allocation buffer also slightly increases internal fragmentation, since a pointer is necessary to locate the next free object in the buffer. 1066 1051 … … 1068 1053 For example, when a container is created, rather than placing all objects within the container on the free list, the objects form an allocation buffer and are allocated from the buffer as allocation requests are made. 1069 1054 This lazy method of constructing objects is beneficial in terms of paging and caching. 1070 For example, although an entire container, possibly spanning several pages, is allocated from the operating system, only a small part of the container is used in the working set of the allocator, reducing the number of pages and cache lines that are brought into higher levels of cache.1071 1072 1073 \subs ection{Lock-Free Operations}1055 For example, although an entire container, possibly spanning several pages, is allocated from the OS, only a small part of the container is used in the working set of the allocator, reducing the number of pages and cache lines that are brought into higher levels of cache. 1056 1057 1058 \subsubsection{Lock-Free Operations} 1074 1059 \label{s:LockFreeOperations} 1075 1060 … … 1194 1179 % A sequence of code that is guaranteed to run to completion before being invoked to accept another input is called serially-reusable code.~\cite{SeriallyReusable}\label{p:SeriallyReusable} 1195 1180 % \end{quote} 1196 % If a KT is preempted during an allocation operation, the operating systemcan schedule another KT on the same CPU, which can begin an allocation operation before the previous operation associated with this CPU has completed, invalidating heap correctness.1181 % If a KT is preempted during an allocation operation, the OS can schedule another KT on the same CPU, which can begin an allocation operation before the previous operation associated with this CPU has completed, invalidating heap correctness. 1197 1182 % Note, the serially-reusable problem can occur in sequential programs with preemption, if the signal handler calls the preempted function, unless the function is serially reusable. 1198 % Essentially, the serially-reusable problem is a race condition on an unprotected critical subsection, where the operating systemis providing the second thread via the signal handler.1183 % Essentially, the serially-reusable problem is a race condition on an unprotected critical subsection, where the OS is providing the second thread via the signal handler. 1199 1184 % 1200 1185 % Library @librseq@~\cite{librseq} was used to perform a fast determination of the CPU and to ensure all memory operations complete on one CPU using @librseq@'s restartable sequences, which restart the critical subsection after undoing its writes, if the critical subsection is preempted. … … 1256 1241 A sequence of code that is guaranteed to run to completion before being invoked to accept another input is called serially-reusable code.~\cite{SeriallyReusable}\label{p:SeriallyReusable} 1257 1242 \end{quote} 1258 If a KT is preempted during an allocation operation, the operating systemcan schedule another KT on the same CPU, which can begin an allocation operation before the previous operation associated with this CPU has completed, invalidating heap correctness.1243 If a KT is preempted during an allocation operation, the OS can schedule another KT on the same CPU, which can begin an allocation operation before the previous operation associated with this CPU has completed, invalidating heap correctness. 1259 1244 Note, the serially-reusable problem can occur in sequential programs with preemption, if the signal handler calls the preempted function, unless the function is serially reusable. 1260 Essentially, the serially-reusable problem is a race condition on an unprotected critical subsection, where the operating systemis providing the second thread via the signal handler.1245 Essentially, the serially-reusable problem is a race condition on an unprotected critical subsection, where the OS is providing the second thread via the signal handler. 1261 1246 1262 1247 Library @librseq@~\cite{librseq} was used to perform a fast determination of the CPU and to ensure all memory operations complete on one CPU using @librseq@'s restartable sequences, which restart the critical subsection after undoing its writes, if the critical subsection is preempted. … … 1273 1258 For the T:H=CPU and 1:1 models, locking is eliminated along the allocation fastpath. 1274 1259 However, T:H=CPU has poor operating-system support to determine the CPU id (heap id) and prevent the serially-reusable problem for KTs. 1275 More operating systemsupport is required to make this model viable, but there is still the serially-reusable problem with user-level threading.1260 More OS support is required to make this model viable, but there is still the serially-reusable problem with user-level threading. 1276 1261 So the 1:1 model had no atomic actions along the fastpath and no special operating-system support requirements. 1277 1262 The 1:1 model still has the serially-reusable problem with user-level threading, which is addressed in Section~\ref{s:UserlevelThreadingSupport}, and the greatest potential for heap blowup for certain allocation patterns. … … 1308 1293 A primary goal of llheap is low latency, hence the name low-latency heap (llheap). 1309 1294 Two forms of latency are internal and external. 1310 Internal latency is the time to perform an allocation, while external latency is time to obtain/return storage from/to the operating system.1295 Internal latency is the time to perform an allocation, while external latency is time to obtain/return storage from/to the OS. 1311 1296 Ideally latency is $O(1)$ with a small constant. 1312 1297 … … 1314 1299 The mitigating factor is that most programs have well behaved allocation patterns, where the majority of allocation operations can be $O(1)$, and heap blowup does not occur without coalescing (although the allocation footprint may be slightly larger). 1315 1300 1316 To obtain $O(1)$ external latency means obtaining one large storage area from the operating systemand subdividing it across all program allocations, which requires a good guess at the program storage high-watermark and potential large external fragmentation.1301 To obtain $O(1)$ external latency means obtaining one large storage area from the OS and subdividing it across all program allocations, which requires a good guess at the program storage high-watermark and potential large external fragmentation. 1317 1302 Excluding real-time operating-systems, operating-system operations are unbounded, and hence some external latency is unavoidable. 1318 1303 The mitigating factor is that operating-system calls can often be reduced if a programmer has a sense of the storage high-watermark and the allocator is capable of using this information (see @malloc_expansion@ \pageref{p:malloc_expansion}). … … 1329 1314 headers per allocation versus containers, 1330 1315 no coalescing to minimize latency, 1331 global heap memory (pool) obtained from the operating systemusing @mmap@ to create and reuse heaps needed by threads,1316 global heap memory (pool) obtained from the OS using @mmap@ to create and reuse heaps needed by threads, 1332 1317 local reserved memory (pool) per heap obtained from global pool, 1333 global reserved memory (pool) obtained from the operating systemusing @sbrk@ call,1318 global reserved memory (pool) obtained from the OS using @sbrk@ call, 1334 1319 optional fast-lookup table for converting allocation requests into bucket sizes, 1335 1320 optional statistic-counters table for accumulating counts of allocation operations. … … 1358 1343 Each heap uses segregated free-buckets that have free objects distributed across 91 different sizes from 16 to 4M. 1359 1344 All objects in a bucket are of the same size. 1360 The number of buckets used is determined dynamically depending on the crossover point from @sbrk@ to @mmap@ allocation using @mallopt( M_MMAP_THRESHOLD )@, \ie small objects managed by the program and large objects managed by the operating system.1345 The number of buckets used is determined dynamically depending on the crossover point from @sbrk@ to @mmap@ allocation using @mallopt( M_MMAP_THRESHOLD )@, \ie small objects managed by the program and large objects managed by the OS. 1361 1346 Each free bucket of a specific size has two lists. 1362 1347 1) A free stack used solely by the KT heap-owner, so push/pop operations do not require locking. … … 1367 1352 Algorithm~\ref{alg:heapObjectAlloc} shows the allocation outline for an object of size $S$. 1368 1353 First, the allocation is divided into small (@sbrk@) or large (@mmap@). 1369 For large allocations, the storage is mapped directly from the operating system.1354 For large allocations, the storage is mapped directly from the OS. 1370 1355 For small allocations, $S$ is quantized into a bucket size. 1371 1356 Quantizing is performed using a binary search over the ordered bucket array. … … 1378 1363 heap's local pool, 1379 1364 global pool, 1380 operating system(@sbrk@).1365 OS (@sbrk@). 1381 1366 1382 1367 \begin{algorithm} … … 1443 1428 Algorithm~\ref{alg:heapObjectFreeOwn} shows the de-allocation (free) outline for an object at address $A$ with ownership. 1444 1429 First, the address is divided into small (@sbrk@) or large (@mmap@). 1445 For large allocations, the storage is unmapped back to the operating system.1430 For large allocations, the storage is unmapped back to the OS. 1446 1431 For small allocations, the bucket associated with the request size is retrieved. 1447 1432 If the bucket is local to the thread, the allocation is pushed onto the thread's associated bucket. … … 3044 3029 3045 3030 \textsf{pt3} is the only memory allocator where the total dynamic memory goes down in the second half of the program lifetime when the memory is freed by the benchmark program. 3046 It makes pt3 the only memory allocator that gives memory back to the operating systemas it is freed by the program.3031 It makes pt3 the only memory allocator that gives memory back to the OS as it is freed by the program. 3047 3032 3048 3033 % FOR 1 THREAD -
doc/papers/llheap/figures/AllocatorComponents.fig
r24d6572 r62d62db 8 8 -2 9 9 1200 2 10 6 1275 2025 2700 262511 10 6 2400 2025 2700 2625 12 11 2 2 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 5 … … 14 13 2 2 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 5 15 14 2700 2025 2700 2325 2400 2325 2400 2025 2700 2025 16 -617 4 2 0 50 -1 2 11 0.0000 2 165 1005 2325 2400 Management\00118 15 -6 19 16 2 2 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 5 … … 61 58 2 2 0 1 0 7 60 -1 13 0.000 0 0 -1 0 0 5 62 59 3300 2700 6300 2700 6300 3000 3300 3000 3300 2700 63 4 0 0 50 -1 2 11 0.0000 2 165 585 3300 1725 Storage\00160 4 0 0 50 -1 2 11 0.0000 2 165 1005 3300 1725 Storage Data\001 64 61 4 2 0 50 -1 0 11 0.0000 2 165 810 3000 1875 free objects\001 65 62 4 2 0 50 -1 0 11 0.0000 2 135 1140 3000 2850 reserve memory\001 66 63 4 1 0 50 -1 0 11 0.0000 2 120 795 2325 1500 Static Zone\001 67 64 4 1 0 50 -1 0 11 0.0000 2 165 1845 4800 1500 Dynamic-Allocation Zone\001 65 4 2 0 50 -1 2 11 0.0000 2 165 1005 2325 2325 Management\001 66 4 2 0 50 -1 2 11 0.0000 2 135 375 2325 2525 Data\001 -
doc/theses/colby_parsons_MMAth/benchmarks/actors/cfa/balance.cfa
r24d6572 r62d62db 31 31 32 32 d_actor ** actor_arr; 33 Allocation receive( d_actor & this, start_msg & msg ) with( this ) {33 allocation receive( d_actor & this, start_msg & msg ) with( this ) { 34 34 for ( i; Set ) { 35 35 *actor_arr[i + gstart] << shared_msg; … … 38 38 } 39 39 40 Allocation receive( d_actor & this, d_msg & msg ) with( this ) {40 allocation receive( d_actor & this, d_msg & msg ) with( this ) { 41 41 if ( recs == rounds ) return Delete; 42 42 if ( recs % Batch == 0 ) { … … 50 50 } 51 51 52 Allocation receive( filler & this, d_msg & msg ) { return Delete; }52 allocation receive( filler & this, d_msg & msg ) { return Delete; } 53 53 54 54 int main( int argc, char * argv[] ) { -
doc/theses/colby_parsons_MMAth/benchmarks/actors/cfa/dynamic.cfa
r24d6572 r62d62db 24 24 25 25 uint64_t start_time; 26 Allocation receive( derived_actor & receiver, derived_msg & msg ) {26 allocation receive( derived_actor & receiver, derived_msg & msg ) { 27 27 if ( msg.cnt >= Times ) { 28 28 printf("%.2f\n", ((double)(bench_time() - start_time)) / ((double)Times) ); // ns -
doc/theses/colby_parsons_MMAth/benchmarks/actors/cfa/executor.cfa
r24d6572 r62d62db 25 25 struct d_msg { inline message; } shared_msg; 26 26 27 Allocation receive( d_actor & this, d_msg & msg ) with( this ) {27 allocation receive( d_actor & this, d_msg & msg ) with( this ) { 28 28 if ( recs == rounds ) return Finished; 29 29 if ( recs % Batch == 0 ) { -
doc/theses/colby_parsons_MMAth/benchmarks/actors/cfa/matrix.cfa
r24d6572 r62d62db 24 24 } 25 25 26 Allocation receive( derived_actor & receiver, derived_msg & msg ) {26 allocation receive( derived_actor & receiver, derived_msg & msg ) { 27 27 for ( unsigned int i = 0; i < yc; i += 1 ) { // multiply X_row by Y_col and sum products 28 28 msg.Z[i] = 0; -
doc/theses/colby_parsons_MMAth/benchmarks/actors/cfa/repeat.cfa
r24d6572 r62d62db 46 46 47 47 Client * cl; 48 Allocation receive( Server & this, IntMsg & msg ) { msg.val = 7; *cl << msg; return Nodelete; }49 Allocation receive( Server & this, CharMsg & msg ) { msg.val = 'x'; *cl << msg; return Nodelete; }50 Allocation receive( Server & this, StateMsg & msg ) { return Finished; }48 allocation receive( Server & this, IntMsg & msg ) { msg.val = 7; *cl << msg; return Nodelete; } 49 allocation receive( Server & this, CharMsg & msg ) { msg.val = 'x'; *cl << msg; return Nodelete; } 50 allocation receive( Server & this, StateMsg & msg ) { return Finished; } 51 51 52 52 void terminateServers( Client & this ) with(this) { … … 56 56 } 57 57 58 Allocation reset( Client & this ) with(this) {58 allocation reset( Client & this ) with(this) { 59 59 times += 1; 60 60 if ( times == Times ) { terminateServers( this ); return Finished; } … … 64 64 } 65 65 66 Allocation process( Client & this ) with(this) {66 allocation process( Client & this ) with(this) { 67 67 this.results++; 68 68 if ( results == 2 * Messages ) { return reset( this ); } … … 70 70 } 71 71 72 Allocation receive( Client & this, IntMsg & msg ) { return process( this ); }73 Allocation receive( Client & this, CharMsg & msg ) { return process( this ); }74 Allocation receive( Client & this, StateMsg & msg ) with(this) {72 allocation receive( Client & this, IntMsg & msg ) { return process( this ); } 73 allocation receive( Client & this, CharMsg & msg ) { return process( this ); } 74 allocation receive( Client & this, StateMsg & msg ) with(this) { 75 75 for ( i; Messages ) { 76 76 servers[i] << intmsg[i]; -
doc/theses/colby_parsons_MMAth/benchmarks/actors/cfa/static.cfa
r24d6572 r62d62db 23 23 24 24 uint64_t start_time; 25 Allocation receive( derived_actor & receiver, derived_msg & msg ) {25 allocation receive( derived_actor & receiver, derived_msg & msg ) { 26 26 if ( msg.cnt >= Times ) { 27 27 printf("%.2f\n", ((double)(bench_time() - start_time)) / ((double)Times) ); // ns -
doc/theses/colby_parsons_MMAth/code/basic_actor_example.cfa
r24d6572 r62d62db 19 19 } 20 20 21 Allocation receive( derived_actor & receiver, derived_msg & msg ) {21 allocation receive( derived_actor & receiver, derived_msg & msg ) { 22 22 printf("The message contained the string: %s\n", msg.word); 23 23 return Finished; // Return allocation status of Finished now that the actor is done work -
doc/user/figures/EHMHierarchy.fig
r24d6572 r62d62db 29 29 1 1 1.00 60.00 90.00 30 30 4950 1950 4950 1725 31 4 1 0 50 -1 0 1 30.0000 2 135 225 1950 1650 IO\00132 4 1 0 50 -1 0 1 30.0000 2 135 915 4950 1650 Arithmetic\00133 4 1 0 50 -1 0 1 30.0000 2 150 330 1350 2100 File\00134 4 1 0 50 -1 0 1 30.0000 2 135 735 2550 2100 Network\00135 4 1 0 50 -1 0 1 30.0000 2 180 1215 3750 2100 DivideByZero\00136 4 1 0 50 -1 0 1 30.0000 2 150 810 4950 2100 Overflow\00137 4 1 0 50 -1 0 1 30.0000 2 150 915 6000 2100 Underflow\00138 4 1 0 50 -1 0 1 30.0000 2 180 855 3450 1200 Exception\00131 4 1 0 50 -1 0 12 0.0000 2 135 225 1950 1650 IO\001 32 4 1 0 50 -1 0 12 0.0000 2 135 915 4950 1650 Arithmetic\001 33 4 1 0 50 -1 0 12 0.0000 2 150 330 1350 2100 File\001 34 4 1 0 50 -1 0 12 0.0000 2 135 735 2550 2100 Network\001 35 4 1 0 50 -1 0 12 0.0000 2 180 1215 3750 2100 DivideByZero\001 36 4 1 0 50 -1 0 12 0.0000 2 150 810 4950 2100 Overflow\001 37 4 1 0 50 -1 0 12 0.0000 2 150 915 6000 2100 Underflow\001 38 4 1 0 50 -1 0 12 0.0000 2 180 855 3450 1200 Exception\001 -
doc/user/user.tex
r24d6572 r62d62db 11 11 %% Created On : Wed Apr 6 14:53:29 2016 12 12 %% Last Modified By : Peter A. Buhr 13 %% Last Modified On : Mon Aug 22 23:43:30 202214 %% Update Count : 55 0313 %% Last Modified On : Mon Jun 5 21:18:29 2023 14 %% Update Count : 5521 15 15 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 16 16 … … 108 108 \huge \CFA Team (past and present) \medskip \\ 109 109 \Large Andrew Beach, Richard Bilson, Michael Brooks, Peter A. Buhr, Thierry Delisle, \smallskip \\ 110 \Large Glen Ditchfield, Rodolfo G. Esteves, Aaron Moss, Colby Parsons, Rob Schluntz,\smallskip \\111 \Large Fangren Yu, Mubeen Zulfiqar110 \Large Glen Ditchfield, Rodolfo G. Esteves, Jiada Liang, Aaron Moss, Colby Parsons \smallskip \\ 111 \Large Rob Schluntz, Fangren Yu, Mubeen Zulfiqar 112 112 }% author 113 113 … … 169 169 Like \Index*[C++]{\CC{}}, there may be both old and new ways to achieve the same effect. 170 170 For example, the following programs compare the C, \CFA, and \CC I/O mechanisms, where the programs output the same result. 171 \begin{ flushleft}172 \begin{tabular}{@{}l @{\hspace{1em}}l@{\hspace{1em}}l@{}}173 \multicolumn{1}{@{}c @{\hspace{1em}}}{\textbf{C}} & \multicolumn{1}{c}{\textbf{\CFA}} & \multicolumn{1}{c@{}}{\textbf{\CC}} \\171 \begin{center} 172 \begin{tabular}{@{}lll@{}} 173 \multicolumn{1}{@{}c}{\textbf{C}} & \multicolumn{1}{c}{\textbf{\CFA}} & \multicolumn{1}{c@{}}{\textbf{\CC}} \\ 174 174 \begin{cfa}[tabsize=3] 175 175 #include <stdio.h>$\indexc{stdio.h}$ … … 199 199 \end{cfa} 200 200 \end{tabular} 201 \end{ flushleft}201 \end{center} 202 202 While \CFA I/O \see{\VRef{s:StreamIOLibrary}} looks similar to \Index*[C++]{\CC{}}, there are important differences, such as automatic spacing between variables and an implicit newline at the end of the expression list, similar to \Index*{Python}~\cite{Python}. 203 203 … … 856 856 still works. 857 857 Nevertheless, reversing the default action would have a non-trivial effect on case actions that compound, such as the above example of processing shell arguments. 858 Therefore, to preserve backwards compatibility, it is necessary to introduce a new kind of ©switch© statement, called \Indexc{choose}, with no implicit fall-through semantics and an explicit fall-through if the last statement of a case-clause ends with the new keyword \Indexc{fallthrough}/\ Indexc{fallthru}, \eg:858 Therefore, to preserve backwards compatibility, it is necessary to introduce a new kind of ©switch© statement, called \Indexc{choose}, with no implicit fall-through semantics and an explicit fall-through if the last statement of a case-clause ends with the new keyword \Indexc{fallthrough}/\-\Indexc{fallthru}, \eg: 859 859 \begin{cfa} 860 860 ®choose® ( i ) { … … 1167 1167 \end{cfa} 1168 1168 \end{itemize} 1169 \R{Warning}: specifying the down-to range maybe unex cepted because the loop control \emph{implicitly} switches the L and H values (and toggles the increment/decrement for I):1169 \R{Warning}: specifying the down-to range maybe unexpected because the loop control \emph{implicitly} switches the L and H values (and toggles the increment/decrement for I): 1170 1170 \begin{cfa} 1171 1171 for ( i; 1 ~ 10 ) ${\C[1.5in]{// up range}$ … … 1173 1173 for ( i; ®10 -~ 1® ) ${\C{// \R{WRONG down range!}}\CRT}$ 1174 1174 \end{cfa} 1175 The reason for this sema tics is that the range direction can be toggled by adding/removing the minus, ©'-'©, versus interchanging the L and H expressions, which has a greater chance of introducing errors.1175 The reason for this semantics is that the range direction can be toggled by adding/removing the minus, ©'-'©, versus interchanging the L and H expressions, which has a greater chance of introducing errors. 1176 1176 1177 1177 … … 2256 2256 Days days = Mon; // enumeration type declaration and initialization 2257 2257 \end{cfa} 2258 The set of enums are injected into the variable namespace at the definition scope. 2259 Hence, enums may be overloaded with enum/variable/function names. 2260 \begin{cfa} 2258 The set of enums is injected into the variable namespace at the definition scope. 2259 Hence, enums may be overloaded with variable, enum, and function names. 2260 \begin{cfa} 2261 int Foo; $\C{// type/variable separate namespaces}$ 2261 2262 enum Foo { Bar }; 2262 2263 enum Goo { Bar }; $\C[1.75in]{// overload Foo.Bar}$ 2263 int Foo; $\C{// type/variable separate namespace}$2264 2264 double Bar; $\C{// overload Foo.Bar, Goo.Bar}\CRT$ 2265 2265 \end{cfa} … … 2301 2301 Hence, the value of enum ©Mon© is 0, ©Tue© is 1, ...\,, ©Sun© is 6. 2302 2302 If an enum value is specified, numbering continues by one from that value for subsequent unnumbered enums. 2303 If an enum value is a nexpression, the compiler performs constant-folding to obtain a constant value.2303 If an enum value is a \emph{constant} expression, the compiler performs constant-folding to obtain a constant value. 2304 2304 2305 2305 \CFA allows other integral types with associated values. … … 2313 2313 \begin{cfa} 2314 2314 // non-integral numeric 2315 enum( ®double® ) Math { PI_2 = 1.570796, PI = 3.141597, 2315 enum( ®double® ) Math { PI_2 = 1.570796, PI = 3.141597, E = 2.718282 } 2316 2316 // pointer 2317 enum( ®char *® ) Name { Fred = "Fred", Mary = "Mary", 2317 enum( ®char *® ) Name { Fred = "Fred", Mary = "Mary", Jane = "Jane" }; 2318 2318 int i, j, k; 2319 2319 enum( ®int *® ) ptr { I = &i, J = &j, K = &k }; 2320 enum( ®int &® ) ref { I = i, J = j,K = k };2320 enum( ®int &® ) ref { I = i, J = j, K = k }; 2321 2321 // tuple 2322 2322 enum( ®[int, int]® ) { T = [ 1, 2 ] }; … … 2361 2361 \begin{cfa} 2362 2362 enum( char * ) Name2 { ®inline Name®, Jack = "Jack", Jill = "Jill" }; 2363 enum ®/* inferred */® 2363 enum ®/* inferred */® Name3 { ®inline Name2®, Sue = "Sue", Tom = "Tom" }; 2364 2364 \end{cfa} 2365 2365 Enumeration ©Name2© inherits all the enums and their values from enumeration ©Name© by containment, and a ©Name© enumeration is a subtype of enumeration ©Name2©. … … 3818 3818 "[ output-file (default stdout) ] ]"; 3819 3819 } // choose 3820 } catch( ® Open_Failure® * ex; ex->istream == &in ) {3820 } catch( ®open_failure® * ex; ex->istream == &in ) { $\C{// input file errors}$ 3821 3821 ®exit® | "Unable to open input file" | argv[1]; 3822 } catch( ® Open_Failure® * ex; ex->ostream == &out ) {3822 } catch( ®open_failure® * ex; ex->ostream == &out ) { $\C{// output file errors}$ 3823 3823 ®close®( in ); $\C{// optional}$ 3824 3824 ®exit® | "Unable to open output file" | argv[2]; … … 4038 4038 4039 4039 \item 4040 \Indexc{sepDisable}\index{manipulator!sepDisable@©sepDisable©} and \Indexc{sepEnable}\index{manipulator!sepEnable@©sepEnable©} toggle printing the separator.4040 \Indexc{sepDisable}\index{manipulator!sepDisable@©sepDisable©} and \Indexc{sepEnable}\index{manipulator!sepEnable@©sepEnable©} globally toggle printing the separator. 4041 4041 \begin{cfa}[belowskip=0pt] 4042 4042 sout | sepDisable | 1 | 2 | 3; $\C{// turn off implicit separator}$ … … 4053 4053 4054 4054 \item 4055 \Indexc{sepOn}\index{manipulator!sepOn@©sepOn©} and \Indexc{sepOff}\index{manipulator!sepOff@©sepOff©} toggle printing the separator with respect to the next printed item, and then return to the global separator setting.4055 \Indexc{sepOn}\index{manipulator!sepOn@©sepOn©} and \Indexc{sepOff}\index{manipulator!sepOff@©sepOff©} locally toggle printing the separator with respect to the next printed item, and then return to the global separator setting. 4056 4056 \begin{cfa}[belowskip=0pt] 4057 4057 sout | 1 | sepOff | 2 | 3; $\C{// turn off implicit separator for the next item}$ … … 4129 4129 6 4130 4130 \end{cfa} 4131 Note, a terminating ©nl© is merged (overrides) with the implicit newline at the end of the ©sout© expression, otherwise it is impossible to toprint a single newline4131 Note, a terminating ©nl© is merged (overrides) with the implicit newline at the end of the ©sout© expression, otherwise it is impossible to print a single newline 4132 4132 \item 4133 4133 \Indexc{nlOn}\index{manipulator!nlOn@©nlOn©} implicitly prints a newline at the end of each output expression. -
driver/cc1.cc
r24d6572 r62d62db 10 10 // Created On : Fri Aug 26 14:23:51 2005 11 11 // Last Modified By : Peter A. Buhr 12 // Last Modified On : Thu Feb 17 18:04:23 202213 // Update Count : 42 212 // Last Modified On : Fri Jun 9 11:36:44 2023 13 // Update Count : 423 14 14 // 15 15 … … 385 385 // strip inappropriate flags with an argument 386 386 387 } else if ( arg == "-auxbase" || arg == "-auxbase-strip" || arg == "-dumpbase" || arg == "-dumpdir" ) { 387 } else if ( arg == "-auxbase" || arg == "-auxbase-strip" || 388 arg == "-dumpbase" || arg == "-dumpbase-ext" || arg == "-dumpdir" ) { 388 389 i += 1; 389 390 #ifdef __DEBUG_H__ -
libcfa/src/concurrency/actor.hfa
r24d6572 r62d62db 13 13 #endif // CFA_DEBUG 14 14 15 #define DEBUG_ABORT( cond, string ) CFA_DEBUG( if ( cond ) abort( string ) ) 16 15 17 // Define the default number of processors created in the executor. Must be greater than 0. 16 18 #define __DEFAULT_EXECUTOR_PROCESSORS__ 2 … … 42 44 struct executor; 43 45 44 enum Allocation { Nodelete, Delete, Destroy, Finished }; // allocation status45 46 typedef Allocation (*__receive_fn)(actor &, message &);46 enum allocation { Nodelete, Delete, Destroy, Finished }; // allocation status 47 48 typedef allocation (*__receive_fn)(actor &, message &); 47 49 struct request { 48 50 actor * receiver; … … 393 395 struct actor { 394 396 size_t ticket; // executor-queue handle 395 Allocation allocation_; // allocation action397 allocation allocation_; // allocation action 396 398 inline virtual_dtor; 397 399 }; … … 400 402 // Once an actor is allocated it must be sent a message or the actor system cannot stop. Hence, its receive 401 403 // member must be called to end it 402 verifyf( __actor_executor_, "Creating actor before calling start_actor_system() can cause undefined behaviour.\n" );404 DEBUG_ABORT( __actor_executor_ == 0p, "Creating actor before calling start_actor_system() can cause undefined behaviour.\n" ); 403 405 allocation_ = Nodelete; 404 406 ticket = __get_next_ticket( *__actor_executor_ ); … … 430 432 431 433 struct message { 432 Allocation allocation_; // allocation action434 allocation allocation_; // allocation action 433 435 inline virtual_dtor; 434 436 }; … … 437 439 this.allocation_ = Nodelete; 438 440 } 439 static inline void ?{}( message & this, Allocation allocation) {440 memcpy( &this.allocation_, &alloc ation, sizeof(allocation) ); // optimization to elide ctor441 verifyf( this.allocation_ != Finished, "The Finished Allocation status is not supported for message types.\n");441 static inline void ?{}( message & this, allocation alloc ) { 442 memcpy( &this.allocation_, &alloc, sizeof(allocation) ); // optimization to elide ctor 443 DEBUG_ABORT( this.allocation_ == Finished, "The Finished allocation status is not supported for message types.\n" ); 442 444 } 443 445 static inline void ^?{}( message & this ) with(this) { … … 453 455 } // switch 454 456 } 455 static inline void set_allocation( message & this, Allocation state ) {457 static inline void set_allocation( message & this, allocation state ) { 456 458 this.allocation_ = state; 457 459 } 458 460 459 461 static inline void deliver_request( request & this ) { 462 DEBUG_ABORT( this.receiver->ticket == (unsigned long int)MAX, "Attempted to send message to deleted/dead actor\n" ); 460 463 this.receiver->allocation_ = this.fn( *this.receiver, *this.msg ); 461 464 check_message( *this.msg ); … … 631 634 632 635 static inline void send( actor & this, request & req ) { 633 verifyf( this.ticket != (unsigned long int)MAX, "Attempted to send message to deleted/dead actor\n" );636 DEBUG_ABORT( this.ticket == (unsigned long int)MAX, "Attempted to send message to deleted/dead actor\n" ); 634 637 send( *__actor_executor_, req, this.ticket ); 635 638 } … … 680 683 // assigned at creation to __base_msg_finished to avoid unused message warning 681 684 message __base_msg_finished @= { .allocation_ : Finished }; 682 struct __ DeleteMsg { inline message; } DeleteMsg = __base_msg_finished;683 struct __ DestroyMsg { inline message; } DestroyMsg = __base_msg_finished;684 struct __ FinishedMsg { inline message; } FinishedMsg = __base_msg_finished;685 686 Allocation receive( actor & this, __DeleteMsg& msg ) { return Delete; }687 Allocation receive( actor & this, __DestroyMsg& msg ) { return Destroy; }688 Allocation receive( actor & this, __FinishedMsg& msg ) { return Finished; }689 685 struct __delete_msg_t { inline message; } delete_msg = __base_msg_finished; 686 struct __destroy_msg_t { inline message; } destroy_msg = __base_msg_finished; 687 struct __finished_msg_t { inline message; } finished_msg = __base_msg_finished; 688 689 allocation receive( actor & this, __delete_msg_t & msg ) { return Delete; } 690 allocation receive( actor & this, __destroy_msg_t & msg ) { return Destroy; } 691 allocation receive( actor & this, __finished_msg_t & msg ) { return Finished; } 692 -
libcfa/src/concurrency/atomic.hfa
r24d6572 r62d62db 10 10 // Created On : Thu May 25 15:22:46 2023 11 11 // Last Modified By : Peter A. Buhr 12 // Last Modified On : Thu May 25 15:24:45202313 // Update Count : 112 // Last Modified On : Fri Jun 9 13:36:47 2023 13 // Update Count : 46 14 14 // 15 15 16 #define LOAD( lock ) (__atomic_load_n( &(lock), __ATOMIC_SEQ_CST )) 17 #define LOADM( lock, memorder ) (__atomic_load_n( &(lock), memorder )) 18 #define STORE( lock, assn ) (__atomic_store_n( &(lock), assn, __ATOMIC_SEQ_CST )) 19 #define STOREM( lock, assn, memorder ) (__atomic_store_n( &(lock), assn, memorder )) 20 #define CLR( lock ) (__atomic_clear( &(lock), __ATOMIC_RELEASE )) 21 #define CLRM( lock, memorder ) (__atomic_clear( &(lock), memorder )) 22 #define TAS( lock ) (__atomic_test_and_set( &(lock), __ATOMIC_ACQUIRE )) 23 #define TASM( lock, memorder ) (__atomic_test_and_set( &(lock), memorder )) 24 #define CAS( change, comp, assn ) ({typeof(comp) __temp = (comp); __atomic_compare_exchange_n( &(change), &(__temp), (assn), false, __ATOMIC_SEQ_CST, __ATOMIC_SEQ_CST ); }) 25 #define CASM( change, comp, assn, memorder... ) ({typeof(comp) * __temp = &(comp); __atomic_compare_exchange_n( &(change), &(__temp), (assn), false, memorder, memorder ); }) 26 #define CASV( change, comp, assn ) (__atomic_compare_exchange_n( &(change), &(comp), (assn), false, __ATOMIC_SEQ_CST, __ATOMIC_SEQ_CST )) 27 #define CASVM( change, comp, assn, memorder... ) (__atomic_compare_exchange_n( &(change), &(comp), (assn), false, memorder, memorder )) 28 #define FAS( change, assn ) (__atomic_exchange_n( &(change), (assn), __ATOMIC_SEQ_CST )) 29 #define FASM( change, assn, memorder ) (__atomic_exchange_n( &(change), (assn), memorder )) 30 #define FAI( change, Inc ) (__atomic_fetch_add( &(change), (Inc), __ATOMIC_SEQ_CST )) 31 #define FAIM( change, Inc, memorder ) (__atomic_fetch_add( &(change), (Inc), memorder )) 16 #define LOAD( val ) (LOADM( val, __ATOMIC_SEQ_CST)) 17 #define LOADM( val, memorder ) (__atomic_load_n( &(val), memorder)) 18 19 #define STORE( val, assn ) (STOREM( val, assn, __ATOMIC_SEQ_CST)) 20 #define STOREM( val, assn, memorder ) (__atomic_store_n( &(val), assn, memorder)) 21 22 #define TAS( lock ) (TASM( lock, __ATOMIC_ACQUIRE)) 23 #define TASM( lock, memorder ) (__atomic_test_and_set( &(lock), memorder)) 24 25 #define TASCLR( lock ) (TASCLRM( lock, __ATOMIC_RELEASE)) 26 #define TASCLRM( lock, memorder ) (__atomic_clear( &(lock), memorder)) 27 28 #define FAS( assn, replace ) (FASM(assn, replace, __ATOMIC_SEQ_CST)) 29 #define FASM( assn, replace, memorder ) (__atomic_exchange_n( &(assn), (replace), memorder)) 30 31 #define FAI( assn, Inc ) (__atomic_fetch_add( &(assn), (Inc), __ATOMIC_SEQ_CST)) 32 #define FAIM( assn, Inc, memorder ) (__atomic_fetch_add( &(assn), (Inc), memorder)) 33 34 // Use __sync because __atomic with 128-bit CAA can result in calls to pthread_mutex_lock. 35 36 // #define CAS( assn, comp, replace ) (CASM( assn, comp, replace, __ATOMIC_SEQ_CST, __ATOMIC_SEQ_CST)) 37 // #define CASM( assn, comp, replace, memorder... ) ({ \ 38 // typeof(comp) __temp = (comp); \ 39 // __atomic_compare_exchange_n( &(assn), &(__temp), (replace), false, memorder ); \ 40 // }) 41 #define CAS( assn, comp, replace ) (__sync_bool_compare_and_swap( &assn, comp, replace)) 42 #define CASM( assn, comp, replace, memorder... ) _Static_assert( false, "memory order unsupported for CAS macro" ); 43 44 // #define CASV( assn, comp, replace ) (__atomic_compare_exchange_n( &(assn), &(comp), (replace), false, __ATOMIC_SEQ_CST, __ATOMIC_SEQ_CST )) 45 // #define CASVM( assn, comp, replace, memorder... ) (__atomic_compare_exchange_n( &(assn), &(comp), (replace), false, memorder, memorder )) 46 #define CASV( assn, comp, replace ) ({ \ 47 typeof(comp) temp = comp; \ 48 typeof(comp) old = __sync_val_compare_and_swap( &(assn), (comp), (replace) ); \ 49 old == temp ? true : (comp = old, false); \ 50 }) 51 #define CASVM( assn, comp, replace, memorder... ) _Static_assert( false, "memory order unsupported for CASV macro" ); -
libcfa/src/concurrency/channel.hfa
r24d6572 r62d62db 51 51 vtable(channel_closed) channel_closed_vt; 52 52 53 static inline bool is_insert( channel_closed & e ) { return e lem != 0p; }54 static inline bool is_remove( channel_closed & e ) { return e lem == 0p; }53 static inline bool is_insert( channel_closed & e ) { return e.elem != 0p; } 54 static inline bool is_remove( channel_closed & e ) { return e.elem == 0p; } 55 55 56 56 // #define CHAN_STATS // define this to get channel stats printed in dtor -
libcfa/src/concurrency/locks.hfa
r24d6572 r62d62db 32 32 #include "select.hfa" 33 33 34 #include <fstream.hfa>35 36 34 // futex headers 37 35 #include <linux/futex.h> /* Definition of FUTEX_* constants */ -
libcfa/src/containers/lockfree.hfa
r24d6572 r62d62db 199 199 200 200 forall( T & ) 201 struct LinkData { 202 T * volatile top; // pointer to stack top 203 uintptr_t count; // count each push 204 }; 205 206 forall( T & ) 201 207 union Link { 202 struct { // 32/64-bit x 2 203 T * volatile top; // pointer to stack top 204 uintptr_t count; // count each push 205 }; 208 LinkData(T) data; 206 209 #if __SIZEOF_INT128__ == 16 207 210 __int128 // gcc, 128-bit integer … … 220 223 void ?{}( StackLF(T) & this ) with(this) { stack.atom = 0; } 221 224 222 T * top( StackLF(T) & this ) with(this) { return stack. top; }225 T * top( StackLF(T) & this ) with(this) { return stack.data.top; } 223 226 224 227 void push( StackLF(T) & this, T & n ) with(this) { 225 228 *( &n )`next = stack; // atomic assignment unnecessary, or use CAA 226 229 for () { // busy wait 227 if ( __atomic_compare_exchange_n( &stack.atom, &( &n )`next->atom, (Link(T))@{ {&n, ( &n )`next->count + 1} }.atom, false, __ATOMIC_SEQ_CST, __ATOMIC_SEQ_CST ) ) break; // attempt to update top node230 if ( __atomic_compare_exchange_n( &stack.atom, &( &n )`next->atom, (Link(T))@{ (LinkData(T))@{ &n, ( &n )`next->data.count + 1} }.atom, false, __ATOMIC_SEQ_CST, __ATOMIC_SEQ_CST ) ) break; // attempt to update top node 228 231 } // for 229 232 } // push … … 232 235 Link(T) t @= stack; // atomic assignment unnecessary, or use CAA 233 236 for () { // busy wait 234 if ( t.top == 0p ) return 0p; // empty stack ? 235 if ( __atomic_compare_exchange_n( &stack.atom, &t.atom, (Link(T))@{ {( t.top )`next->top, t.count} }.atom, false, __ATOMIC_SEQ_CST, __ATOMIC_SEQ_CST ) ) return t.top; // attempt to update top node 237 if ( t.data.top == 0p ) return 0p; // empty stack ? 238 Link(T) * next = ( t.data.top )`next; 239 if ( __atomic_compare_exchange_n( &stack.atom, &t.atom, (Link(T))@{ (LinkData(T))@{ next->data.top, t.data.count } }.atom, false, __ATOMIC_SEQ_CST, __ATOMIC_SEQ_CST ) ) return t.data.top; // attempt to update top node 236 240 } // for 237 241 } // pop … … 239 243 bool unsafe_remove( StackLF(T) & this, T * node ) with(this) { 240 244 Link(T) * link = &stack; 241 for() { 242 T * next = link->top; 243 if( next == node ) { 244 link->top = ( node )`next->top; 245 for () { 246 // TODO: Avoiding some problems with double fields access. 247 LinkData(T) * data = &link->data; 248 T * next = (T *)&(*data).top; 249 if ( next == node ) { 250 data->top = ( node )`next->data.top; 245 251 return true; 246 252 } 247 if ( next == 0p ) return false;253 if ( next == 0p ) return false; 248 254 link = ( next )`next; 249 255 } -
libcfa/src/fstream.cfa
r24d6572 r62d62db 10 10 // Created On : Wed May 27 17:56:53 2015 11 11 // Last Modified By : Peter A. Buhr 12 // Last Modified On : Sat Apr 9 14:55:54 202213 // Update Count : 51 512 // Last Modified On : Mon Jun 5 22:00:23 2023 13 // Update Count : 518 14 14 // 15 15 … … 117 117 } // for 118 118 if ( file == 0p ) { 119 throw ( Open_Failure){ os };119 throw (open_failure){ os }; 120 120 // abort | IO_MSG "open output file \"" | name | "\"" | nl | strerror( errno ); 121 121 } // if … … 137 137 } // for 138 138 if ( ret == EOF ) { 139 throw ( Close_Failure){ os };139 throw (close_failure){ os }; 140 140 // abort | IO_MSG "close output" | nl | strerror( errno ); 141 141 } // if … … 145 145 ofstream & write( ofstream & os, const char data[], size_t size ) { 146 146 if ( fail( os ) ) { 147 throw ( Write_Failure){ os };147 throw (write_failure){ os }; 148 148 // abort | IO_MSG "attempt write I/O on failed stream"; 149 149 } // if 150 150 151 151 if ( fwrite( data, 1, size, (FILE *)(os.file$) ) != size ) { 152 throw ( Write_Failure){ os };152 throw (write_failure){ os }; 153 153 // abort | IO_MSG "write" | nl | strerror( errno ); 154 154 } // if … … 240 240 } // for 241 241 if ( file == 0p ) { 242 throw ( Open_Failure){ is };242 throw (open_failure){ is }; 243 243 // abort | IO_MSG "open input file \"" | name | "\"" | nl | strerror( errno ); 244 244 } // if … … 260 260 } // for 261 261 if ( ret == EOF ) { 262 throw ( Close_Failure){ is };262 throw (close_failure){ is }; 263 263 // abort | IO_MSG "close input" | nl | strerror( errno ); 264 264 } // if … … 268 268 ifstream & read( ifstream & is, char data[], size_t size ) { 269 269 if ( fail( is ) ) { 270 throw ( Read_Failure){ is };270 throw (read_failure){ is }; 271 271 // abort | IO_MSG "attempt read I/O on failed stream"; 272 272 } // if 273 273 274 274 if ( fread( data, size, 1, (FILE *)(is.file$) ) == 0 ) { 275 throw ( Read_Failure){ is };275 throw (read_failure){ is }; 276 276 // abort | IO_MSG "read" | nl | strerror( errno ); 277 277 } // if … … 318 318 319 319 320 static vtable( Open_Failure) Open_Failure_vt;320 static vtable(open_failure) open_failure_vt; 321 321 322 322 // exception I/O constructors 323 void ?{}( Open_Failure & ex, ofstream & ostream ) with(ex) {324 virtual_table = & Open_Failure_vt;323 void ?{}( open_failure & ex, ofstream & ostream ) with(ex) { 324 virtual_table = &open_failure_vt; 325 325 ostream = &ostream; 326 326 tag = 1; 327 327 } // ?{} 328 328 329 void ?{}( Open_Failure & ex, ifstream & istream ) with(ex) {330 virtual_table = & Open_Failure_vt;329 void ?{}( open_failure & ex, ifstream & istream ) with(ex) { 330 virtual_table = &open_failure_vt; 331 331 istream = &istream; 332 332 tag = 0; … … 334 334 335 335 336 static vtable( Close_Failure) Close_Failure_vt;336 static vtable(close_failure) close_failure_vt; 337 337 338 338 // exception I/O constructors 339 void ?{}( Close_Failure & ex, ofstream & ostream ) with(ex) {340 virtual_table = & Close_Failure_vt;339 void ?{}( close_failure & ex, ofstream & ostream ) with(ex) { 340 virtual_table = &close_failure_vt; 341 341 ostream = &ostream; 342 342 tag = 1; 343 343 } // ?{} 344 344 345 void ?{}( Close_Failure & ex, ifstream & istream ) with(ex) {346 virtual_table = & Close_Failure_vt;345 void ?{}( close_failure & ex, ifstream & istream ) with(ex) { 346 virtual_table = &close_failure_vt; 347 347 istream = &istream; 348 348 tag = 0; … … 350 350 351 351 352 static vtable( Write_Failure) Write_Failure_vt;352 static vtable(write_failure) write_failure_vt; 353 353 354 354 // exception I/O constructors 355 void ?{}( Write_Failure & ex, ofstream & ostream ) with(ex) {356 virtual_table = & Write_Failure_vt;355 void ?{}( write_failure & ex, ofstream & ostream ) with(ex) { 356 virtual_table = &write_failure_vt; 357 357 ostream = &ostream; 358 358 tag = 1; 359 359 } // ?{} 360 360 361 void ?{}( Write_Failure & ex, ifstream & istream ) with(ex) {362 virtual_table = & Write_Failure_vt;361 void ?{}( write_failure & ex, ifstream & istream ) with(ex) { 362 virtual_table = &write_failure_vt; 363 363 istream = &istream; 364 364 tag = 0; … … 366 366 367 367 368 static vtable( Read_Failure) Read_Failure_vt;368 static vtable(read_failure) read_failure_vt; 369 369 370 370 // exception I/O constructors 371 void ?{}( Read_Failure & ex, ofstream & ostream ) with(ex) {372 virtual_table = & Read_Failure_vt;371 void ?{}( read_failure & ex, ofstream & ostream ) with(ex) { 372 virtual_table = &read_failure_vt; 373 373 ostream = &ostream; 374 374 tag = 1; 375 375 } // ?{} 376 376 377 void ?{}( Read_Failure & ex, ifstream & istream ) with(ex) {378 virtual_table = & Read_Failure_vt;377 void ?{}( read_failure & ex, ifstream & istream ) with(ex) { 378 virtual_table = &read_failure_vt; 379 379 istream = &istream; 380 380 tag = 0; 381 381 } // ?{} 382 382 383 // void throw Open_Failure( ofstream & ostream ) {384 // Open_Failure exc = { ostream };383 // void throwopen_failure( ofstream & ostream ) { 384 // open_failure exc = { ostream }; 385 385 // } 386 386 387 // void throw Open_Failure( ifstream & istream ) {388 // Open_Failure exc = { istream };387 // void throwopen_failure( ifstream & istream ) { 388 // open_failure exc = { istream }; 389 389 // } 390 390 -
libcfa/src/fstream.hfa
r24d6572 r62d62db 10 10 // Created On : Wed May 27 17:56:53 2015 11 11 // Last Modified By : Peter A. Buhr 12 // Last Modified On : Sun Oct 10 09:37:32 202113 // Update Count : 24 312 // Last Modified On : Mon Jun 5 22:00:20 2023 13 // Update Count : 246 14 14 // 15 15 … … 137 137 138 138 139 exception Open_Failure {139 exception open_failure { 140 140 union { 141 141 ofstream * ostream; … … 146 146 }; 147 147 148 void ?{}( Open_Failure & this, ofstream & );149 void ?{}( Open_Failure & this, ifstream & );148 void ?{}( open_failure & this, ofstream & ); 149 void ?{}( open_failure & this, ifstream & ); 150 150 151 exception Close_Failure {151 exception close_failure { 152 152 union { 153 153 ofstream * ostream; … … 158 158 }; 159 159 160 void ?{}( Close_Failure & this, ofstream & );161 void ?{}( Close_Failure & this, ifstream & );160 void ?{}( close_failure & this, ofstream & ); 161 void ?{}( close_failure & this, ifstream & ); 162 162 163 exception Write_Failure {163 exception write_failure { 164 164 union { 165 165 ofstream * ostream; … … 170 170 }; 171 171 172 void ?{}( Write_Failure & this, ofstream & );173 void ?{}( Write_Failure & this, ifstream & );172 void ?{}( write_failure & this, ofstream & ); 173 void ?{}( write_failure & this, ifstream & ); 174 174 175 exception Read_Failure {175 exception read_failure { 176 176 union { 177 177 ofstream * ostream; … … 182 182 }; 183 183 184 void ?{}( Read_Failure & this, ofstream & );185 void ?{}( Read_Failure & this, ifstream & );184 void ?{}( read_failure & this, ofstream & ); 185 void ?{}( read_failure & this, ifstream & ); 186 186 187 187 // Local Variables: // -
libcfa/src/math.trait.hfa
r24d6572 r62d62db 10 10 // Created On : Fri Jul 16 15:40:52 2021 11 11 // Last Modified By : Peter A. Buhr 12 // Last Modified On : T hu Feb 2 11:36:56202313 // Update Count : 2 012 // Last Modified On : Tue Jun 6 07:59:17 2023 13 // Update Count : 24 14 14 // 15 15 … … 17 17 18 18 forall( U ) 19 trait Not {19 trait not { 20 20 void ?{}( U &, zero_t ); 21 21 int !?( U ); 22 }; // Not22 }; // not 23 23 24 forall( T | Not( T ) )25 trait Equality {24 forall( T | not( T ) ) 25 trait equality { 26 26 int ?==?( T, T ); 27 27 int ?!=?( T, T ); 28 }; // Equality28 }; // equality 29 29 30 forall( U | Equality( U ) )31 trait Relational {30 forall( U | equality( U ) ) 31 trait relational { 32 32 int ?<?( U, U ); 33 33 int ?<=?( U, U ); 34 34 int ?>?( U, U ); 35 35 int ?>=?( U, U ); 36 }; // Relational36 }; // relational 37 37 38 38 forall ( T ) 39 trait Signed { 39 trait Signed { // must be capitalized, conflict with keyword signed 40 40 T +?( T ); 41 41 T -?( T ); … … 44 44 45 45 forall( U | Signed( U ) ) 46 trait Additive {46 trait additive { 47 47 U ?+?( U, U ); 48 48 U ?-?( U, U ); 49 49 U ?+=?( U &, U ); 50 50 U ?-=?( U &, U ); 51 }; // Additive51 }; // additive 52 52 53 forall( T | Additive( T ) )54 trait Incdec {53 forall( T | additive( T ) ) 54 trait inc_dec { 55 55 void ?{}( T &, one_t ); 56 56 // T ?++( T & ); … … 58 58 // T ?--( T & ); 59 59 // T --?( T & ); 60 }; // Incdec60 }; // inc_dec 61 61 62 forall( U | Incdec( U ) )63 trait Multiplicative {62 forall( U | inc_dec( U ) ) 63 trait multiplicative { 64 64 U ?*?( U, U ); 65 65 U ?/?( U, U ); 66 66 U ?%?( U, U ); 67 67 U ?/=?( U &, U ); 68 }; // Multiplicative68 }; // multiplicative 69 69 70 forall( T | Relational( T ) | Multiplicative( T ) )71 trait Arithmetic {72 }; // Arithmetic70 forall( T | relational( T ) | multiplicative( T ) ) 71 trait arithmetic { 72 }; // arithmetic 73 73 74 74 // Local Variables: // -
libcfa/src/parseconfig.cfa
r24d6572 r62d62db 144 144 in | nl; // ignore remainder of line 145 145 } // for 146 } catch( Open_Failure * ex; ex->istream == &in ) {146 } catch( open_failure * ex; ex->istream == &in ) { 147 147 delete( kv_pairs ); 148 148 throw *ex; … … 203 203 204 204 205 forall(T | Relational( T ))205 forall(T | relational( T )) 206 206 [ bool ] is_nonnegative( & T value ) { 207 207 T zero_val = 0; … … 209 209 } 210 210 211 forall(T | Relational( T ))211 forall(T | relational( T )) 212 212 [ bool ] is_positive( & T value ) { 213 213 T zero_val = 0; … … 215 215 } 216 216 217 forall(T | Relational( T ))217 forall(T | relational( T )) 218 218 [ bool ] is_nonpositive( & T value ) { 219 219 T zero_val = 0; … … 221 221 } 222 222 223 forall(T | Relational( T ))223 forall(T | relational( T )) 224 224 [ bool ] is_negative( & T value ) { 225 225 T zero_val = 0; -
libcfa/src/parseconfig.hfa
r24d6572 r62d62db 107 107 108 108 109 forall(T | Relational( T ))109 forall(T | relational( T )) 110 110 [ bool ] is_nonnegative( & T ); 111 111 112 forall(T | Relational( T ))112 forall(T | relational( T )) 113 113 [ bool ] is_positive( & T ); 114 114 115 forall(T | Relational( T ))115 forall(T | relational( T )) 116 116 [ bool ] is_nonpositive( & T ); 117 117 118 forall(T | Relational( T ))118 forall(T | relational( T )) 119 119 [ bool ] is_negative( & T ); 120 120 -
libcfa/src/rational.cfa
r24d6572 r62d62db 10 10 // Created On : Wed Apr 6 17:54:28 2016 11 11 // Last Modified By : Peter A. Buhr 12 // Last Modified On : Thu Aug 25 18:09:58 202213 // Update Count : 19 412 // Last Modified On : Mon Jun 5 22:49:06 2023 13 // Update Count : 196 14 14 // 15 15 … … 20 20 #pragma GCC visibility push(default) 21 21 22 forall( T | Arithmetic( T ) ) {22 forall( T | arithmetic( T ) ) { 23 23 // helper routines 24 24 … … 39 39 abort | "Invalid rational number construction: denominator cannot be equal to 0."; 40 40 } // exit 41 if ( d < (T){0} ) { d = -d; n = -n; } 41 if ( d < (T){0} ) { d = -d; n = -n; } // move sign to numerator 42 42 return gcd( abs( n ), d ); // simplify 43 } // Rationalnumber::simplify43 } // simplify 44 44 45 45 // constructors 46 46 47 void ?{}( Rational(T) & r, zero_t ) {47 void ?{}( rational(T) & r, zero_t ) { 48 48 r{ (T){0}, (T){1} }; 49 49 } // rational 50 50 51 void ?{}( Rational(T) & r, one_t ) {51 void ?{}( rational(T) & r, one_t ) { 52 52 r{ (T){1}, (T){1} }; 53 53 } // rational 54 54 55 void ?{}( Rational(T) & r ) {55 void ?{}( rational(T) & r ) { 56 56 r{ (T){0}, (T){1} }; 57 57 } // rational 58 58 59 void ?{}( Rational(T) & r, T n ) {59 void ?{}( rational(T) & r, T n ) { 60 60 r{ n, (T){1} }; 61 61 } // rational 62 62 63 void ?{}( Rational(T) & r, T n, T d ) {64 T t = simplify( n, d ); // simplify63 void ?{}( rational(T) & r, T n, T d ) { 64 T t = simplify( n, d ); // simplify 65 65 r.[numerator, denominator] = [n / t, d / t]; 66 66 } // rational … … 68 68 // getter for numerator/denominator 69 69 70 T numerator( Rational(T) r ) {70 T numerator( rational(T) r ) { 71 71 return r.numerator; 72 72 } // numerator 73 73 74 T denominator( Rational(T) r ) {74 T denominator( rational(T) r ) { 75 75 return r.denominator; 76 76 } // denominator 77 77 78 [ T, T ] ?=?( & [ T, T ] dest, Rational(T) src ) {78 [ T, T ] ?=?( & [ T, T ] dest, rational(T) src ) { 79 79 return dest = src.[ numerator, denominator ]; 80 80 } // ?=? … … 82 82 // setter for numerator/denominator 83 83 84 T numerator( Rational(T) r, T n ) {84 T numerator( rational(T) r, T n ) { 85 85 T prev = r.numerator; 86 T t = gcd( abs( n ), r.denominator ); 86 T t = gcd( abs( n ), r.denominator ); // simplify 87 87 r.[numerator, denominator] = [n / t, r.denominator / t]; 88 88 return prev; 89 89 } // numerator 90 90 91 T denominator( Rational(T) r, T d ) {91 T denominator( rational(T) r, T d ) { 92 92 T prev = r.denominator; 93 T t = simplify( r.numerator, d ); // simplify93 T t = simplify( r.numerator, d ); // simplify 94 94 r.[numerator, denominator] = [r.numerator / t, d / t]; 95 95 return prev; … … 98 98 // comparison 99 99 100 int ?==?( Rational(T) l, Rational(T) r ) {100 int ?==?( rational(T) l, rational(T) r ) { 101 101 return l.numerator * r.denominator == l.denominator * r.numerator; 102 102 } // ?==? 103 103 104 int ?!=?( Rational(T) l, Rational(T) r ) {104 int ?!=?( rational(T) l, rational(T) r ) { 105 105 return ! ( l == r ); 106 106 } // ?!=? 107 107 108 int ?!=?( Rational(T) l, zero_t ) {109 return ! ( l == ( Rational(T)){ 0 } );108 int ?!=?( rational(T) l, zero_t ) { 109 return ! ( l == (rational(T)){ 0 } ); 110 110 } // ?!=? 111 111 112 int ?<?( Rational(T) l, Rational(T) r ) {112 int ?<?( rational(T) l, rational(T) r ) { 113 113 return l.numerator * r.denominator < l.denominator * r.numerator; 114 114 } // ?<? 115 115 116 int ?<=?( Rational(T) l, Rational(T) r ) {116 int ?<=?( rational(T) l, rational(T) r ) { 117 117 return l.numerator * r.denominator <= l.denominator * r.numerator; 118 118 } // ?<=? 119 119 120 int ?>?( Rational(T) l, Rational(T) r ) {120 int ?>?( rational(T) l, rational(T) r ) { 121 121 return ! ( l <= r ); 122 122 } // ?>? 123 123 124 int ?>=?( Rational(T) l, Rational(T) r ) {124 int ?>=?( rational(T) l, rational(T) r ) { 125 125 return ! ( l < r ); 126 126 } // ?>=? … … 128 128 // arithmetic 129 129 130 Rational(T) +?( Rational(T) r ) {131 return ( Rational(T)){ r.numerator, r.denominator };130 rational(T) +?( rational(T) r ) { 131 return (rational(T)){ r.numerator, r.denominator }; 132 132 } // +? 133 133 134 Rational(T) -?( Rational(T) r ) {135 return ( Rational(T)){ -r.numerator, r.denominator };134 rational(T) -?( rational(T) r ) { 135 return (rational(T)){ -r.numerator, r.denominator }; 136 136 } // -? 137 137 138 Rational(T) ?+?( Rational(T) l, Rational(T) r ) {138 rational(T) ?+?( rational(T) l, rational(T) r ) { 139 139 if ( l.denominator == r.denominator ) { // special case 140 return ( Rational(T)){ l.numerator + r.numerator, l.denominator };140 return (rational(T)){ l.numerator + r.numerator, l.denominator }; 141 141 } else { 142 return ( Rational(T)){ l.numerator * r.denominator + l.denominator * r.numerator, l.denominator * r.denominator };142 return (rational(T)){ l.numerator * r.denominator + l.denominator * r.numerator, l.denominator * r.denominator }; 143 143 } // if 144 144 } // ?+? 145 145 146 Rational(T) ?+=?( Rational(T) & l, Rational(T) r ) {146 rational(T) ?+=?( rational(T) & l, rational(T) r ) { 147 147 l = l + r; 148 148 return l; 149 149 } // ?+? 150 150 151 Rational(T) ?+=?( Rational(T) & l, one_t ) {152 l = l + ( Rational(T)){ 1 };151 rational(T) ?+=?( rational(T) & l, one_t ) { 152 l = l + (rational(T)){ 1 }; 153 153 return l; 154 154 } // ?+? 155 155 156 Rational(T) ?-?( Rational(T) l, Rational(T) r ) {156 rational(T) ?-?( rational(T) l, rational(T) r ) { 157 157 if ( l.denominator == r.denominator ) { // special case 158 return ( Rational(T)){ l.numerator - r.numerator, l.denominator };158 return (rational(T)){ l.numerator - r.numerator, l.denominator }; 159 159 } else { 160 return ( Rational(T)){ l.numerator * r.denominator - l.denominator * r.numerator, l.denominator * r.denominator };160 return (rational(T)){ l.numerator * r.denominator - l.denominator * r.numerator, l.denominator * r.denominator }; 161 161 } // if 162 162 } // ?-? 163 163 164 Rational(T) ?-=?( Rational(T) & l, Rational(T) r ) {164 rational(T) ?-=?( rational(T) & l, rational(T) r ) { 165 165 l = l - r; 166 166 return l; 167 167 } // ?-? 168 168 169 Rational(T) ?-=?( Rational(T) & l, one_t ) {170 l = l - ( Rational(T)){ 1 };169 rational(T) ?-=?( rational(T) & l, one_t ) { 170 l = l - (rational(T)){ 1 }; 171 171 return l; 172 172 } // ?-? 173 173 174 Rational(T) ?*?( Rational(T) l, Rational(T) r ) {175 return ( Rational(T)){ l.numerator * r.numerator, l.denominator * r.denominator };174 rational(T) ?*?( rational(T) l, rational(T) r ) { 175 return (rational(T)){ l.numerator * r.numerator, l.denominator * r.denominator }; 176 176 } // ?*? 177 177 178 Rational(T) ?*=?( Rational(T) & l, Rational(T) r ) {178 rational(T) ?*=?( rational(T) & l, rational(T) r ) { 179 179 return l = l * r; 180 180 } // ?*? 181 181 182 Rational(T) ?/?( Rational(T) l, Rational(T) r ) {182 rational(T) ?/?( rational(T) l, rational(T) r ) { 183 183 if ( r.numerator < (T){0} ) { 184 184 r.[numerator, denominator] = [-r.numerator, -r.denominator]; 185 185 } // if 186 return ( Rational(T)){ l.numerator * r.denominator, l.denominator * r.numerator };186 return (rational(T)){ l.numerator * r.denominator, l.denominator * r.numerator }; 187 187 } // ?/? 188 188 189 Rational(T) ?/=?( Rational(T) & l, Rational(T) r ) {189 rational(T) ?/=?( rational(T) & l, rational(T) r ) { 190 190 return l = l / r; 191 191 } // ?/? … … 194 194 195 195 forall( istype & | istream( istype ) | { istype & ?|?( istype &, T & ); } ) 196 istype & ?|?( istype & is, Rational(T) & r ) {196 istype & ?|?( istype & is, rational(T) & r ) { 197 197 is | r.numerator | r.denominator; 198 198 T t = simplify( r.numerator, r.denominator ); … … 203 203 204 204 forall( ostype & | ostream( ostype ) | { ostype & ?|?( ostype &, T ); } ) { 205 ostype & ?|?( ostype & os, Rational(T) r ) {205 ostype & ?|?( ostype & os, rational(T) r ) { 206 206 return os | r.numerator | '/' | r.denominator; 207 207 } // ?|? 208 208 209 void ?|?( ostype & os, Rational(T) r ) {209 void ?|?( ostype & os, rational(T) r ) { 210 210 (ostype &)(os | r); ends( os ); 211 211 } // ?|? … … 213 213 } // distribution 214 214 215 forall( T | Arithmetic( T ) | { T ?\?( T, unsigned long ); } ) {216 Rational(T) ?\?( Rational(T) x, long int y ) {215 forall( T | arithmetic( T ) | { T ?\?( T, unsigned long ); } ) { 216 rational(T) ?\?( rational(T) x, long int y ) { 217 217 if ( y < 0 ) { 218 return ( Rational(T)){ x.denominator \ -y, x.numerator \ -y };218 return (rational(T)){ x.denominator \ -y, x.numerator \ -y }; 219 219 } else { 220 return ( Rational(T)){ x.numerator \ y, x.denominator \ y };220 return (rational(T)){ x.numerator \ y, x.denominator \ y }; 221 221 } // if 222 222 } // ?\? 223 223 224 Rational(T) ?\=?( Rational(T) & x, long int y ) {224 rational(T) ?\=?( rational(T) & x, long int y ) { 225 225 return x = x \ y; 226 226 } // ?\? … … 229 229 // conversion 230 230 231 forall( T | Arithmetic( T ) | { double convert( T ); } )232 double widen( Rational(T) r ) {231 forall( T | arithmetic( T ) | { double convert( T ); } ) 232 double widen( rational(T) r ) { 233 233 return convert( r.numerator ) / convert( r.denominator ); 234 234 } // widen 235 235 236 forall( T | Arithmetic( T ) | { double convert( T ); T convert( double ); } )237 Rational(T) narrow( double f, T md ) {236 forall( T | arithmetic( T ) | { double convert( T ); T convert( double ); } ) 237 rational(T) narrow( double f, T md ) { 238 238 // http://www.ics.uci.edu/~eppstein/numth/frap.c 239 if ( md <= (T){1} ) { // maximum fractional digits too small?240 return ( Rational(T)){ convert( f ), (T){1}};// truncate fraction239 if ( md <= (T){1} ) { // maximum fractional digits too small? 240 return (rational(T)){ convert( f ), (T){1}}; // truncate fraction 241 241 } // if 242 242 … … 260 260 if ( f > (double)0x7FFFFFFF ) break; // representation failure 261 261 } // for 262 return ( Rational(T)){ m00, m10 };262 return (rational(T)){ m00, m10 }; 263 263 } // narrow 264 264 -
libcfa/src/rational.hfa
r24d6572 r62d62db 12 12 // Created On : Wed Apr 6 17:56:25 2016 13 13 // Last Modified By : Peter A. Buhr 14 // Last Modified On : Tue Jul 20 17:45:29 202115 // Update Count : 11 814 // Last Modified On : Mon Jun 5 22:49:05 2023 15 // Update Count : 119 16 16 // 17 17 … … 19 19 20 20 #include "iostream.hfa" 21 #include "math.trait.hfa" // Arithmetic21 #include "math.trait.hfa" // arithmetic 22 22 23 23 // implementation 24 24 25 forall( T | Arithmetic( T ) ) {26 struct Rational {25 forall( T | arithmetic( T ) ) { 26 struct rational { 27 27 T numerator, denominator; // invariant: denominator > 0 28 }; // Rational28 }; // rational 29 29 30 30 // constructors 31 31 32 void ?{}( Rational(T) & r );33 void ?{}( Rational(T) & r, zero_t );34 void ?{}( Rational(T) & r, one_t );35 void ?{}( Rational(T) & r, T n );36 void ?{}( Rational(T) & r, T n, T d );32 void ?{}( rational(T) & r ); 33 void ?{}( rational(T) & r, zero_t ); 34 void ?{}( rational(T) & r, one_t ); 35 void ?{}( rational(T) & r, T n ); 36 void ?{}( rational(T) & r, T n, T d ); 37 37 38 38 // numerator/denominator getter 39 39 40 T numerator( Rational(T) r );41 T denominator( Rational(T) r );42 [ T, T ] ?=?( & [ T, T ] dest, Rational(T) src );40 T numerator( rational(T) r ); 41 T denominator( rational(T) r ); 42 [ T, T ] ?=?( & [ T, T ] dest, rational(T) src ); 43 43 44 44 // numerator/denominator setter 45 45 46 T numerator( Rational(T) r, T n );47 T denominator( Rational(T) r, T d );46 T numerator( rational(T) r, T n ); 47 T denominator( rational(T) r, T d ); 48 48 49 49 // comparison 50 50 51 int ?==?( Rational(T) l, Rational(T) r );52 int ?!=?( Rational(T) l, Rational(T) r );53 int ?!=?( Rational(T) l, zero_t ); // => !54 int ?<?( Rational(T) l, Rational(T) r );55 int ?<=?( Rational(T) l, Rational(T) r );56 int ?>?( Rational(T) l, Rational(T) r );57 int ?>=?( Rational(T) l, Rational(T) r );51 int ?==?( rational(T) l, rational(T) r ); 52 int ?!=?( rational(T) l, rational(T) r ); 53 int ?!=?( rational(T) l, zero_t ); // => ! 54 int ?<?( rational(T) l, rational(T) r ); 55 int ?<=?( rational(T) l, rational(T) r ); 56 int ?>?( rational(T) l, rational(T) r ); 57 int ?>=?( rational(T) l, rational(T) r ); 58 58 59 59 // arithmetic 60 60 61 Rational(T) +?( Rational(T) r );62 Rational(T) -?( Rational(T) r );63 Rational(T) ?+?( Rational(T) l, Rational(T) r );64 Rational(T) ?+=?( Rational(T) & l, Rational(T) r );65 Rational(T) ?+=?( Rational(T) & l, one_t ); // => ++?, ?++66 Rational(T) ?-?( Rational(T) l, Rational(T) r );67 Rational(T) ?-=?( Rational(T) & l, Rational(T) r );68 Rational(T) ?-=?( Rational(T) & l, one_t ); // => --?, ?--69 Rational(T) ?*?( Rational(T) l, Rational(T) r );70 Rational(T) ?*=?( Rational(T) & l, Rational(T) r );71 Rational(T) ?/?( Rational(T) l, Rational(T) r );72 Rational(T) ?/=?( Rational(T) & l, Rational(T) r );61 rational(T) +?( rational(T) r ); 62 rational(T) -?( rational(T) r ); 63 rational(T) ?+?( rational(T) l, rational(T) r ); 64 rational(T) ?+=?( rational(T) & l, rational(T) r ); 65 rational(T) ?+=?( rational(T) & l, one_t ); // => ++?, ?++ 66 rational(T) ?-?( rational(T) l, rational(T) r ); 67 rational(T) ?-=?( rational(T) & l, rational(T) r ); 68 rational(T) ?-=?( rational(T) & l, one_t ); // => --?, ?-- 69 rational(T) ?*?( rational(T) l, rational(T) r ); 70 rational(T) ?*=?( rational(T) & l, rational(T) r ); 71 rational(T) ?/?( rational(T) l, rational(T) r ); 72 rational(T) ?/=?( rational(T) & l, rational(T) r ); 73 73 74 74 // I/O 75 75 forall( istype & | istream( istype ) | { istype & ?|?( istype &, T & ); } ) 76 istype & ?|?( istype &, Rational(T) & );76 istype & ?|?( istype &, rational(T) & ); 77 77 78 78 forall( ostype & | ostream( ostype ) | { ostype & ?|?( ostype &, T ); } ) { 79 ostype & ?|?( ostype &, Rational(T) );80 void ?|?( ostype &, Rational(T) );79 ostype & ?|?( ostype &, rational(T) ); 80 void ?|?( ostype &, rational(T) ); 81 81 } // distribution 82 82 } // distribution 83 83 84 forall( T | Arithmetic( T ) | { T ?\?( T, unsigned long ); } ) {85 Rational(T) ?\?( Rational(T) x, long int y );86 Rational(T) ?\=?( Rational(T) & x, long int y );84 forall( T | arithmetic( T ) | { T ?\?( T, unsigned long ); } ) { 85 rational(T) ?\?( rational(T) x, long int y ); 86 rational(T) ?\=?( rational(T) & x, long int y ); 87 87 } // distribution 88 88 89 89 // conversion 90 forall( T | Arithmetic( T ) | { double convert( T ); } )91 double widen( Rational(T) r );92 forall( T | Arithmetic( T ) | { double convert( T ); T convert( double );} )93 Rational(T) narrow( double f, T md );90 forall( T | arithmetic( T ) | { double convert( T ); } ) 91 double widen( rational(T) r ); 92 forall( T | arithmetic( T ) | { double convert( T ); T convert( double );} ) 93 rational(T) narrow( double f, T md ); 94 94 95 95 // Local Variables: // -
src/AST/DeclReplacer.hpp
r24d6572 r62d62db 18 18 #include <unordered_map> 19 19 20 #include "Node.hpp" 20 namespace ast { 21 class DeclWithType; 22 class Expr; 23 class Node; 24 class TypeDecl; 25 } 21 26 22 27 namespace ast { 23 class DeclWithType;24 class TypeDecl;25 class Expr;26 28 27 namespace DeclReplacer { 28 using DeclMap = std::unordered_map< const DeclWithType *, const DeclWithType * >; 29 using TypeMap = std::unordered_map< const TypeDecl *, const TypeDecl * >; 30 using ExprMap = std::unordered_map< const DeclWithType *, const Expr * >; 29 namespace DeclReplacer { 31 30 32 const Node * replace( const Node * node, const DeclMap & declMap, bool debug = false ); 33 const Node * replace( const Node * node, const TypeMap & typeMap, bool debug = false ); 34 const Node * replace( const Node * node, const DeclMap & declMap, const TypeMap & typeMap, bool debug = false ); 35 const Node * replace( const Node * node, const ExprMap & exprMap); 36 } 31 using DeclMap = std::unordered_map< const DeclWithType *, const DeclWithType * >; 32 using TypeMap = std::unordered_map< const TypeDecl *, const TypeDecl * >; 33 using ExprMap = std::unordered_map< const DeclWithType *, const Expr * >; 34 35 const Node * replace( const Node * node, const DeclMap & declMap, bool debug = false ); 36 const Node * replace( const Node * node, const TypeMap & typeMap, bool debug = false ); 37 const Node * replace( const Node * node, const DeclMap & declMap, const TypeMap & typeMap, bool debug = false ); 38 const Node * replace( const Node * node, const ExprMap & exprMap); 39 40 } 41 37 42 } 38 43 -
src/AST/Pass.hpp
r24d6572 r62d62db 414 414 }; 415 415 416 /// Use when the templated visitor should update the symbol table 416 /// Use when the templated visitor should update the symbol table, 417 /// that is, when your pass core needs to query the symbol table. 418 /// Expected setups: 419 /// - For master passes that kick off at the compilation unit 420 /// - before resolver: extend WithSymbolTableX<IgnoreErrors> 421 /// - after resolver: extend WithSymbolTable and use defaults 422 /// - (FYI, for completeness, the resolver's main pass uses ValidateOnAdd when it kicks off) 423 /// - For helper passes that kick off at arbitrary points in the AST: 424 /// - take an existing symbol table as a parameter, extend WithSymbolTable, 425 /// and construct with WithSymbolTable(const SymbolTable &) 417 426 struct WithSymbolTable { 418 SymbolTable symtab; 427 WithSymbolTable(const ast::SymbolTable & from) : symtab(from) {} 428 WithSymbolTable(ast::SymbolTable::ErrorDetection errorMode = ast::SymbolTable::ErrorDetection::AssertClean) : symtab(errorMode) {} 429 ast::SymbolTable symtab; 430 }; 431 template <ast::SymbolTable::ErrorDetection errorMode> 432 struct WithSymbolTableX : WithSymbolTable { 433 WithSymbolTableX() : WithSymbolTable(errorMode) {} 419 434 }; 420 435 -
src/AST/Pass.impl.hpp
r24d6572 r62d62db 72 72 template<typename it_t, template <class...> class container_t> 73 73 static inline void take_all( it_t it, container_t<ast::ptr<ast::Decl>> * decls, bool * mutated = nullptr ) { 74 if (empty(decls)) return;74 if ( empty( decls ) ) return; 75 75 76 76 std::transform(decls->begin(), decls->end(), it, [](const ast::Decl * decl) -> auto { … … 78 78 }); 79 79 decls->clear(); 80 if (mutated) *mutated = true;80 if ( mutated ) *mutated = true; 81 81 } 82 82 83 83 template<typename it_t, template <class...> class container_t> 84 84 static inline void take_all( it_t it, container_t<ast::ptr<ast::Stmt>> * stmts, bool * mutated = nullptr ) { 85 if (empty(stmts)) return;85 if ( empty( stmts ) ) return; 86 86 87 87 std::move(stmts->begin(), stmts->end(), it); 88 88 stmts->clear(); 89 if (mutated) *mutated = true;89 if ( mutated ) *mutated = true; 90 90 } 91 91 … … 93 93 /// Check if should be skipped, different for pointers and containers 94 94 template<typename node_t> 95 bool skip( const ast::ptr<node_t> & val ) {95 bool skip( const ast::ptr<node_t> & val ) { 96 96 return !val; 97 97 } … … 110 110 111 111 template<typename node_t> 112 const node_t & get( const node_t & val, long ) {112 const node_t & get( const node_t & val, long ) { 113 113 return val; 114 114 } … … 126 126 } 127 127 } 128 129 template< typename core_t > 130 template< typename node_t > 131 auto ast::Pass< core_t >::call_accept( const node_t * node ) 132 -> typename ast::Pass< core_t >::template generic_call_accept_result<node_t>::type 133 { 134 __pedantic_pass_assert( __visit_children() ); 135 __pedantic_pass_assert( node ); 136 137 static_assert( !std::is_base_of<ast::Expr, node_t>::value, "ERROR"); 138 static_assert( !std::is_base_of<ast::Stmt, node_t>::value, "ERROR"); 139 140 auto nval = node->accept( *this ); 141 __pass::result1< 142 typename std::remove_pointer< decltype( node->accept(*this) ) >::type 143 > res; 144 res.differs = nval != node; 145 res.value = nval; 146 return res; 147 } 148 149 template< typename core_t > 150 __pass::template result1<ast::Expr> ast::Pass< core_t >::call_accept( const ast::Expr * expr ) { 151 __pedantic_pass_assert( __visit_children() ); 152 __pedantic_pass_assert( expr ); 153 154 auto nval = expr->accept( *this ); 155 return { nval != expr, nval }; 156 } 157 158 template< typename core_t > 159 __pass::template result1<ast::Stmt> ast::Pass< core_t >::call_accept( const ast::Stmt * stmt ) { 160 __pedantic_pass_assert( __visit_children() ); 161 __pedantic_pass_assert( stmt ); 162 163 const ast::Stmt * nval = stmt->accept( *this ); 164 return { nval != stmt, nval }; 165 } 166 167 template< typename core_t > 168 __pass::template result1<ast::Expr> ast::Pass< core_t >::call_accept_top( const ast::Expr * expr ) { 169 __pedantic_pass_assert( __visit_children() ); 170 __pedantic_pass_assert( expr ); 171 172 const ast::TypeSubstitution ** typeSubs_ptr = __pass::typeSubs( core, 0 ); 173 if ( typeSubs_ptr && expr->env ) { 174 *typeSubs_ptr = expr->env; 175 } 176 177 auto nval = expr->accept( *this ); 178 return { nval != expr, nval }; 179 } 180 181 template< typename core_t > 182 __pass::template result1<ast::Stmt> ast::Pass< core_t >::call_accept_as_compound( const ast::Stmt * stmt ) { 183 __pedantic_pass_assert( __visit_children() ); 184 __pedantic_pass_assert( stmt ); 185 186 // add a few useful symbols to the scope 187 using __pass::empty; 188 189 // get the stmts/decls that will need to be spliced in 190 auto stmts_before = __pass::stmtsToAddBefore( core, 0 ); 191 auto stmts_after = __pass::stmtsToAddAfter ( core, 0 ); 192 auto decls_before = __pass::declsToAddBefore( core, 0 ); 193 auto decls_after = __pass::declsToAddAfter ( core, 0 ); 194 195 // These may be modified by subnode but most be restored once we exit this statemnet. 196 ValueGuardPtr< const ast::TypeSubstitution * > __old_env ( __pass::typeSubs( core, 0 ) ); 197 ValueGuardPtr< typename std::remove_pointer< decltype(stmts_before) >::type > __old_decls_before( stmts_before ); 198 ValueGuardPtr< typename std::remove_pointer< decltype(stmts_after ) >::type > __old_decls_after ( stmts_after ); 199 ValueGuardPtr< typename std::remove_pointer< decltype(decls_before) >::type > __old_stmts_before( decls_before ); 200 ValueGuardPtr< typename std::remove_pointer< decltype(decls_after ) >::type > __old_stmts_after ( decls_after ); 201 202 // Now is the time to actually visit the node 203 const ast::Stmt * nstmt = stmt->accept( *this ); 204 205 // If the pass doesn't want to add anything then we are done 206 if( empty(stmts_before) && empty(stmts_after) && empty(decls_before) && empty(decls_after) ) { 207 return { nstmt != stmt, nstmt }; 208 } 209 210 // Make sure that it is either adding statements or declartions but not both 211 // this is because otherwise the order would be awkward to predict 212 assert(( empty( stmts_before ) && empty( stmts_after )) 213 || ( empty( decls_before ) && empty( decls_after )) ); 214 215 // Create a new Compound Statement to hold the new decls/stmts 216 ast::CompoundStmt * compound = new ast::CompoundStmt( stmt->location ); 217 218 // Take all the declarations that go before 219 __pass::take_all( std::back_inserter( compound->kids ), decls_before ); 220 __pass::take_all( std::back_inserter( compound->kids ), stmts_before ); 221 222 // Insert the original declaration 223 compound->kids.emplace_back( nstmt ); 224 225 // Insert all the declarations that go before 226 __pass::take_all( std::back_inserter( compound->kids ), decls_after ); 227 __pass::take_all( std::back_inserter( compound->kids ), stmts_after ); 228 229 return {true, compound}; 230 } 231 232 template< typename core_t > 233 template< template <class...> class container_t > 234 __pass::template resultNstmt<container_t> ast::Pass< core_t >::call_accept( const container_t< ptr<Stmt> > & statements ) { 235 __pedantic_pass_assert( __visit_children() ); 236 if( statements.empty() ) return {}; 237 238 // We are going to aggregate errors for all these statements 239 SemanticErrorException errors; 240 241 // add a few useful symbols to the scope 242 using __pass::empty; 243 244 // get the stmts/decls that will need to be spliced in 245 auto stmts_before = __pass::stmtsToAddBefore( core, 0 ); 246 auto stmts_after = __pass::stmtsToAddAfter ( core, 0 ); 247 auto decls_before = __pass::declsToAddBefore( core, 0 ); 248 auto decls_after = __pass::declsToAddAfter ( core, 0 ); 249 250 // These may be modified by subnode but most be restored once we exit this statemnet. 251 ValueGuardPtr< typename std::remove_pointer< decltype(stmts_before) >::type > __old_decls_before( stmts_before ); 252 ValueGuardPtr< typename std::remove_pointer< decltype(stmts_after ) >::type > __old_decls_after ( stmts_after ); 253 ValueGuardPtr< typename std::remove_pointer< decltype(decls_before) >::type > __old_stmts_before( decls_before ); 254 ValueGuardPtr< typename std::remove_pointer< decltype(decls_after ) >::type > __old_stmts_after ( decls_after ); 255 256 // update pass statitistics 257 pass_visitor_stats.depth++; 258 pass_visitor_stats.max->push(pass_visitor_stats.depth); 259 pass_visitor_stats.avg->push(pass_visitor_stats.depth); 260 261 __pass::resultNstmt<container_t> new_kids; 262 for( auto value : enumerate( statements ) ) { 263 try { 264 size_t i = value.idx; 265 const Stmt * stmt = value.val; 266 __pedantic_pass_assert( stmt ); 267 const ast::Stmt * new_stmt = stmt->accept( *this ); 268 assert( new_stmt ); 269 if(new_stmt != stmt ) { new_kids.differs = true; } 270 271 // Make sure that it is either adding statements or declartions but not both 272 // this is because otherwise the order would be awkward to predict 273 assert(( empty( stmts_before ) && empty( stmts_after )) 274 || ( empty( decls_before ) && empty( decls_after )) ); 275 276 // Take all the statements which should have gone after, N/A for first iteration 277 new_kids.take_all( decls_before ); 278 new_kids.take_all( stmts_before ); 279 280 // Now add the statement if there is one 281 if(new_stmt != stmt) { 282 new_kids.values.emplace_back( new_stmt, i, false ); 283 } else { 284 new_kids.values.emplace_back( nullptr, i, true ); 285 } 286 287 // Take all the declarations that go before 288 new_kids.take_all( decls_after ); 289 new_kids.take_all( stmts_after ); 128 } 129 130 template< typename core_t > 131 template< typename node_t > 132 auto ast::Pass< core_t >::call_accept( const node_t * node ) -> 133 typename ast::Pass< core_t >::template generic_call_accept_result<node_t>::type 134 { 135 __pedantic_pass_assert( __visit_children() ); 136 __pedantic_pass_assert( node ); 137 138 static_assert( !std::is_base_of<ast::Expr, node_t>::value, "ERROR" ); 139 static_assert( !std::is_base_of<ast::Stmt, node_t>::value, "ERROR" ); 140 141 auto nval = node->accept( *this ); 142 __pass::result1< 143 typename std::remove_pointer< decltype( node->accept(*this) ) >::type 144 > res; 145 res.differs = nval != node; 146 res.value = nval; 147 return res; 148 } 149 150 template< typename core_t > 151 ast::__pass::template result1<ast::Expr> ast::Pass< core_t >::call_accept( const ast::Expr * expr ) { 152 __pedantic_pass_assert( __visit_children() ); 153 __pedantic_pass_assert( expr ); 154 155 auto nval = expr->accept( *this ); 156 return { nval != expr, nval }; 157 } 158 159 template< typename core_t > 160 ast::__pass::template result1<ast::Stmt> ast::Pass< core_t >::call_accept( const ast::Stmt * stmt ) { 161 __pedantic_pass_assert( __visit_children() ); 162 __pedantic_pass_assert( stmt ); 163 164 const ast::Stmt * nval = stmt->accept( *this ); 165 return { nval != stmt, nval }; 166 } 167 168 template< typename core_t > 169 ast::__pass::template result1<ast::Expr> ast::Pass< core_t >::call_accept_top( const ast::Expr * expr ) { 170 __pedantic_pass_assert( __visit_children() ); 171 __pedantic_pass_assert( expr ); 172 173 const ast::TypeSubstitution ** typeSubs_ptr = __pass::typeSubs( core, 0 ); 174 if ( typeSubs_ptr && expr->env ) { 175 *typeSubs_ptr = expr->env; 176 } 177 178 auto nval = expr->accept( *this ); 179 return { nval != expr, nval }; 180 } 181 182 template< typename core_t > 183 ast::__pass::template result1<ast::Stmt> ast::Pass< core_t >::call_accept_as_compound( const ast::Stmt * stmt ) { 184 __pedantic_pass_assert( __visit_children() ); 185 __pedantic_pass_assert( stmt ); 186 187 // add a few useful symbols to the scope 188 using __pass::empty; 189 190 // get the stmts/decls that will need to be spliced in 191 auto stmts_before = __pass::stmtsToAddBefore( core, 0 ); 192 auto stmts_after = __pass::stmtsToAddAfter ( core, 0 ); 193 auto decls_before = __pass::declsToAddBefore( core, 0 ); 194 auto decls_after = __pass::declsToAddAfter ( core, 0 ); 195 196 // These may be modified by subnode but most be restored once we exit this statemnet. 197 ValueGuardPtr< const ast::TypeSubstitution * > __old_env ( __pass::typeSubs( core, 0 ) ); 198 ValueGuardPtr< typename std::remove_pointer< decltype(stmts_before) >::type > __old_decls_before( stmts_before ); 199 ValueGuardPtr< typename std::remove_pointer< decltype(stmts_after ) >::type > __old_decls_after ( stmts_after ); 200 ValueGuardPtr< typename std::remove_pointer< decltype(decls_before) >::type > __old_stmts_before( decls_before ); 201 ValueGuardPtr< typename std::remove_pointer< decltype(decls_after ) >::type > __old_stmts_after ( decls_after ); 202 203 // Now is the time to actually visit the node 204 const ast::Stmt * nstmt = stmt->accept( *this ); 205 206 // If the pass doesn't want to add anything then we are done 207 if ( empty(stmts_before) && empty(stmts_after) && empty(decls_before) && empty(decls_after) ) { 208 return { nstmt != stmt, nstmt }; 209 } 210 211 // Make sure that it is either adding statements or declartions but not both 212 // this is because otherwise the order would be awkward to predict 213 assert(( empty( stmts_before ) && empty( stmts_after )) 214 || ( empty( decls_before ) && empty( decls_after )) ); 215 216 // Create a new Compound Statement to hold the new decls/stmts 217 ast::CompoundStmt * compound = new ast::CompoundStmt( stmt->location ); 218 219 // Take all the declarations that go before 220 __pass::take_all( std::back_inserter( compound->kids ), decls_before ); 221 __pass::take_all( std::back_inserter( compound->kids ), stmts_before ); 222 223 // Insert the original declaration 224 compound->kids.emplace_back( nstmt ); 225 226 // Insert all the declarations that go before 227 __pass::take_all( std::back_inserter( compound->kids ), decls_after ); 228 __pass::take_all( std::back_inserter( compound->kids ), stmts_after ); 229 230 return { true, compound }; 231 } 232 233 template< typename core_t > 234 template< template <class...> class container_t > 235 ast::__pass::template resultNstmt<container_t> ast::Pass< core_t >::call_accept( const container_t< ptr<Stmt> > & statements ) { 236 __pedantic_pass_assert( __visit_children() ); 237 if ( statements.empty() ) return {}; 238 239 // We are going to aggregate errors for all these statements 240 SemanticErrorException errors; 241 242 // add a few useful symbols to the scope 243 using __pass::empty; 244 245 // get the stmts/decls that will need to be spliced in 246 auto stmts_before = __pass::stmtsToAddBefore( core, 0 ); 247 auto stmts_after = __pass::stmtsToAddAfter ( core, 0 ); 248 auto decls_before = __pass::declsToAddBefore( core, 0 ); 249 auto decls_after = __pass::declsToAddAfter ( core, 0 ); 250 251 // These may be modified by subnode but most be restored once we exit this statemnet. 252 ValueGuardPtr< typename std::remove_pointer< decltype(stmts_before) >::type > __old_decls_before( stmts_before ); 253 ValueGuardPtr< typename std::remove_pointer< decltype(stmts_after ) >::type > __old_decls_after ( stmts_after ); 254 ValueGuardPtr< typename std::remove_pointer< decltype(decls_before) >::type > __old_stmts_before( decls_before ); 255 ValueGuardPtr< typename std::remove_pointer< decltype(decls_after ) >::type > __old_stmts_after ( decls_after ); 256 257 // update pass statitistics 258 pass_visitor_stats.depth++; 259 pass_visitor_stats.max->push(pass_visitor_stats.depth); 260 pass_visitor_stats.avg->push(pass_visitor_stats.depth); 261 262 __pass::resultNstmt<container_t> new_kids; 263 for ( auto value : enumerate( statements ) ) { 264 try { 265 size_t i = value.idx; 266 const Stmt * stmt = value.val; 267 __pedantic_pass_assert( stmt ); 268 const ast::Stmt * new_stmt = stmt->accept( *this ); 269 assert( new_stmt ); 270 if ( new_stmt != stmt ) { new_kids.differs = true; } 271 272 // Make sure that it is either adding statements or declartions but not both 273 // this is because otherwise the order would be awkward to predict 274 assert(( empty( stmts_before ) && empty( stmts_after )) 275 || ( empty( decls_before ) && empty( decls_after )) ); 276 277 // Take all the statements which should have gone after, N/A for first iteration 278 new_kids.take_all( decls_before ); 279 new_kids.take_all( stmts_before ); 280 281 // Now add the statement if there is one 282 if ( new_stmt != stmt ) { 283 new_kids.values.emplace_back( new_stmt, i, false ); 284 } else { 285 new_kids.values.emplace_back( nullptr, i, true ); 290 286 } 291 catch ( SemanticErrorException &e ) { 292 errors.append( e ); 287 288 // Take all the declarations that go before 289 new_kids.take_all( decls_after ); 290 new_kids.take_all( stmts_after ); 291 } catch ( SemanticErrorException &e ) { 292 errors.append( e ); 293 } 294 } 295 pass_visitor_stats.depth--; 296 if ( !errors.isEmpty() ) { throw errors; } 297 298 return new_kids; 299 } 300 301 template< typename core_t > 302 template< template <class...> class container_t, typename node_t > 303 ast::__pass::template resultN<container_t, node_t> ast::Pass< core_t >::call_accept( const container_t< ast::ptr<node_t> > & container ) { 304 __pedantic_pass_assert( __visit_children() ); 305 if ( container.empty() ) return {}; 306 SemanticErrorException errors; 307 308 pass_visitor_stats.depth++; 309 pass_visitor_stats.max->push(pass_visitor_stats.depth); 310 pass_visitor_stats.avg->push(pass_visitor_stats.depth); 311 312 bool mutated = false; 313 container_t<ptr<node_t>> new_kids; 314 for ( const node_t * node : container ) { 315 try { 316 __pedantic_pass_assert( node ); 317 const node_t * new_stmt = strict_dynamic_cast< const node_t * >( node->accept( *this ) ); 318 if ( new_stmt != node ) { 319 mutated = true; 320 new_kids.emplace_back( new_stmt ); 321 } else { 322 new_kids.emplace_back( nullptr ); 293 323 } 294 } 295 pass_visitor_stats.depth--; 296 if ( !errors.isEmpty() ) { throw errors; } 297 298 return new_kids; 299 } 300 301 template< typename core_t > 302 template< template <class...> class container_t, typename node_t > 303 __pass::template resultN<container_t, node_t> ast::Pass< core_t >::call_accept( const container_t< ast::ptr<node_t> > & container ) { 304 __pedantic_pass_assert( __visit_children() ); 305 if( container.empty() ) return {}; 306 SemanticErrorException errors; 307 308 pass_visitor_stats.depth++; 309 pass_visitor_stats.max->push(pass_visitor_stats.depth); 310 pass_visitor_stats.avg->push(pass_visitor_stats.depth); 311 312 bool mutated = false; 313 container_t<ptr<node_t>> new_kids; 314 for ( const node_t * node : container ) { 315 try { 316 __pedantic_pass_assert( node ); 317 const node_t * new_stmt = strict_dynamic_cast< const node_t * >( node->accept( *this ) ); 318 if(new_stmt != node ) { 319 mutated = true; 320 new_kids.emplace_back( new_stmt ); 321 } else { 322 new_kids.emplace_back( nullptr ); 323 } 324 325 } 326 catch( SemanticErrorException &e ) { 327 errors.append( e ); 328 } 329 } 330 331 __pedantic_pass_assert( new_kids.size() == container.size() ); 332 pass_visitor_stats.depth--; 333 if ( ! errors.isEmpty() ) { throw errors; } 334 335 return ast::__pass::resultN<container_t, node_t>{ mutated, new_kids }; 336 } 337 338 template< typename core_t > 339 template<typename node_t, typename super_t, typename field_t> 340 void ast::Pass< core_t >::maybe_accept( 341 const node_t * & parent, 342 field_t super_t::*field 343 ) { 344 static_assert( std::is_base_of<super_t, node_t>::value, "Error deducing member object" ); 345 346 if(__pass::skip(parent->*field)) return; 347 const auto & old_val = __pass::get(parent->*field, 0); 348 349 static_assert( !std::is_same<const ast::Node * &, decltype(old_val)>::value, "ERROR"); 350 351 auto new_val = call_accept( old_val ); 352 353 static_assert( !std::is_same<const ast::Node *, decltype(new_val)>::value /* || std::is_same<int, decltype(old_val)>::value */, "ERROR"); 354 355 if( new_val.differs ) { 356 auto new_parent = __pass::mutate<core_t>(parent); 357 new_val.apply(new_parent, field); 358 parent = new_parent; 359 } 360 } 361 362 template< typename core_t > 363 template<typename node_t, typename super_t, typename field_t> 364 void ast::Pass< core_t >::maybe_accept_top( 365 const node_t * & parent, 366 field_t super_t::*field 367 ) { 368 static_assert( std::is_base_of<super_t, node_t>::value, "Error deducing member object" ); 369 370 if(__pass::skip(parent->*field)) return; 371 const auto & old_val = __pass::get(parent->*field, 0); 372 373 static_assert( !std::is_same<const ast::Node * &, decltype(old_val)>::value, "ERROR"); 374 375 auto new_val = call_accept_top( old_val ); 376 377 static_assert( !std::is_same<const ast::Node *, decltype(new_val)>::value /* || std::is_same<int, decltype(old_val)>::value */, "ERROR"); 378 379 if( new_val.differs ) { 380 auto new_parent = __pass::mutate<core_t>(parent); 381 new_val.apply(new_parent, field); 382 parent = new_parent; 383 } 384 } 385 386 template< typename core_t > 387 template<typename node_t, typename super_t, typename field_t> 388 void ast::Pass< core_t >::maybe_accept_as_compound( 389 const node_t * & parent, 390 field_t super_t::*child 391 ) { 392 static_assert( std::is_base_of<super_t, node_t>::value, "Error deducing member object" ); 393 394 if(__pass::skip(parent->*child)) return; 395 const auto & old_val = __pass::get(parent->*child, 0); 396 397 static_assert( !std::is_same<const ast::Node * &, decltype(old_val)>::value, "ERROR"); 398 399 auto new_val = call_accept_as_compound( old_val ); 400 401 static_assert( !std::is_same<const ast::Node *, decltype(new_val)>::value || std::is_same<int, decltype(old_val)>::value, "ERROR"); 402 403 if( new_val.differs ) { 404 auto new_parent = __pass::mutate<core_t>(parent); 405 new_val.apply( new_parent, child ); 406 parent = new_parent; 407 } 408 } 409 324 } catch ( SemanticErrorException &e ) { 325 errors.append( e ); 326 } 327 } 328 329 __pedantic_pass_assert( new_kids.size() == container.size() ); 330 pass_visitor_stats.depth--; 331 if ( !errors.isEmpty() ) { throw errors; } 332 333 return ast::__pass::resultN<container_t, node_t>{ mutated, new_kids }; 334 } 335 336 template< typename core_t > 337 template<typename node_t, typename super_t, typename field_t> 338 void ast::Pass< core_t >::maybe_accept( 339 const node_t * & parent, 340 field_t super_t::*field 341 ) { 342 static_assert( std::is_base_of<super_t, node_t>::value, "Error deducing member object" ); 343 344 if ( __pass::skip( parent->*field ) ) return; 345 const auto & old_val = __pass::get(parent->*field, 0); 346 347 static_assert( !std::is_same<const ast::Node * &, decltype(old_val)>::value, "ERROR" ); 348 349 auto new_val = call_accept( old_val ); 350 351 static_assert( !std::is_same<const ast::Node *, decltype(new_val)>::value /* || std::is_same<int, decltype(old_val)>::value */, "ERROR" ); 352 353 if ( new_val.differs ) { 354 auto new_parent = __pass::mutate<core_t>(parent); 355 new_val.apply(new_parent, field); 356 parent = new_parent; 357 } 358 } 359 360 template< typename core_t > 361 template<typename node_t, typename super_t, typename field_t> 362 void ast::Pass< core_t >::maybe_accept_top( 363 const node_t * & parent, 364 field_t super_t::*field 365 ) { 366 static_assert( std::is_base_of<super_t, node_t>::value, "Error deducing member object" ); 367 368 if ( __pass::skip( parent->*field ) ) return; 369 const auto & old_val = __pass::get(parent->*field, 0); 370 371 static_assert( !std::is_same<const ast::Node * &, decltype(old_val)>::value, "ERROR" ); 372 373 auto new_val = call_accept_top( old_val ); 374 375 static_assert( !std::is_same<const ast::Node *, decltype(new_val)>::value /* || std::is_same<int, decltype(old_val)>::value */, "ERROR" ); 376 377 if ( new_val.differs ) { 378 auto new_parent = __pass::mutate<core_t>(parent); 379 new_val.apply(new_parent, field); 380 parent = new_parent; 381 } 382 } 383 384 template< typename core_t > 385 template<typename node_t, typename super_t, typename field_t> 386 void ast::Pass< core_t >::maybe_accept_as_compound( 387 const node_t * & parent, 388 field_t super_t::*child 389 ) { 390 static_assert( std::is_base_of<super_t, node_t>::value, "Error deducing member object" ); 391 392 if ( __pass::skip( parent->*child ) ) return; 393 const auto & old_val = __pass::get(parent->*child, 0); 394 395 static_assert( !std::is_same<const ast::Node * &, decltype(old_val)>::value, "ERROR" ); 396 397 auto new_val = call_accept_as_compound( old_val ); 398 399 static_assert( !std::is_same<const ast::Node *, decltype(new_val)>::value || std::is_same<int, decltype(old_val)>::value, "ERROR" ); 400 401 if ( new_val.differs ) { 402 auto new_parent = __pass::mutate<core_t>(parent); 403 new_val.apply( new_parent, child ); 404 parent = new_parent; 405 } 410 406 } 411 407 … … 761 757 762 758 if ( __visit_children() ) { 763 // Do not enter (or leave) a new scope if atFunctionTop. Remember to save the result. 764 auto guard1 = makeFuncGuard( [this, enterScope = !this->atFunctionTop]() { 765 if ( enterScope ) { 766 __pass::symtab::enter(core, 0); 767 } 768 }, [this, leaveScope = !this->atFunctionTop]() { 769 if ( leaveScope ) { 770 __pass::symtab::leave(core, 0); 771 } 772 }); 773 ValueGuard< bool > guard2( atFunctionTop ); 774 atFunctionTop = false; 775 guard_scope guard3 { *this }; 776 maybe_accept( node, &CompoundStmt::kids ); 759 // Do not enter (or leave) a new symbol table scope if atFunctionTop. 760 // But always enter (and leave) a new general scope. 761 if ( atFunctionTop ) { 762 ValueGuard< bool > guard1( atFunctionTop ); 763 atFunctionTop = false; 764 guard_scope guard2( *this ); 765 maybe_accept( node, &CompoundStmt::kids ); 766 } else { 767 guard_symtab guard1( *this ); 768 guard_scope guard2( *this ); 769 maybe_accept( node, &CompoundStmt::kids ); 770 } 777 771 } 778 772 -
src/AST/SymbolTable.cpp
r24d6572 r62d62db 91 91 } 92 92 93 SymbolTable::SymbolTable( )93 SymbolTable::SymbolTable( ErrorDetection errorMode ) 94 94 : idTable(), typeTable(), structTable(), enumTable(), unionTable(), traitTable(), 95 prevScope(), scope( 0 ), repScope( 0 ) { ++*stats().count; }95 prevScope(), scope( 0 ), repScope( 0 ), errorMode(errorMode) { ++*stats().count; } 96 96 97 97 SymbolTable::~SymbolTable() { stats().size->push( idTable ? idTable->size() : 0 ); } 98 99 void SymbolTable::OnFindError( CodeLocation location, std::string error ) const { 100 assertf( errorMode != AssertClean, "Name collision/redefinition, found during a compilation phase where none should be possible. Detail: %s", error.c_str() ); 101 if (errorMode == ValidateOnAdd) { 102 SemanticError(location, error); 103 } 104 assertf( errorMode == IgnoreErrors, "Unrecognized symbol-table error mode %d", errorMode ); 105 } 98 106 99 107 void SymbolTable::enterScope() { … … 274 282 } 275 283 276 namespace { 277 /// true if redeclaration conflict between two types 278 bool addedTypeConflicts( const NamedTypeDecl * existing, const NamedTypeDecl * added ) { 279 if ( existing->base == nullptr ) { 280 return false; 281 } else if ( added->base == nullptr ) { 282 return true; 283 } else { 284 // typedef redeclarations are errors only if types are different 285 if ( ! ResolvExpr::typesCompatible( existing->base, added->base ) ) { 286 SemanticError( added->location, "redeclaration of " + added->name ); 287 } 288 } 289 // does not need to be added to the table if both existing and added have a base that are 290 // the same 284 bool SymbolTable::addedTypeConflicts( 285 const NamedTypeDecl * existing, const NamedTypeDecl * added ) const { 286 if ( existing->base == nullptr ) { 287 return false; 288 } else if ( added->base == nullptr ) { 291 289 return true; 292 } 293 294 /// true if redeclaration conflict between two aggregate declarations 295 bool addedDeclConflicts( const AggregateDecl * existing, const AggregateDecl * added ) { 296 if ( ! existing->body ) { 297 return false; 298 } else if ( added->body ) { 299 SemanticError( added, "redeclaration of " ); 300 } 301 return true; 302 } 290 } else { 291 // typedef redeclarations are errors only if types are different 292 if ( ! ResolvExpr::typesCompatible( existing->base, added->base ) ) { 293 OnFindError( added->location, "redeclaration of " + added->name ); 294 } 295 } 296 // does not need to be added to the table if both existing and added have a base that are 297 // the same 298 return true; 299 } 300 301 bool SymbolTable::addedDeclConflicts( 302 const AggregateDecl * existing, const AggregateDecl * added ) const { 303 if ( ! existing->body ) { 304 return false; 305 } else if ( added->body ) { 306 OnFindError( added, "redeclaration of " ); 307 } 308 return true; 303 309 } 304 310 … … 653 659 if ( deleter && ! existing.deleter ) { 654 660 if ( handleConflicts.mode == OnConflict::Error ) { 655 SemanticError( added, "deletion of defined identifier " );661 OnFindError( added, "deletion of defined identifier " ); 656 662 } 657 663 return true; 658 664 } else if ( ! deleter && existing.deleter ) { 659 665 if ( handleConflicts.mode == OnConflict::Error ) { 660 SemanticError( added, "definition of deleted identifier " );666 OnFindError( added, "definition of deleted identifier " ); 661 667 } 662 668 return true; … … 666 672 if ( isDefinition( added ) && isDefinition( existing.id ) ) { 667 673 if ( handleConflicts.mode == OnConflict::Error ) { 668 SemanticError( added,674 OnFindError( added, 669 675 isFunction( added ) ? 670 676 "duplicate function definition for " : … … 675 681 } else { 676 682 if ( handleConflicts.mode == OnConflict::Error ) { 677 SemanticError( added, "duplicate definition for " );683 OnFindError( added, "duplicate definition for " ); 678 684 } 679 685 return true; … … 727 733 // Check that a Cforall declaration doesn't override any C declaration 728 734 if ( hasCompatibleCDecl( name, mangleName ) ) { 729 SemanticError( decl, "Cforall declaration hides C function " );735 OnFindError( decl, "Cforall declaration hides C function " ); 730 736 } 731 737 } else { … … 733 739 // type-compatibility, which it may not be. 734 740 if ( hasIncompatibleCDecl( name, mangleName ) ) { 735 SemanticError( decl, "conflicting overload of C function " );741 OnFindError( decl, "conflicting overload of C function " ); 736 742 } 737 743 } -
src/AST/SymbolTable.hpp
r24d6572 r62d62db 93 93 94 94 public: 95 SymbolTable(); 95 96 /// Mode to control when (during which pass) user-caused name-declaration errors get reported. 97 /// The default setting `AssertClean` supports, "I expect all user-caused errors to have been 98 /// reported by now," or, "I wouldn't know what to do with an error; are there even any here?" 99 enum ErrorDetection { 100 AssertClean, ///< invalid user decls => assert fails during addFoo (default) 101 ValidateOnAdd, ///< invalid user decls => calls SemanticError during addFoo 102 IgnoreErrors ///< acts as if unspecified decls were removed, forcing validity 103 }; 104 105 explicit SymbolTable( 106 ErrorDetection ///< mode for the lifetime of the symbol table (whole pass) 107 ); 108 SymbolTable() : SymbolTable(AssertClean) {} 96 109 ~SymbolTable(); 110 111 ErrorDetection getErrorMode() const { 112 return errorMode; 113 } 97 114 98 115 // when using an indexer manually (e.g., within a mutator traversal), it is necessary to … … 158 175 159 176 private: 177 void OnFindError( CodeLocation location, std::string error ) const; 178 179 template< typename T > 180 void OnFindError( const T * obj, const std::string & error ) const { 181 OnFindError( obj->location, toString( error, obj ) ); 182 } 183 184 template< typename T > 185 void OnFindError( CodeLocation location, const T * obj, const std::string & error ) const { 186 OnFindError( location, toString( error, obj ) ); 187 } 188 160 189 /// Ensures that a proper backtracking scope exists before a mutation 161 190 void lazyInitScope(); … … 168 197 bool removeSpecialOverrides( IdData & decl, MangleTable::Ptr & mangleTable ); 169 198 170 /// Options for handling identifier conflicts 199 /// Error detection mode given at construction (pass-specific). 200 /// Logically const, except that the symbol table's push-pop is achieved by autogenerated 201 /// assignment onto self. The feield is left motuable to keep this code-gen simple. 202 /// Conceptual constness is preserved by all SymbolTable in a stack sharing the same mode. 203 ErrorDetection errorMode; 204 205 /// Options for handling identifier conflicts. 206 /// Varies according to AST location during traversal: captures semantics of the construct 207 /// being visited as "would shadow" vs "must not collide." 208 /// At a given AST location, is the same for every pass. 171 209 struct OnConflict { 172 210 enum { 173 Error, ///< Throw a semantic error211 Error, ///< Follow the current pass's ErrorDetection mode (may throw a semantic error) 174 212 Delete ///< Delete the earlier version with the delete statement 175 213 } mode; … … 191 229 const Decl * deleter ); 192 230 231 /// true if redeclaration conflict between two types 232 bool addedTypeConflicts( const NamedTypeDecl * existing, const NamedTypeDecl * added ) const; 233 234 /// true if redeclaration conflict between two aggregate declarations 235 bool addedDeclConflicts( const AggregateDecl * existing, const AggregateDecl * added ) const; 236 193 237 /// common code for addId, addDeletedId, etc. 194 238 void addIdCommon( … … 213 257 } 214 258 259 215 260 // Local Variables: // 216 261 // tab-width: 4 // -
src/AST/Util.cpp
r24d6572 r62d62db 83 83 } 84 84 85 /// Check that the MemberExpr has an aggregate type and matching member. 86 void memberMatchesAggregate( const MemberExpr * expr ) { 87 const Type * aggrType = expr->aggregate->result->stripReferences(); 88 const AggregateDecl * decl = nullptr; 89 if ( auto inst = dynamic_cast<const StructInstType *>( aggrType ) ) { 90 decl = inst->base; 91 } else if ( auto inst = dynamic_cast<const UnionInstType *>( aggrType ) ) { 92 decl = inst->base; 93 } 94 assertf( decl, "Aggregate of member not correct type." ); 95 96 for ( auto aggrMember : decl->members ) { 97 if ( expr->member == aggrMember ) { 98 return; 99 } 100 } 101 assertf( false, "Member not found." ); 102 } 103 85 104 struct InvariantCore { 86 105 // To save on the number of visits: this is a kind of composed core. … … 108 127 } 109 128 129 void previsit( const MemberExpr * node ) { 130 previsit( (const ParseNode *)node ); 131 memberMatchesAggregate( node ); 132 } 133 110 134 void postvisit( const Node * node ) { 111 135 no_strong_cycles.postvisit( node ); -
src/Concurrency/Actors.cpp
r24d6572 r62d62db 38 38 bool namedDecl = false; 39 39 40 // finds and sets a ptr to the Allocation enum, which is needed in the next pass40 // finds and sets a ptr to the allocation enum, which is needed in the next pass 41 41 void previsit( const EnumDecl * decl ) { 42 if( decl->name == " Allocation" ) *allocationDecl = decl;42 if( decl->name == "allocation" ) *allocationDecl = decl; 43 43 } 44 44 … … 227 227 static inline derived_actor & ?|?( derived_actor & receiver, derived_msg & msg ) { 228 228 request new_req; 229 Allocation (*my_work_fn)( derived_actor &, derived_msg & ) = receive;229 allocation (*my_work_fn)( derived_actor &, derived_msg & ) = receive; 230 230 __receive_fn fn = (__receive_fn)my_work_fn; 231 231 new_req{ &receiver, &msg, fn }; … … 246 246 )); 247 247 248 // Function type is: Allocation (*)( derived_actor &, derived_msg & )248 // Function type is: allocation (*)( derived_actor &, derived_msg & ) 249 249 FunctionType * derivedReceive = new FunctionType(); 250 250 derivedReceive->params.push_back( ast::deepCopy( derivedActorRef ) ); … … 252 252 derivedReceive->returns.push_back( new EnumInstType( *allocationDecl ) ); 253 253 254 // Generates: Allocation (*my_work_fn)( derived_actor &, derived_msg & ) = receive;254 // Generates: allocation (*my_work_fn)( derived_actor &, derived_msg & ) = receive; 255 255 sendBody->push_back( new DeclStmt( 256 256 decl->location, … … 263 263 )); 264 264 265 // Function type is: Allocation (*)( actor &, message & )265 // Function type is: allocation (*)( actor &, message & ) 266 266 FunctionType * genericReceive = new FunctionType(); 267 267 genericReceive->params.push_back( new ReferenceType( new StructInstType( *actorDecl ) ) ); … … 269 269 genericReceive->returns.push_back( new EnumInstType( *allocationDecl ) ); 270 270 271 // Generates: Allocation (*fn)( actor &, message & ) = (Allocation (*)( actor &, message & ))my_work_fn;271 // Generates: allocation (*fn)( actor &, message & ) = (allocation (*)( actor &, message & ))my_work_fn; 272 272 // More readable synonymous code: 273 // typedef Allocation (*__receive_fn)(actor &, message &);273 // typedef allocation (*__receive_fn)(actor &, message &); 274 274 // __receive_fn fn = (__receive_fn)my_work_fn; 275 275 sendBody->push_back( new DeclStmt( … … 422 422 const StructDecl ** msgDecl = &msgDeclPtr; 423 423 424 // first pass collects ptrs to Allocation enum, request type, and generic receive fn typedef424 // first pass collects ptrs to allocation enum, request type, and generic receive fn typedef 425 425 // also populates maps of all derived actors and messages 426 426 Pass<CollectactorStructDecls>::run( translationUnit, actorStructDecls, messageStructDecls, requestDecl, -
src/Parser/lex.ll
r24d6572 r62d62db 10 10 * Created On : Sat Sep 22 08:58:10 2001 11 11 * Last Modified By : Peter A. Buhr 12 * Last Modified On : Tue May 2 08:45:21202313 * Update Count : 7 6912 * Last Modified On : Fri Jun 9 10:04:00 2023 13 * Update Count : 770 14 14 */ 15 15 … … 319 319 static { KEYWORD_RETURN(STATIC); } 320 320 _Static_assert { KEYWORD_RETURN(STATICASSERT); } // C11 321 _static_assert { KEYWORD_RETURN(STATICASSERT); } // C23 321 322 struct { KEYWORD_RETURN(STRUCT); } 322 323 suspend { KEYWORD_RETURN(SUSPEND); } // CFA -
src/Parser/parser.yy
r24d6572 r62d62db 10 10 // Created On : Sat Sep 1 20:22:55 2001 11 11 // Last Modified By : Peter A. Buhr 12 // Last Modified On : Wed Apr 26 16:45:37202313 // Update Count : 63 3012 // Last Modified On : Wed Jun 7 14:32:28 2023 13 // Update Count : 6341 14 14 // 15 15 … … 108 108 assert( declList ); 109 109 // printf( "distAttr1 typeSpec %p\n", typeSpec ); typeSpec->print( std::cout ); 110 DeclarationNode * c ur = declList, * cl = (new DeclarationNode)->addType( typeSpec );110 DeclarationNode * cl = (new DeclarationNode)->addType( typeSpec ); 111 111 // printf( "distAttr2 cl %p\n", cl ); cl->type->print( std::cout ); 112 112 // cl->type->aggregate.name = cl->type->aggInst.aggregate->aggregate.name; 113 113 114 for ( cur = dynamic_cast<DeclarationNode *>( cur->get_next() ); cur != nullptr; cur = dynamic_cast<DeclarationNode *>( cur->get_next() ) ) {114 for ( DeclarationNode * cur = dynamic_cast<DeclarationNode *>( declList->get_next() ); cur != nullptr; cur = dynamic_cast<DeclarationNode *>( cur->get_next() ) ) { 115 115 cl->cloneBaseType( cur ); 116 116 } // for … … 206 206 #define NEW_ONE new ExpressionNode( build_constantInteger( yylloc, *new string( "1" ) ) ) 207 207 #define UPDOWN( compop, left, right ) (compop == OperKinds::LThan || compop == OperKinds::LEThan ? left : right) 208 #define MISSING_ANON_FIELD " Missing loop fields with an anonymous loop index is meaningless as loop index is unavailable in loop body."209 #define MISSING_LOW " Missing low value for up-to range so index is uninitialized."210 #define MISSING_HIGH " Missing high value for down-to range so index is uninitialized."208 #define MISSING_ANON_FIELD "syntax error, missing loop fields with an anonymous loop index is meaningless as loop index is unavailable in loop body." 209 #define MISSING_LOW "syntax error, missing low value for up-to range so index is uninitialized." 210 #define MISSING_HIGH "syntax error, missing high value for down-to range so index is uninitialized." 211 211 212 212 static ForCtrl * makeForCtrl( … … 232 232 ForCtrl * forCtrl( const CodeLocation & location, DeclarationNode * index, ExpressionNode * start, enum OperKinds compop, ExpressionNode * comp, ExpressionNode * inc ) { 233 233 if ( index->initializer ) { 234 SemanticError( yylloc, " Direct initialization disallowed. Use instead: type var; initialization ~ comparison ~ increment." );234 SemanticError( yylloc, "syntax error, direct initialization disallowed. Use instead: type var; initialization ~ comparison ~ increment." ); 235 235 } // if 236 236 if ( index->next ) { 237 SemanticError( yylloc, " Multiple loop indexes disallowed in for-loop declaration." );237 SemanticError( yylloc, "syntax error, multiple loop indexes disallowed in for-loop declaration." ); 238 238 } // if 239 239 DeclarationNode * initDecl = index->addInitializer( new InitializerNode( start ) ); … … 260 260 return forCtrl( location, type, new string( identifier->name ), start, compop, comp, inc ); 261 261 } else { 262 SemanticError( yylloc, " Expression disallowed. Only loop-index nameallowed." ); return nullptr;262 SemanticError( yylloc, "syntax error, loop-index name missing. Expression disallowed." ); return nullptr; 263 263 } // if 264 264 } else { 265 SemanticError( yylloc, " Expression disallowed. Only loop-index name allowed." ); return nullptr;265 SemanticError( yylloc, "syntax error, loop-index name missing. Expression disallowed. ." ); return nullptr; 266 266 } // if 267 267 } // forCtrl 268 268 269 269 static void IdentifierBeforeIdentifier( string & identifier1, string & identifier2, const char * kind ) { 270 SemanticError( yylloc, ::toString( " Adjacent identifiers \"", identifier1, "\" and \"", identifier2, "\" are not meaningful in a", kind, ".\n"270 SemanticError( yylloc, ::toString( "syntax error, adjacent identifiers \"", identifier1, "\" and \"", identifier2, "\" are not meaningful in a", kind, ".\n" 271 271 "Possible cause is misspelled type name or missing generic parameter." ) ); 272 272 } // IdentifierBeforeIdentifier 273 273 274 274 static void IdentifierBeforeType( string & identifier, const char * kind ) { 275 SemanticError( yylloc, ::toString( " Identifier \"", identifier, "\" cannot appear before a ", kind, ".\n"275 SemanticError( yylloc, ::toString( "syntax error, identifier \"", identifier, "\" cannot appear before a ", kind, ".\n" 276 276 "Possible cause is misspelled storage/CV qualifier, misspelled typename, or missing generic parameter." ) ); 277 277 } // IdentifierBeforeType … … 689 689 // | RESUME '(' comma_expression ')' compound_statement 690 690 // { SemanticError( yylloc, "Resume expression is currently unimplemented." ); $$ = nullptr; } 691 | IDENTIFIER IDENTIFIER // syntax error691 | IDENTIFIER IDENTIFIER // invalid syntax rules 692 692 { IdentifierBeforeIdentifier( *$1.str, *$2.str, "n expression" ); $$ = nullptr; } 693 | IDENTIFIER type_qualifier // syntax error693 | IDENTIFIER type_qualifier // invalid syntax rules 694 694 { IdentifierBeforeType( *$1.str, "type qualifier" ); $$ = nullptr; } 695 | IDENTIFIER storage_class // syntax error695 | IDENTIFIER storage_class // invalid syntax rules 696 696 { IdentifierBeforeType( *$1.str, "storage class" ); $$ = nullptr; } 697 | IDENTIFIER basic_type_name // syntax error697 | IDENTIFIER basic_type_name // invalid syntax rules 698 698 { IdentifierBeforeType( *$1.str, "type" ); $$ = nullptr; } 699 | IDENTIFIER TYPEDEFname // syntax error699 | IDENTIFIER TYPEDEFname // invalid syntax rules 700 700 { IdentifierBeforeType( *$1.str, "type" ); $$ = nullptr; } 701 | IDENTIFIER TYPEGENname // syntax error701 | IDENTIFIER TYPEGENname // invalid syntax rules 702 702 { IdentifierBeforeType( *$1.str, "type" ); $$ = nullptr; } 703 703 ; … … 1152 1152 identifier_or_type_name ':' attribute_list_opt statement 1153 1153 { $$ = $4->add_label( yylloc, $1, $3 ); } 1154 | identifier_or_type_name ':' attribute_list_opt error // syntax error1155 { 1156 SemanticError( yylloc, ::toString( " Label \"", *$1.str, "\" must be associated with a statement, "1154 | identifier_or_type_name ':' attribute_list_opt error // invalid syntax rule 1155 { 1156 SemanticError( yylloc, ::toString( "syntx error, label \"", *$1.str, "\" must be associated with a statement, " 1157 1157 "where a declaration, case, or default is not a statement. " 1158 1158 "Move the label or terminate with a semi-colon." ) ); … … 1193 1193 | statement_list_nodecl statement 1194 1194 { assert( $1 ); $1->set_last( $2 ); $$ = $1; } 1195 | statement_list_nodecl error // syntax error1196 { SemanticError( yylloc, " Declarations only allowed at the start of the switch body, i.e., after the '{'." ); $$ = nullptr; }1195 | statement_list_nodecl error // invalid syntax rule 1196 { SemanticError( yylloc, "syntax error, declarations only allowed at the start of the switch body, i.e., after the '{'." ); $$ = nullptr; } 1197 1197 ; 1198 1198 … … 1219 1219 $$ = $7 ? new StatementNode( build_compound( yylloc, (StatementNode *)((new StatementNode( $7 ))->set_last( sw )) ) ) : sw; 1220 1220 } 1221 | SWITCH '(' comma_expression ')' '{' error '}' // CFA, syntaxerror1222 { SemanticError( yylloc, " Only declarations canappear before the list of case clauses." ); $$ = nullptr; }1221 | SWITCH '(' comma_expression ')' '{' error '}' // CFA, invalid syntax rule error 1222 { SemanticError( yylloc, "synatx error, declarations can only appear before the list of case clauses." ); $$ = nullptr; } 1223 1223 | CHOOSE '(' comma_expression ')' case_clause // CFA 1224 1224 { $$ = new StatementNode( build_switch( yylloc, false, $3, $5 ) ); } … … 1228 1228 $$ = $7 ? new StatementNode( build_compound( yylloc, (StatementNode *)((new StatementNode( $7 ))->set_last( sw )) ) ) : sw; 1229 1229 } 1230 | CHOOSE '(' comma_expression ')' '{' error '}' // CFA, syntax error1231 { SemanticError( yylloc, " Only declarations canappear before the list of case clauses." ); $$ = nullptr; }1230 | CHOOSE '(' comma_expression ')' '{' error '}' // CFA, invalid syntax rule 1231 { SemanticError( yylloc, "syntax error, declarations can only appear before the list of case clauses." ); $$ = nullptr; } 1232 1232 ; 1233 1233 … … 1268 1268 1269 1269 case_label: // CFA 1270 CASE error // syntax error1271 { SemanticError( yylloc, " Missing case listafter case." ); $$ = nullptr; }1270 CASE error // invalid syntax rule 1271 { SemanticError( yylloc, "syntax error, case list missing after case." ); $$ = nullptr; } 1272 1272 | CASE case_value_list ':' { $$ = $2; } 1273 | CASE case_value_list error // syntax error1274 { SemanticError( yylloc, " Missing colonafter case list." ); $$ = nullptr; }1273 | CASE case_value_list error // invalid syntax rule 1274 { SemanticError( yylloc, "syntax error, colon missing after case list." ); $$ = nullptr; } 1275 1275 | DEFAULT ':' { $$ = new ClauseNode( build_default( yylloc ) ); } 1276 1276 // A semantic check is required to ensure only one default clause per switch/choose statement. 1277 | DEFAULT error // syntax error1278 { SemanticError( yylloc, " Missing colonafter default." ); $$ = nullptr; }1277 | DEFAULT error // invalid syntax rules 1278 { SemanticError( yylloc, "syntax error, colon missing after default." ); $$ = nullptr; } 1279 1279 ; 1280 1280 … … 1405 1405 else { SemanticError( yylloc, MISSING_HIGH ); $$ = nullptr; } 1406 1406 } 1407 | comma_expression updowneq comma_expression '~' '@' // CFA, error1407 | comma_expression updowneq comma_expression '~' '@' // CFA, invalid syntax rules 1408 1408 { SemanticError( yylloc, MISSING_ANON_FIELD ); $$ = nullptr; } 1409 | '@' updowneq '@' // CFA, error1409 | '@' updowneq '@' // CFA, invalid syntax rules 1410 1410 { SemanticError( yylloc, MISSING_ANON_FIELD ); $$ = nullptr; } 1411 | '@' updowneq comma_expression '~' '@' // CFA, error1411 | '@' updowneq comma_expression '~' '@' // CFA, invalid syntax rules 1412 1412 { SemanticError( yylloc, MISSING_ANON_FIELD ); $$ = nullptr; } 1413 | comma_expression updowneq '@' '~' '@' // CFA, error1413 | comma_expression updowneq '@' '~' '@' // CFA, invalid syntax rules 1414 1414 { SemanticError( yylloc, MISSING_ANON_FIELD ); $$ = nullptr; } 1415 | '@' updowneq '@' '~' '@' // CFA, error1415 | '@' updowneq '@' '~' '@' // CFA, invalid syntax rules 1416 1416 { SemanticError( yylloc, MISSING_ANON_FIELD ); $$ = nullptr; } 1417 1417 … … 1431 1431 { 1432 1432 if ( $4 == OperKinds::GThan || $4 == OperKinds::GEThan ) { SemanticError( yylloc, MISSING_HIGH ); $$ = nullptr; } 1433 else if ( $4 == OperKinds::LEThan ) { SemanticError( yylloc, " Equality with missing high value is meaningless. Use \"~\"." ); $$ = nullptr; }1433 else if ( $4 == OperKinds::LEThan ) { SemanticError( yylloc, "syntax error, equality with missing high value is meaningless. Use \"~\"." ); $$ = nullptr; } 1434 1434 else $$ = forCtrl( yylloc, $3, $1, $3->clone(), $4, nullptr, NEW_ONE ); 1435 1435 } 1436 | comma_expression ';' '@' updowneq '@' // CFA, error1437 { SemanticError( yylloc, " Missing low/high value for up/down-to range so index is uninitialized." ); $$ = nullptr; }1436 | comma_expression ';' '@' updowneq '@' // CFA, invalid syntax rules 1437 { SemanticError( yylloc, "syntax error, missing low/high value for up/down-to range so index is uninitialized." ); $$ = nullptr; } 1438 1438 1439 1439 | comma_expression ';' comma_expression updowneq comma_expression '~' comma_expression // CFA 1440 1440 { $$ = forCtrl( yylloc, $3, $1, UPDOWN( $4, $3->clone(), $5 ), $4, UPDOWN( $4, $5->clone(), $3->clone() ), $7 ); } 1441 | comma_expression ';' '@' updowneq comma_expression '~' comma_expression // CFA, error1441 | comma_expression ';' '@' updowneq comma_expression '~' comma_expression // CFA, invalid syntax rules 1442 1442 { 1443 1443 if ( $4 == OperKinds::LThan || $4 == OperKinds::LEThan ) { SemanticError( yylloc, MISSING_LOW ); $$ = nullptr; } … … 1447 1447 { 1448 1448 if ( $4 == OperKinds::GThan || $4 == OperKinds::GEThan ) { SemanticError( yylloc, MISSING_HIGH ); $$ = nullptr; } 1449 else if ( $4 == OperKinds::LEThan ) { SemanticError( yylloc, " Equality with missing high value is meaningless. Use \"~\"." ); $$ = nullptr; }1449 else if ( $4 == OperKinds::LEThan ) { SemanticError( yylloc, "syntax error, equality with missing high value is meaningless. Use \"~\"." ); $$ = nullptr; } 1450 1450 else $$ = forCtrl( yylloc, $3, $1, $3->clone(), $4, nullptr, $7 ); 1451 1451 } 1452 1452 | comma_expression ';' comma_expression updowneq comma_expression '~' '@' // CFA 1453 1453 { $$ = forCtrl( yylloc, $3, $1, UPDOWN( $4, $3->clone(), $5 ), $4, UPDOWN( $4, $5->clone(), $3->clone() ), nullptr ); } 1454 | comma_expression ';' '@' updowneq comma_expression '~' '@' // CFA, error1454 | comma_expression ';' '@' updowneq comma_expression '~' '@' // CFA, invalid syntax rules 1455 1455 { 1456 1456 if ( $4 == OperKinds::LThan || $4 == OperKinds::LEThan ) { SemanticError( yylloc, MISSING_LOW ); $$ = nullptr; } … … 1460 1460 { 1461 1461 if ( $4 == OperKinds::GThan || $4 == OperKinds::GEThan ) { SemanticError( yylloc, MISSING_HIGH ); $$ = nullptr; } 1462 else if ( $4 == OperKinds::LEThan ) { SemanticError( yylloc, " Equality with missing high value is meaningless. Use \"~\"." ); $$ = nullptr; }1462 else if ( $4 == OperKinds::LEThan ) { SemanticError( yylloc, "syntax error, equality with missing high value is meaningless. Use \"~\"." ); $$ = nullptr; } 1463 1463 else $$ = forCtrl( yylloc, $3, $1, $3->clone(), $4, nullptr, nullptr ); 1464 1464 } 1465 1465 | comma_expression ';' '@' updowneq '@' '~' '@' // CFA 1466 { SemanticError( yylloc, " Missing low/high value for up/down-to range so index is uninitialized." ); $$ = nullptr; }1466 { SemanticError( yylloc, "syntax error, missing low/high value for up/down-to range so index is uninitialized." ); $$ = nullptr; } 1467 1467 1468 1468 | declaration comma_expression // CFA … … 1481 1481 { 1482 1482 if ( $3 == OperKinds::GThan || $3 == OperKinds::GEThan ) { SemanticError( yylloc, MISSING_HIGH ); $$ = nullptr; } 1483 else if ( $3 == OperKinds::LEThan ) { SemanticError( yylloc, " Equality with missing high value is meaningless. Use \"~\"." ); $$ = nullptr; }1483 else if ( $3 == OperKinds::LEThan ) { SemanticError( yylloc, "syntax error, equality with missing high value is meaningless. Use \"~\"." ); $$ = nullptr; } 1484 1484 else $$ = forCtrl( yylloc, $1, $2, $3, nullptr, NEW_ONE ); 1485 1485 } … … 1495 1495 { 1496 1496 if ( $3 == OperKinds::GThan || $3 == OperKinds::GEThan ) { SemanticError( yylloc, MISSING_HIGH ); $$ = nullptr; } 1497 else if ( $3 == OperKinds::LEThan ) { SemanticError( yylloc, " Equality with missing high value is meaningless. Use \"~\"." ); $$ = nullptr; }1497 else if ( $3 == OperKinds::LEThan ) { SemanticError( yylloc, "syntax error, equality with missing high value is meaningless. Use \"~\"." ); $$ = nullptr; } 1498 1498 else $$ = forCtrl( yylloc, $1, $2, $3, nullptr, $6 ); 1499 1499 } … … 1508 1508 { 1509 1509 if ( $3 == OperKinds::GThan || $3 == OperKinds::GEThan ) { SemanticError( yylloc, MISSING_HIGH ); $$ = nullptr; } 1510 else if ( $3 == OperKinds::LEThan ) { SemanticError( yylloc, " Equality with missing high value is meaningless. Use \"~\"." ); $$ = nullptr; }1510 else if ( $3 == OperKinds::LEThan ) { SemanticError( yylloc, "syntax error, equality with missing high value is meaningless. Use \"~\"." ); $$ = nullptr; } 1511 1511 else $$ = forCtrl( yylloc, $1, $2, $3, nullptr, nullptr ); 1512 1512 } 1513 | declaration '@' updowneq '@' '~' '@' // CFA, error1514 { SemanticError( yylloc, " Missing low/high value for up/down-to range so index is uninitialized." ); $$ = nullptr; }1513 | declaration '@' updowneq '@' '~' '@' // CFA, invalid syntax rules 1514 { SemanticError( yylloc, "syntax error, missing low/high value for up/down-to range so index is uninitialized." ); $$ = nullptr; } 1515 1515 1516 1516 | comma_expression ';' TYPEDEFname // CFA, array type … … 1521 1521 | comma_expression ';' downupdowneq TYPEDEFname // CFA, array type 1522 1522 { 1523 if ( $3 == OperKinds::LEThan || $3 == OperKinds::GEThan ) { SemanticError( yylloc, "All enumation ranges are equal (all values). Remove \"=~\"." ); $$ = nullptr; } 1523 if ( $3 == OperKinds::LEThan || $3 == OperKinds::GEThan ) { 1524 SemanticError( yylloc, "syntax error, all enumeration ranges are equal (all values). Remove \"=~\"." ); $$ = nullptr; 1525 } 1524 1526 SemanticError( yylloc, "Type iterator is currently unimplemented." ); $$ = nullptr; 1525 1527 } … … 1616 1618 MUTEX '(' argument_expression_list_opt ')' statement 1617 1619 { 1618 if ( ! $3 ) { SemanticError( yylloc, " mutex argument list cannot be empty." ); $$ = nullptr; }1620 if ( ! $3 ) { SemanticError( yylloc, "syntax error, mutex argument list cannot be empty." ); $$ = nullptr; } 1619 1621 $$ = new StatementNode( build_mutex( yylloc, $3, $5 ) ); 1620 1622 } … … 1664 1666 { $$ = build_waitfor_timeout( yylloc, $1, $3, $4, maybe_build_compound( yylloc, $5 ) ); } 1665 1667 // "else" must be conditional after timeout or timeout is never triggered (i.e., it is meaningless) 1666 | wor_waitfor_clause wor when_clause_opt timeout statement wor ELSE statement // syntax error1667 { SemanticError( yylloc, " else clause must be conditional after timeout or timeout never triggered." ); $$ = nullptr; }1668 | wor_waitfor_clause wor when_clause_opt timeout statement wor ELSE statement // invalid syntax rules 1669 { SemanticError( yylloc, "syntax error, else clause must be conditional after timeout or timeout never triggered." ); $$ = nullptr; } 1668 1670 | wor_waitfor_clause wor when_clause_opt timeout statement wor when_clause ELSE statement 1669 1671 { $$ = build_waitfor_else( yylloc, build_waitfor_timeout( yylloc, $1, $3, $4, maybe_build_compound( yylloc, $5 ) ), $7, maybe_build_compound( yylloc, $9 ) ); } … … 1709 1711 { $$ = new ast::WaitUntilStmt::ClauseNode( ast::WaitUntilStmt::ClauseNode::Op::LEFT_OR, $1, build_waituntil_timeout( yylloc, $3, $4, maybe_build_compound( yylloc, $5 ) ) ); } 1710 1712 // "else" must be conditional after timeout or timeout is never triggered (i.e., it is meaningless) 1711 | wor_waituntil_clause wor when_clause_opt timeout statement wor ELSE statement // syntax error1712 { SemanticError( yylloc, " else clause must be conditional after timeout or timeout never triggered." ); $$ = nullptr; }1713 | wor_waituntil_clause wor when_clause_opt timeout statement wor ELSE statement // invalid syntax rules 1714 { SemanticError( yylloc, "syntax error, else clause must be conditional after timeout or timeout never triggered." ); $$ = nullptr; } 1713 1715 | wor_waituntil_clause wor when_clause_opt timeout statement wor when_clause ELSE statement 1714 1716 { $$ = new ast::WaitUntilStmt::ClauseNode( ast::WaitUntilStmt::ClauseNode::Op::LEFT_OR, $1, … … 2065 2067 assert( $1->type ); 2066 2068 if ( $1->type->qualifiers.any() ) { // CV qualifiers ? 2067 SemanticError( yylloc, " Useless type qualifier(s) in empty declaration." ); $$ = nullptr;2069 SemanticError( yylloc, "syntax error, useless type qualifier(s) in empty declaration." ); $$ = nullptr; 2068 2070 } 2069 2071 // enums are never empty declarations because there must have at least one enumeration. 2070 2072 if ( $1->type->kind == TypeData::AggregateInst && $1->storageClasses.any() ) { // storage class ? 2071 SemanticError( yylloc, " Useless storage qualifier(s) in empty aggregate declaration." ); $$ = nullptr;2073 SemanticError( yylloc, "syntax error, useless storage qualifier(s) in empty aggregate declaration." ); $$ = nullptr; 2072 2074 } 2073 2075 } … … 2100 2102 | type_declaration_specifier 2101 2103 | sue_declaration_specifier 2102 | sue_declaration_specifier invalid_types 2103 { 2104 SemanticError( yylloc, ::toString( " Missing ';' afterend of ",2104 | sue_declaration_specifier invalid_types // invalid syntax rule 2105 { 2106 SemanticError( yylloc, ::toString( "syntax error, expecting ';' at end of ", 2105 2107 $1->type->enumeration.name ? "enum" : ast::AggregateDecl::aggrString( $1->type->aggregate.kind ), 2106 " declaration " ) );2108 " declaration." ) ); 2107 2109 $$ = nullptr; 2108 2110 } … … 2584 2586 // } // for 2585 2587 } 2588 | type_specifier field_declaring_list_opt '}' // invalid syntax rule 2589 { 2590 SemanticError( yylloc, ::toString( "syntax error, expecting ';' at end of previous declaration." ) ); 2591 $$ = nullptr; 2592 } 2586 2593 | EXTENSION type_specifier field_declaring_list_opt ';' // GCC 2587 2594 { $$ = fieldDecl( $2, $3 ); distExt( $$ ); } … … 2682 2689 | ENUM '(' cfa_abstract_parameter_declaration ')' attribute_list_opt '{' enumerator_list comma_opt '}' 2683 2690 { 2684 if ( $3->storageClasses.val != 0 || $3->type->qualifiers.any() ) 2685 { SemanticError( yylloc, "storage-class and CV qualifiers are not meaningful for enumeration constants, which are const." ); }2686 2691 if ( $3->storageClasses.val != 0 || $3->type->qualifiers.any() ) { 2692 SemanticError( yylloc, "syntax error, storage-class and CV qualifiers are not meaningful for enumeration constants, which are const." ); 2693 } 2687 2694 $$ = DeclarationNode::newEnum( nullptr, $7, true, true, $3 )->addQualifiers( $5 ); 2688 2695 } … … 2693 2700 | ENUM '(' cfa_abstract_parameter_declaration ')' attribute_list_opt identifier attribute_list_opt 2694 2701 { 2695 if ( $3->storageClasses.any() || $3->type->qualifiers.val != 0 ) { SemanticError( yylloc, "storage-class and CV qualifiers are not meaningful for enumeration constants, which are const." ); } 2702 if ( $3->storageClasses.any() || $3->type->qualifiers.val != 0 ) { 2703 SemanticError( yylloc, "syntax error, storage-class and CV qualifiers are not meaningful for enumeration constants, which are const." ); 2704 } 2696 2705 typedefTable.makeTypedef( *$6 ); 2697 2706 } … … 3166 3175 | IDENTIFIER IDENTIFIER 3167 3176 { IdentifierBeforeIdentifier( *$1.str, *$2.str, " declaration" ); $$ = nullptr; } 3168 | IDENTIFIER type_qualifier // syntax error3177 | IDENTIFIER type_qualifier // invalid syntax rules 3169 3178 { IdentifierBeforeType( *$1.str, "type qualifier" ); $$ = nullptr; } 3170 | IDENTIFIER storage_class // syntax error3179 | IDENTIFIER storage_class // invalid syntax rules 3171 3180 { IdentifierBeforeType( *$1.str, "storage class" ); $$ = nullptr; } 3172 | IDENTIFIER basic_type_name // syntax error3181 | IDENTIFIER basic_type_name // invalid syntax rules 3173 3182 { IdentifierBeforeType( *$1.str, "type" ); $$ = nullptr; } 3174 | IDENTIFIER TYPEDEFname // syntax error3183 | IDENTIFIER TYPEDEFname // invalid syntax rules 3175 3184 { IdentifierBeforeType( *$1.str, "type" ); $$ = nullptr; } 3176 | IDENTIFIER TYPEGENname // syntax error3185 | IDENTIFIER TYPEGENname // invalid syntax rules 3177 3186 { IdentifierBeforeType( *$1.str, "type" ); $$ = nullptr; } 3178 3187 | external_function_definition … … 3209 3218 | type_qualifier_list 3210 3219 { 3211 if ( $1->type->qualifiers.any() ) { SemanticError( yylloc, "CV qualifiers cannot be distributed; only storage-class and forall qualifiers." ); } 3220 if ( $1->type->qualifiers.any() ) { 3221 SemanticError( yylloc, "syntax error, CV qualifiers cannot be distributed; only storage-class and forall qualifiers." ); 3222 } 3212 3223 if ( $1->type->forall ) forall = true; // remember generic type 3213 3224 } … … 3220 3231 | declaration_qualifier_list 3221 3232 { 3222 if ( $1->type && $1->type->qualifiers.any() ) { SemanticError( yylloc, "CV qualifiers cannot be distributed; only storage-class and forall qualifiers." ); } 3233 if ( $1->type && $1->type->qualifiers.any() ) { 3234 SemanticError( yylloc, "syntax error, CV qualifiers cannot be distributed; only storage-class and forall qualifiers." ); 3235 } 3223 3236 if ( $1->type && $1->type->forall ) forall = true; // remember generic type 3224 3237 } … … 3231 3244 | declaration_qualifier_list type_qualifier_list 3232 3245 { 3233 if ( ($1->type && $1->type->qualifiers.any()) || ($2->type && $2->type->qualifiers.any()) ) { SemanticError( yylloc, "CV qualifiers cannot be distributed; only storage-class and forall qualifiers." ); } 3246 if ( ($1->type && $1->type->qualifiers.any()) || ($2->type && $2->type->qualifiers.any()) ) { 3247 SemanticError( yylloc, "syntax error, CV qualifiers cannot be distributed; only storage-class and forall qualifiers." ); 3248 } 3234 3249 if ( ($1->type && $1->type->forall) || ($2->type && $2->type->forall) ) forall = true; // remember generic type 3235 3250 } … … 3262 3277 $$ = $3; forall = false; 3263 3278 if ( $5 ) { 3264 SemanticError( yylloc, " Attributes cannot be associated with function body. Move attribute(s) before \"with\" clause." );3279 SemanticError( yylloc, "syntax error, attributes cannot be associated with function body. Move attribute(s) before \"with\" clause." ); 3265 3280 $$ = nullptr; 3266 3281 } // if -
src/ResolvExpr/CommonType.cc
r24d6572 r62d62db 1035 1035 void postvisit( const ast::TraitInstType * ) {} 1036 1036 1037 void postvisit( const ast::TypeInstType * inst) {}1038 1039 void postvisit( const ast::TupleType * tuple ) {1037 void postvisit( const ast::TypeInstType * ) {} 1038 1039 void postvisit( const ast::TupleType * tuple ) { 1040 1040 tryResolveWithTypedEnum( tuple ); 1041 1041 } -
src/ResolvExpr/Resolver.cc
r24d6572 r62d62db 1107 1107 1108 1108 /// Removes cast to type of argument (unlike StripCasts, also handles non-generated casts) 1109 void removeExtraneousCast( ast::ptr<ast::Expr> & expr , const ast::SymbolTable & symtab) {1109 void removeExtraneousCast( ast::ptr<ast::Expr> & expr ) { 1110 1110 if ( const ast::CastExpr * castExpr = expr.as< ast::CastExpr >() ) { 1111 1111 if ( typesCompatible( castExpr->arg->result, castExpr->result ) ) { … … 1197 1197 ast::ptr< ast::Expr > castExpr = new ast::CastExpr{ untyped, type }; 1198 1198 ast::ptr< ast::Expr > newExpr = findSingleExpression( castExpr, context ); 1199 removeExtraneousCast( newExpr , context.symtab);1199 removeExtraneousCast( newExpr ); 1200 1200 return newExpr; 1201 1201 } … … 1262 1262 static size_t traceId; 1263 1263 Resolver_new( const ast::TranslationGlobal & global ) : 1264 ast::WithSymbolTable(ast::SymbolTable::ErrorDetection::ValidateOnAdd), 1264 1265 context{ symtab, global } {} 1265 1266 Resolver_new( const ResolveContext & context ) : … … 2041 2042 const ast::Type * initContext = currentObject.getCurrentType(); 2042 2043 2043 removeExtraneousCast( newExpr , symtab);2044 removeExtraneousCast( newExpr ); 2044 2045 2045 2046 // check if actual object's type is char[] -
src/Validate/HoistStruct.cpp
r24d6572 r62d62db 18 18 #include <sstream> 19 19 20 #include "AST/DeclReplacer.hpp" 20 21 #include "AST/Pass.hpp" 21 22 #include "AST/TranslationUnit.hpp" 23 #include "AST/Vector.hpp" 22 24 23 25 namespace Validate { … … 51 53 template<typename AggrDecl> 52 54 AggrDecl const * postAggregate( AggrDecl const * ); 55 template<typename InstType> 56 InstType const * preCollectionInstType( InstType const * type ); 53 57 54 58 ast::AggregateDecl const * parent = nullptr; … … 66 70 qualifiedName( decl, ss ); 67 71 return ss.str(); 72 } 73 74 void extendParams( ast::vector<ast::TypeDecl> & dstParams, 75 ast::vector<ast::TypeDecl> const & srcParams ) { 76 if ( srcParams.empty() ) return; 77 78 ast::DeclReplacer::TypeMap newToOld; 79 ast::vector<ast::TypeDecl> params; 80 for ( ast::ptr<ast::TypeDecl> const & srcParam : srcParams ) { 81 ast::TypeDecl * dstParam = ast::deepCopy( srcParam.get() ); 82 dstParam->init = nullptr; 83 newToOld.emplace( srcParam, dstParam ); 84 for ( auto assertion : dstParam->assertions ) { 85 assertion = ast::DeclReplacer::replace( assertion, newToOld ); 86 } 87 params.emplace_back( dstParam ); 88 } 89 spliceBegin( dstParams, params ); 68 90 } 69 91 … … 74 96 mut->parent = parent; 75 97 mut->name = qualifiedName( mut ); 76 return mut;77 } else {78 GuardValue( parent ) = decl;79 returndecl;80 }98 extendParams( mut->params, parent->params ); 99 decl = mut; 100 } 101 GuardValue( parent ) = decl; 102 return decl; 81 103 } 82 104 … … 112 134 } 113 135 136 ast::AggregateDecl const * commonParent( 137 ast::AggregateDecl const * lhs, ast::AggregateDecl const * rhs ) { 138 for ( auto outer = lhs ; outer ; outer = outer->parent ) { 139 for ( auto inner = rhs ; inner ; inner = inner->parent ) { 140 if ( outer == inner ) { 141 return outer; 142 } 143 } 144 } 145 return nullptr; 146 } 147 148 template<typename InstType> 149 InstType const * HoistStructCore::preCollectionInstType( InstType const * type ) { 150 if ( !type->base->parent ) return type; 151 if ( type->base->params.empty() ) return type; 152 153 InstType * mut = ast::mutate( type ); 154 ast::AggregateDecl const * parent = 155 commonParent( this->parent, mut->base->parent ); 156 assert( parent ); 157 158 std::vector<ast::ptr<ast::Expr>> args; 159 for ( const ast::ptr<ast::TypeDecl> & param : parent->params ) { 160 args.emplace_back( new ast::TypeExpr( param->location, 161 new ast::TypeInstType( param ) 162 ) ); 163 } 164 spliceBegin( mut->params, args ); 165 return mut; 166 } 167 114 168 template<typename InstType> 115 169 InstType const * preInstType( InstType const * type ) { … … 121 175 122 176 ast::StructInstType const * HoistStructCore::previsit( ast::StructInstType const * type ) { 123 return preInstType( type);177 return preInstType( preCollectionInstType( type ) ); 124 178 } 125 179 126 180 ast::UnionInstType const * HoistStructCore::previsit( ast::UnionInstType const * type ) { 127 return preInstType( type);181 return preInstType( preCollectionInstType( type ) ); 128 182 } 129 183 -
tests/.expect/copyfile.txt
r24d6572 r62d62db 10 10 // Created On : Fri Jun 19 13:44:05 2020 11 11 // Last Modified By : Peter A. Buhr 12 // Last Modified On : Fri Jun 19 17:58:03 202013 // Update Count : 412 // Last Modified On : Mon Jun 5 21:20:07 2023 13 // Update Count : 5 14 14 // 15 15 … … 30 30 exit | "Usage" | argv[0] | "[ input-file (default stdin) [ output-file (default stdout) ] ]"; 31 31 } // choose 32 } catch( Open_Failure * ex ; ex->istream == &in ) {32 } catch( open_failure * ex ; ex->istream == &in ) { 33 33 exit | "Unable to open input file" | argv[1]; 34 } catch( Open_Failure * ex ; ex->ostream == &out ) {34 } catch( open_failure * ex ; ex->ostream == &out ) { 35 35 close( in ); // optional 36 36 exit | "Unable to open output file" | argv[2]; -
tests/.in/copyfile.txt
r24d6572 r62d62db 10 10 // Created On : Fri Jun 19 13:44:05 2020 11 11 // Last Modified By : Peter A. Buhr 12 // Last Modified On : Fri Jun 19 17:58:03 202013 // Update Count : 412 // Last Modified On : Mon Jun 5 21:20:07 2023 13 // Update Count : 5 14 14 // 15 15 … … 30 30 exit | "Usage" | argv[0] | "[ input-file (default stdin) [ output-file (default stdout) ] ]"; 31 31 } // choose 32 } catch( Open_Failure * ex ; ex->istream == &in ) {32 } catch( open_failure * ex ; ex->istream == &in ) { 33 33 exit | "Unable to open input file" | argv[1]; 34 } catch( Open_Failure * ex ; ex->ostream == &out ) {34 } catch( open_failure * ex ; ex->ostream == &out ) { 35 35 close( in ); // optional 36 36 exit | "Unable to open output file" | argv[2]; -
tests/concurrency/actors/dynamic.cfa
r24d6572 r62d62db 19 19 void ?{}( derived_msg & this ) { ((derived_msg &)this){ 0 }; } 20 20 21 Allocation receive( derived_actor & receiver, derived_msg & msg ) {21 allocation receive( derived_actor & receiver, derived_msg & msg ) { 22 22 if ( msg.cnt >= Times ) { 23 23 sout | "Done"; -
tests/concurrency/actors/executor.cfa
r24d6572 r62d62db 24 24 struct d_msg { inline message; } shared_msg; 25 25 26 Allocation receive( d_actor & this, d_msg & msg ) with( this ) {26 allocation receive( d_actor & this, d_msg & msg ) with( this ) { 27 27 if ( recs == rounds ) return Finished; 28 28 if ( recs % Batch == 0 ) { -
tests/concurrency/actors/inherit.cfa
r24d6572 r62d62db 15 15 void ^?{}( D_msg & this ) { mutex(sout) sout | 'A'; } 16 16 17 Allocation handle() {17 allocation handle() { 18 18 return Finished; 19 19 } 20 20 21 Allocation receive( Server & receiver, D_msg & msg ) { return handle(); }22 Allocation receive( Server & receiver, D_msg2 & msg ) { return handle(); }23 Allocation receive( Server2 & receiver, D_msg & msg ) { return Delete; }24 Allocation receive( Server2 & receiver, D_msg2 & msg ) { return Delete; }21 allocation receive( Server & receiver, D_msg & msg ) { return handle(); } 22 allocation receive( Server & receiver, D_msg2 & msg ) { return handle(); } 23 allocation receive( Server2 & receiver, D_msg & msg ) { return Delete; } 24 allocation receive( Server2 & receiver, D_msg2 & msg ) { return Delete; } 25 25 26 26 int main() { -
tests/concurrency/actors/matrix.cfa
r24d6572 r62d62db 24 24 } 25 25 26 Allocation receive( derived_actor & receiver, derived_msg & msg ) {26 allocation receive( derived_actor & receiver, derived_msg & msg ) { 27 27 for ( unsigned int i = 0; i < yc; i += 1 ) { // multiply X_row by Y_col and sum products 28 28 msg.Z[i] = 0; -
tests/concurrency/actors/pingpong.cfa
r24d6572 r62d62db 19 19 size_t times = 100000; 20 20 21 Allocation receive( ping & receiver, p_msg & msg ) {21 allocation receive( ping & receiver, p_msg & msg ) { 22 22 msg.count++; 23 23 if ( msg.count > times ) return Finished; 24 24 25 Allocation retval = Nodelete;25 allocation retval = Nodelete; 26 26 if ( msg.count == times ) retval = Finished; 27 27 *po << msg; … … 29 29 } 30 30 31 Allocation receive( pong & receiver, p_msg & msg ) {31 allocation receive( pong & receiver, p_msg & msg ) { 32 32 msg.count++; 33 33 if ( msg.count > times ) return Finished; 34 34 35 Allocation retval = Nodelete;35 allocation retval = Nodelete; 36 36 if ( msg.count == times ) retval = Finished; 37 37 *pi << msg; -
tests/concurrency/actors/poison.cfa
r24d6572 r62d62db 18 18 Server s[10]; 19 19 for ( i; 10 ) { 20 s[i] << FinishedMsg;20 s[i] << finished_msg; 21 21 } 22 22 stop_actor_system(); … … 29 29 Server * s = alloc(); 30 30 (*s){}; 31 (*s) << DeleteMsg;31 (*s) << delete_msg; 32 32 } 33 33 stop_actor_system(); … … 39 39 Server s[10]; 40 40 for ( i; 10 ) 41 s[i] << DestroyMsg;41 s[i] << destroy_msg; 42 42 stop_actor_system(); 43 43 for ( i; 10 ) -
tests/concurrency/actors/static.cfa
r24d6572 r62d62db 19 19 void ?{}( derived_msg & this ) { ((derived_msg &)this){ 0 }; } 20 20 21 Allocation receive( derived_actor & receiver, derived_msg & msg ) {21 allocation receive( derived_actor & receiver, derived_msg & msg ) { 22 22 if ( msg.cnt >= Times ) { 23 23 sout | "Done"; -
tests/concurrency/actors/types.cfa
r24d6572 r62d62db 20 20 21 21 // this isn't a valid receive routine since int is not a message type 22 Allocation receive( derived_actor & receiver, int i ) with( receiver ) {22 allocation receive( derived_actor & receiver, int i ) with( receiver ) { 23 23 mutex(sout) sout | i; 24 24 counter++; … … 27 27 } 28 28 29 Allocation receive( derived_actor & receiver, d_msg & msg ) {29 allocation receive( derived_actor & receiver, d_msg & msg ) { 30 30 return receive( receiver, msg.num ); 31 31 } … … 36 36 }; 37 37 38 Allocation receive( derived_actor2 & receiver, d_msg & msg ) {38 allocation receive( derived_actor2 & receiver, d_msg & msg ) { 39 39 mutex(sout) sout | msg.num; 40 40 return Finished; … … 48 48 }; 49 49 50 Allocation receive( derived_actor3 & receiver, d_msg & msg ) {50 allocation receive( derived_actor3 & receiver, d_msg & msg ) { 51 51 mutex(sout) sout | msg.num; 52 52 if ( msg.num == -1 ) return Nodelete; … … 54 54 } 55 55 56 Allocation receive( derived_actor3 & receiver, d_msg2 & msg ) {56 allocation receive( derived_actor3 & receiver, d_msg2 & msg ) { 57 57 mutex(sout) sout | msg.num; 58 58 return Finished; -
tests/concurrency/lockfree_stack.cfa
r24d6572 r62d62db 10 10 // Created On : Thu May 25 15:36:50 2023 11 11 // Last Modified By : Peter A. Buhr 12 // Last Modified On : Tue May 30 19:02:32202313 // Update Count : 1812 // Last Modified On : Fri Jun 9 14:01:07 2023 13 // Update Count : 68 14 14 // 15 15 … … 29 29 int64_t atom; 30 30 #endif // __SIZEOF_INT128__ 31 } __attribute__(( aligned( 16 ) ));31 }; 32 32 33 33 struct Node { … … 42 42 n.next = stack; // atomic assignment unnecessary 43 43 for () { // busy wait 44 if ( CASV( stack.atom, n.next.atom, ((Link){ &n, n.next.count + 1 }.atom) ) ) break; // attempt to update top node 44 Link temp{ &n, n.next.count + 1 }; 45 if ( CASV( s.stack.atom, n.next.atom, temp.atom ) ) break; // attempt to update top node 45 46 } 46 47 } … … 50 51 for () { // busy wait 51 52 if ( t.top == NULL ) return NULL; // empty stack ? 52 if ( CASV( stack.atom, t.atom, ((Link){ t.top->next.top, t.count }.atom) ) ) return t.top; // attempt to update top node 53 Link temp{ t.top->next.top, t.count }; 54 if ( CASV( stack.atom, t.atom, temp.atom ) ) return t.top; // attempt to update top node 53 55 } 54 56 } … … 57 59 Stack stack; // global stack 58 60 59 enum { Times = 60 #if defined( __ARM_ARCH ) // ARM CASV is very slow 61 10_000 62 #else 63 1_000_000 64 #endif // __arm_64__ 65 }; 61 enum { Times = 2_000_000 }; 66 62 67 63 thread Worker {}; … … 82 78 83 79 for ( i; N ) { // push N values on stack 84 // storage must be 16-bytes aligned for cmpxchg16b 85 push( stack, *(Node *)memalign( 16, sizeof( Node ) ) ); 80 push( stack, *(Node *)new() ); // must be 16-byte aligned 86 81 } 87 82 { -
tests/concurrency/waituntil/locks.cfa
r24d6572 r62d62db 2 2 #include <thread.hfa> 3 3 #include <locks.hfa> 4 #include <fstream.hfa> 4 5 #include <mutex_stmt.hfa> 5 6 -
tests/configs/.expect/parseconfig.txt
r24d6572 r62d62db 12 12 Maximum student trips: 3 13 13 14 Open_Failure thrown when config file does not exist14 open_failure thrown when config file does not exist 15 15 Failed to open the config file 16 16 -
tests/configs/parseconfig.cfa
r24d6572 r62d62db 66 66 67 67 68 sout | " Open_Failure thrown when config file does not exist";68 sout | "open_failure thrown when config file does not exist"; 69 69 try { 70 70 parse_config( xstr(IN_DIR) "doesnt-exist.txt", entries, NUM_ENTRIES, parse_tabular_config_format ); 71 } catch( Open_Failure * ex ) {71 } catch( open_failure * ex ) { 72 72 sout | "Failed to open the config file"; 73 73 } -
tests/copyfile.cfa
r24d6572 r62d62db 10 10 // Created On : Fri Jun 19 13:44:05 2020 11 11 // Last Modified By : Peter A. Buhr 12 // Last Modified On : Sat Aug 15 15:00:48 202013 // Update Count : 612 // Last Modified On : Mon Jun 5 21:20:19 2023 13 // Update Count : 7 14 14 // 15 15 … … 30 30 exit | "Usage" | argv[0] | "[ input-file (default stdin) [ output-file (default stdout) ] ]"; 31 31 } // choose 32 } catch( Open_Failure * ex ; ex->istream == &in ) {32 } catch( open_failure * ex ; ex->istream == &in ) { 33 33 exit | "Unable to open input file" | argv[1]; 34 } catch( Open_Failure * ex ; ex->ostream == &out ) {34 } catch( open_failure * ex ; ex->ostream == &out ) { 35 35 close( in ); // optional 36 36 exit | "Unable to open output file" | argv[2]; -
tests/rational.cfa
r24d6572 r62d62db 10 10 // Created On : Mon Mar 28 08:43:12 2016 11 11 // Last Modified By : Peter A. Buhr 12 // Last Modified On : Tue Jul 20 18:13:40 202113 // Update Count : 10 712 // Last Modified On : Mon Jun 5 22:58:09 2023 13 // Update Count : 108 14 14 // 15 15 … … 19 19 #include <fstream.hfa> 20 20 21 typedef Rational(int) RatInt;21 typedef rational(int) rat_int; 22 22 double convert( int i ) { return (double)i; } // used by narrow/widen 23 23 int convert( double d ) { return (int)d; } … … 25 25 int main() { 26 26 sout | "constructor"; 27 RatInt a = { 3 }, b = { 4 }, c, d = 0, e = 1;27 rat_int a = { 3 }, b = { 4 }, c, d = 0, e = 1; 28 28 sout | "a : " | a | "b : " | b | "c : " | c | "d : " | d | "e : " | e; 29 29 30 a = ( RatInt){ 4, 8 };31 b = ( RatInt){ 5, 7 };30 a = (rat_int){ 4, 8 }; 31 b = (rat_int){ 5, 7 }; 32 32 sout | "a : " | a | "b : " | b; 33 a = ( RatInt){ -2, -3 };34 b = ( RatInt){ 3, -2 };33 a = (rat_int){ -2, -3 }; 34 b = (rat_int){ 3, -2 }; 35 35 sout | "a : " | a | "b : " | b; 36 a = ( RatInt){ -2, 3 };37 b = ( RatInt){ 3, 2 };36 a = (rat_int){ -2, 3 }; 37 b = (rat_int){ 3, 2 }; 38 38 sout | "a : " | a | "b : " | b; 39 39 sout | nl; 40 40 41 41 sout | "comparison"; 42 a = ( RatInt){ -2 };43 b = ( RatInt){ -3, 2 };42 a = (rat_int){ -2 }; 43 b = (rat_int){ -3, 2 }; 44 44 sout | "a : " | a | "b : " | b; 45 sout | "a == 0 : " | a == ( Rational(int)){0}; // FIX ME46 sout | "a == 1 : " | a == ( Rational(int)){1}; // FIX ME45 sout | "a == 0 : " | a == (rational(int)){0}; // FIX ME 46 sout | "a == 1 : " | a == (rational(int)){1}; // FIX ME 47 47 sout | "a != 0 : " | a != 0; 48 48 sout | "! a : " | ! a; … … 73 73 74 74 sout | "conversion"; 75 a = ( RatInt){ 3, 4 };75 a = (rat_int){ 3, 4 }; 76 76 sout | widen( a ); 77 a = ( RatInt){ 1, 7 };77 a = (rat_int){ 1, 7 }; 78 78 sout | widen( a ); 79 a = ( RatInt){ 355, 113 };79 a = (rat_int){ 355, 113 }; 80 80 sout | widen( a ); 81 81 sout | narrow( 0.75, 4 ); … … 90 90 91 91 sout | "more tests"; 92 RatInt x = { 1, 2 }, y = { 2 };92 rat_int x = { 1, 2 }, y = { 2 }; 93 93 sout | x - y; 94 94 sout | x > y; … … 96 96 sout | y | denominator( y, -2 ) | y; 97 97 98 RatInt z = { 0, 5 };98 rat_int z = { 0, 5 }; 99 99 sout | z; 100 100 101 101 sout | x | numerator( x, 0 ) | x; 102 102 103 x = ( RatInt){ 1, MAX } + (RatInt){ 1, MAX };103 x = (rat_int){ 1, MAX } + (rat_int){ 1, MAX }; 104 104 sout | x; 105 x = ( RatInt){ 3, MAX } + (RatInt){ 2, MAX };105 x = (rat_int){ 3, MAX } + (rat_int){ 2, MAX }; 106 106 sout | x; 107 107
Note: See TracChangeset
for help on using the changeset viewer.