Started 3 mo 24 days ago
Took 14 min

Build #109 (Jun 1, 2021, 11:51:48 PM)

Distributing Tarball
Build Artifacts
cfa-cc- MB view
  1. add figure for call/return and exceptional control-flow (details)
  2. augment introduction chapter with places for extension (details)
  3. Changed how new pass visitor gets the return type of previsit to avoid a warning in gcc-11 (details)
  4. Removed unnecessary copy in loop (details)
  5. Changed signal handling to avoid incompatible pointer cast (details)
  6. Removed more gcc11 warnings (details)
  7. Changed sum type to uint32_t to make sure integer overflow isn't the problem with fast_lock (details)
  8. Fix cc1 for use with gcc11 (details)
  9. Fix cc1 -E flag for use with gcc11 (details)
  1. Fixed error in pivot ring figure (details)
  2. commiting before merge with peter's comments (details)
  3. Performance fixes to http benchmark (details)
  4. Fix a copy/paste mistake and removed a warning in nodebug (details)
  5. Fix bugs in workstealing edge cases: (details)
  6. formatting (details)
  7. update sections enumeration, coroutine, monitor, math, duration, time, clock (details)
  8. fix declaration ordering (details)
  9. Benchmarks now print stats at exit (details)
  10. Added cforall test for fairness. (details)
  11. Moved single_sem to rq_bench.hpp which was duplicated across multiple fibre benchmarks. (details)
  12. Added the transfer benchmark for libfibre (details)
  13. Recovered the polymorphic CoroutineCancelled exception. The interface might need a bit of polish. (details)
  14. Changed printing of stats to show percentage of pops rather than probe length. (details)
  15. Moved coroutine to global scope to ease debugging (details)
  16. SomeThreadCancelled -> ThreadCancelled: thread cancellations carry type information again. (details)
  17. divide ostream trait into basic_ostream and ostream where the former is used by strstream (details)
  18. formatting, rename public enum sepSize to ofstream_sepSize, change return type to bool for function fail (details)
  19. add first draft of strstream type (details)
  20. Rewrote the stats to use stringstream rather than fprintf. (details)
  21. io_drain now returns true if anything was unparked (details)
  22. Seperated semphore and scheduling logic in unpark (details)
  23. Changed enable interrupts: (details)
  24. Fudge variable in __schedule_thread so it's not accidently used. (details)
  25. Separate schedule_thread from the scheduler lock (details)
  26. Added stat for unparks that can migrate (details)
  27. change default cfathread_mutex to fast_lock (details)
  28. Split ready-queue routines in 3 instead of 2. (details)
  29. Some changes to stats and added back preferred (details)
  30. futures now optionally don't unpark the target thread (details)
  31. Fix to futures after last change (details)
  32. I/O drain now keeps the schedlock for the duration of the call. (details)
  33. change type of manipulator precision parameter from unsigned char to unsigned int (details)
  34. formatting, change lstlisting character to print in red, document manipulators eng and unit (details)
  35. A bit of clean-up, wrapping the detail of the default vtable up in a macro. (details)
  36. Andrew MMath: Switch from common.tex to cfalab.sty. Still work to do but it is almost everything I had before. (details)
  37. parse aggregate_control *, remove spurious yyy variable (details)
  38. move flush from basic_ostream to ostream, refactor istream into basic_istream and istream (details)
  39. remember basic_ostream nl and then override it using the basic_ostream version (details)
  40. rename strstream to ostrstream, first draft of istrstream but problem parsing (details)
  41. rename old name strstream to new name ostrstream (details)
  42. formatting (details)
  43. renaming, comment out non-working istrstream (details)
  44. Andrew MMath: LaTeX clean-up. Went through existing to make some of the examples more managable. (details)
  45. Andrew MMath: \CFAStyle now uses an actual listings style. (details)
  46. Pushing changed to RWlock with io_drain. (details)
  47. Fixed missing return. (details)
  48. add stream function clear, change eof to return bool (details)
  49. formatting (details)
  50. update stream documentation and add strstream documentation (details)
  51. change Latin-1 highlight character for color read in lstlisting, add subparagraph adjustment (details)
  52. Andrew MMath: Went over the features chapter again. (details)
  53. Changed RW lock to avoid hitting the global array on schedule. (details)
  54. Changed comments to ifdef (details)
  55. Improved support for new arrays subscripting by tuples, -[[-,-,-]]. (details)
  56. Activated syntax, forall( [N] ). (details)
  57. Andrew MMath: Cleaned out some cruft from cfalab.sty, simplifying the code there. (details)
  58. Added new subqueue implementation. (details)
  59. Andrew MMath: I'm going in circles with the introduction so here is the first version. (details)
  60. Replaced todonotes with the lighter weight todo package. (details)
  61. Removed old sub-queue (details)
  62. Forgot some changes for new sub-queue (details)
  63. More clean-up after new subqueue (details)
  64. Fixed new-array subscripting to work on 32-bit builds, partly with a workaround for #247. (details)
  65. raise minus sign in typewriter font to middle of line (details)
  66. add package enumitem to adjust spacing (details)
  67. proofread implementation chapter (details)
  68. Andrew MMath: Added command signatures in cfalab.sty. (details)
  69. Baseline commit of new linked-list implementation and test. (details)
  70. Revert "proofread implementation chapter", I've copied out the changes. (details)
  71. The cheap and chearful linked-list performance test (details)
  72. Disabling the part of the new-array subscripting test that isn't working on 32-bit (details)
  73. Removed call to polymorphic 'min' which caused MASSIVE slowdown. (details)
  74. Fix timestamp with new subqueue which was read after being cleared. (details)
  75. formatting (details)
  76. update to min/max to optimize (unsigned) int/long/long-long (details)
  77. Fixing two bugs in new linked list, which last night's build failure revealed. (details)
  78. Processors now have a pointer to their tls. (details)
  79. Changed stats to be more concise (details)
  80. 'print_stats_now' now pulls stats from processors. (details)
  81. Fixing a false failure of the new array test on x86-32 -O3. (details)
  82. Start ready queue cutoff at zero to avoid spurious steals. (details)
  83. add sty version of CFA common latex macros (details)
  84. harmonize common.tex and common.sty macros (details)
  85. update user manual to test common.sty (details)
  86. Adding linked-list convenience functions and testing a corner case. (details)
  87. Changing how the dlist API specifies which list direction you're working on. (details)
  88. Changed ready_queue to use forloop instead of hardcoded unfolded loop of 2 (details)
  89. No different types of enums are not interchangable. (Instance of _default_vtable is no longer extern.) (details)
  90. Replacing "Mike's old linked list" with "Mike's new linked list," including replatforming uses of the old one. (details)
  91. Moved timestamp declaration to kernel.hfa (details)
  92. Added count to the ready sub_queues (details)
  93. Updated alternate main for processors (details)
  94. Removed push stats in alternate main (details)
  95. New main now uses proper define. (details)
  96. Added function to poll ready-queue length. (details)
  97. Added a new attribute 'cfa_linkonce'. (details)
  98. Andrew MMath: First draft of the latest round of fixes to implement complete. More will be needed. (details)
  99. Fix sequential handling of timers (details)
  100. Moved stats cnt to fix 32-bit build (details)
  101. check for log10(0) in output manipulator eng (details)
  102. add 32-bit test for (details)
  103. remove architecture specific tests for manipulatorsOutput4 as the output is the same (details)
  104. put timeval calculation into alarm_node ctor to alleviate slowdowns due to preemption (details)
  105. Added writing points for each chapter of Mubeen's thesis. (details)
  106. Combined all the \code* commands and removed glossaries-extra's automake, which was generating a bunch of warnings. (details)
  107. Updated the virtual module to prefix the names with the new convention and change the parent_vtable into type_info. (details)
  108. made a naming improvement (details)
  109. Temporarily disabling test of new-array multidim subscripting on 32-bit, where it is failing. (details)
  110. Removed calls to wait functions which no longer exist (details)
  111. Small fixes to lock and alarm around possible race and the thunk problem. (details)
  112. Added test to detect regressions on bug #250 (details)
  113. In box pass that creates pointer + VLA for generics: (details)
  114. Adding isListed accessor function to new linked list. (details)
  115. comment on Mubeen's thesis writing-points (details)
  116. Re-worked new array declarations to avoid suggestion that a type-system managed value needs to appear in type position.  The arpk S parameter is now clearly a type, not a length. (details)
  117. Cleanup stats for concision. (details)
  118. Drain on I/O (details)
  119. Small fixes to how timestamp is handled in workstealing queue. (details)
  120. Added option to ready-queue to push ignoring locality. (details)
  121. Added stat array tools for processing (details)
  122. Andrew MMath: Work on figures and linkonce. (details)
  123. Disable preemption in clib_tls. (details)
  124. added updated timeout_lock text (details)
  125. Commit current version (details)
  126. add missing heap statistic counters for free calls and free storage (details)
  127. Andrew MMath: Converted all the ASCII diagrams to xfig diagrams. (details)
  128. Andrew MMath: Handled some too long lines caused by code snipits. Added a new command: \snake. (details)
  129. removed semaphore usage and used a lock when checking if empty (details)
  130. Added input script for setting up archive debugging. (details)
  131. Jenkins now generates a script when archiving test failures. (details)
  132. Build description now cotains node name (details)
  133. Fix error archival to match's expectation (details)
  134. alarm.hfa is now a public header (details)
  135. Tentative fix to make it more robust (details)
  136. Don't archive librairies if no test crashed. (details)
  137. Andrew MMath: Rewrote the existing features/references piece. (details)
  138. Tentative fix to make more robust (details)
  139. Fixed how names com files. (details)
  140. Full build no longer builds the old ast (details)
  141. Fixed typo in jenkins file (details)
  142. update uC++ keywords for lstlisting language (details)
  143. clean up macros and add defaultdialect={CFA} (details)
  144. update macros to reflect changed to common macros, remove otype (details)
  145. Fixed INSTALL file which had out-of-date directions. (details)
  146. Improved INSTALL doc to take into account both devs and users. (details)
  147. revert change from \usepackage{common} to \input{common} (details)
  148. add figure for call/return and exceptional control-flow (details)
  149. augment introduction chapter with places for extension (details)
  150. Changed how new pass visitor gets the return type of previsit to avoid a warning in gcc-11 (details)
  151. Removed unnecessary copy in loop (details)
  152. Changed signal handling to avoid incompatible pointer cast (details)
  153. Removed more gcc11 warnings (details)
  154. Changed sum type to uint32_t to make sure integer overflow isn't the problem with fast_lock (details)
  155. Fix cc1 for use with gcc11 (details)

Started by upstream project Cforall Full Build build number 1749
originally caused by:

  • Started by timer
Revision: 6da81c7ca2f43e2698f1a0886ad3e350cba32a2e
  • refs/remotes/origin/master
Revision: a804f81298e916805241c4eb6b0be21bac7c648e
  • detached