Branch updates

Reini Urban rurban at x-ray.at
Wed Jan 2 16:38:29 UTC 2013


I merged the following branches:

rurban/makefile-codas-gh901
 - Provide missing makefile codas

rurban/llvm-gh853
         + Improved Configure.pl probes for --with-llvm

rurban/track-generated-gh890
       + Install forgotten runtime include files:
          cloneflags.pasm
          packfile_annotation_keytype.pasm
          packfile_segments.pasm
        + Install include/parrot/pbcversion.h with the main package, not devel.
        + Install headers for dynpmc files.
        + Install libraries tgc.pbc, NCI/Utils.pbc, OpenGL/Math.pbc,
PCT/Dumper.pbc,
          postgres.pbc, ProfTest.pbc and ProfTest/*pbc,
String/Utils.pbc, URI/Escape.pbc,
          YAML/Dumper/Base.pbc, YAML/Dumper/Default.pbc
        + Generate MANIFEST.generated, cleanup MANIFEST,
          removed MANIFEST_configure.generated, add installation of
forgotten files
          [GH #890]

rurban/win-replace_slashes-gh891
        + Replace / slashes with \ on windows for all Makefiles automatically.
          This improves cygwin interaction, like remote or cronjob
smoking on windows.

library_path-gh368
    + Remove duplicate library search path entries, move "." from the
beginning to
       the end. [GH 368]

I also fixed these tickets:
* GH #906 errors in MANIFEST.generated
  Thanks to fperrrad
* GH #887, failing make testr with native_pbc
  Thanks to jim keenan
* GH #905 Do not add apple -Wno-long-double with custom --cc=
* GH #902 Add callcontext.pmc to pmc.num
  I kept the ticket open, if 53 is the best index for callcontext.


I worked on these branches, which are ready to be merged.
I'll do that until Saturday.

rurban/threaded-say-gh893
  only concat if strlen < 80
  realloc was unavoidable since our string registers are always
constant. (thanks
  But the GC is fast. So improve at least the most common cases.

line-directives-gh641
  Renable line-directives for pmc. I found no problem and debugging
and error reporting was vastly improved.

rurban/pcre-dynext_libs-gh302
  new dynext_dirs config entry, new ENV var PARROT_DYNEXT, new add_env_paths()
  test in lib_search_path.t

rurban/mult-parrot-env-gh903
  Support multiple paths also for PARROT_LIBRARY, PARROT_INCLUDE

rurban/nci-void-gh904
  Fix various NCI issues:
    Fix "p v" processing GH #904,
    Fix declarations, check --core, use --dynext for glut
    Fix broken src/nci/extra_thunks Makefile rules
    Fix broken nci examples
    Update pdd16_native_call.pod

rurban/skipped-gh884
   Improved Configure steps reporting


Almost done, and will be merged until Saturday:

rurban/pod-copyright-gh851
   Add copyright checks for pod files. One ext/ case clarified and missing.

GH 886 needs to be addressed also. killing a thread in task.t works
but is still not stable enough to be tested.
I skipped windows/cygwin since there we have no signals yet, but other
platforms also reported spurious out of
order results. Re-adding the busy sleep will help in stabilization,
but it is not foolproof.


I continue to work on these branches, help appreciated:

!!! rurban/fix-gc-thr-gh880+gh875
  GC should stop walking into another thread interp.
  Someone should help me with this blocker. I also get work_list asserts,
  and there's not even a threaded GC stress test yet.
  Easy to repro: parrot -t7 and some threaded code.

rurban/ops2c-lines-gh641
  I also started with the missing ops2c line directives

http-daemon-gh898
  fix the library version (90% done)

rurban/test-capture-gh892
  get rid of lib/IO/CaptureOutput.pm and fix config step tests (minor)

cyg-desparse-gh888
  cygwin upstream reported massive success, but I don't see the big
advantage yet.

native_pbc2

-- 
Reini Urban
http://cpanel.net/   http://www.perl-compiler.org/


More information about the parrot-dev mailing list