[libindirect-perl] annotated tag upstream/0.33 created (now 513aace)

gregor herrmann gregoa at debian.org
Fri Oct 17 14:14:14 UTC 2014


This is an automated email from the git hooks/post-receive script.

gregoa pushed a change to annotated tag upstream/0.33
in repository libindirect-perl.

        at  513aace   (tag)
   tagging  ee22364ac200616d0db2c8dfe77862c30a79b027 (commit)
  replaces  upstream/0.31
 tagged by  gregor herrmann
        on  Fri Oct 17 16:08:37 2014 +0200

- Log -----------------------------------------------------------------
Upstream version 0.33

Andrew Main (2):
      Make indirect play nice with Devel::CallParser
      Partially revert 502e3aac.

Florian Ragwitz (1):
      Remove unused variable.

Vincent Pit (287):
      Importing indirect-0.01.tar.gz
      Just test the order. This should be enough with our implementation
      In ck_entersub, only try to recover the method name if it's actually a method
      Show more information when we couldn't run the tests data files
      Consting
      Commenting
      Prefer SvPV_*_const functions
      Skip when we couldn't capture buffers
      Remove the possibly miscompiled syntax ((hints == 2) ? croak : warn)(msg)
      Force UOK flag onto op->src values
      Make the hint a IV
      This is 0.02
      Actually exit when the child didn't succeed
      Add PERL5OPT to the flags passed to the forked perls
      Add undef guards to $stderr in tests
      This is 0.03
      Better logic for finding oo and om in ck_entersub
      No longer fork to test. IPC::Cmd is no longer required
      Remove all the kid op tags from the map when we're done with an entersub
      Call the previous check function earlier when possible
      Test method call inside method call
      Add a link to the coverage report
      This is 0.04
      Now the old check function is always called before storing an op into the map
      Test quotelike environments
      Get rid of an unused variable
      Be proud about not using a source filter
      This is 0.05
      Discrepancy @args/@opts
      Preferring to say 'to say'
      Test 'no indirect "hlagh"'
      Better be on irc.perl.org
      Test 'package A; sub foo; foo A->new' that gets deparsed as 'A->foo->new'
      Test system and exec
      This is 0.06
      Get rid of Debian_CPANTS.txt
      Test 'print' that segfaults
      Fix getting the name of a GV(SV) kid of RV2SV that caused 'print' to segfault
      This is 0.07_01
      More tests with $_
      Test good uses of our variables
      Add support and tests for package variables
      Even if len is 0 in indirect_find(), it'll still return s
      Hide a very unlikely failure from coverage
      Tidy up the new code in ck_rv2sv
      This is 0.07_02
      Factorize the regexp in t/20-bad.t
      Support variables with spaces after the sigil
      padmy names may begin with a non-alnum. I think
      Make test descriptions shorter
      This is 0.07_03
      Test $$ as object
      Brr. Those strstr() really were fugly
      Fix a rare edge case for package whose names are prefix of 'main'
      Document the failure to see an indirect syntax at the very end of the file (without semicolon)
      Test state variables
      This is 0.08
      Really test that no indirect 'hlagh' does not croak
      In tests, say 'no/use indirect' inside the eval string, in preparation of 5.8 backport
      Remove the 'inside s///' test for now, as it seems to randomly fail
      Allow specifying a prefix and skipping individual tests
      "use feature 'state'" as a prefix, so that we'll be able to skip it on older perls
      Support for perl 5.8
      Test that the code compiled fine in t/30-scope.t
      Test if the pragma propagates through eval string
      Test eval block
      Skip the segfaulting "${\(new Hlagh)}" test
      Document the pragma not propagating into 'eval STRING' on 5.8
      This is 0.09
      In t/30-scope.t, avoid failures for "eval 'no indirect; $x = new Foo'" for being seen as 'more than two warnings' because of an uninitialized $w
      Really fix the "${\(new Hlagh)}" bug by ensuring that the kid of a rv2sv is a SVOP before looking for its sv
      Phrasing fix
      Better handling of extraneous warnings in t/30-scope.t
      This is 0.09_01
      Cleanup indirect_map_clean() recursion
      Move the CAVEATS paragraph above DEPENDENCIES
      Bump copyright year
      Yet more diagnostics for t/30-scope.t
      This is 0.10
      Fix a possible collisions by hashing pointers with the format %u
      Silence a visibility warning
      Add a cast to make icc happy
      Use recent ExtUtils::MakeMaker features to generate a better META.yml
      This is 0.11
      Consting fix
      Silence an "unused return value" warning
      Fix negation precedence
      Somewhat improve the readability of indirect_ck_entersub()
      Don't even try to fetch the object if the method couldn't be found
      Point the repository to the gitweb instead of the bare repo
      Test use/no indirect while parsing an indirect construct
      Tighten the scope of pl_linestr in indirect_map_store()
      Slashes are reserved and hence must be encoded in the search part of an URL
      Switch to a ptable-based OP map
      Don't hardcode the package name as the hints key
      Make the hint thread-safe and store a coderef called each time an indirect construct is spotted
      Introduce the I_THREADSAFE() constant
      Make the op map thread safe
      Disable thread safety for 5.8.x on Windows
      Rename some test files
      Introduce the 'hook' unimport option
      Freshen the synopsis
      POD cleanup
      This is 0.12
      Make names stored in the op map readonly
      Pass copies as hook arguments
      Also pass the file and the line number to the hook
      5.9.5 fix
      Lower the ptable initial size
      Get rid of t/90-boilerplate.t
      Add a TODO test for the "pragma propagating into require" issue
      We want to test code in ck_entersub(), not hint
      Stop leaking the package name passed to the hook
      Rename 'id' to 'hint' for consistency
      Work around the "pragma propagating into require" bug
      Test the workaround more extensively
      This is 0.13
      Prevent bogus invalid syntaxes caused by reallocated memory chunks
      Fix a small leak of hints objects
      Don't actually create an entry when storing NULL in the ptable
      Use a better metasyntactic name instead of C<sub>
      Be proud about our workarounds
      This is 0.14
      A few POD nits
      Factor the cloning logic in a separate function
      Store a proper op info instead of hijacking a SV
      Fix line number for multiline indirect constructs
      Add a stress test
      Fix "meth meth" not being correctly reported
      This is 0.15
      Revamp t/21-bad.t
      Only enable the pragma during compile time
      Make the default error message public
      Handle indirect calls on blocks
      Document that we know handle blocks
      Simplify the warning handler in t/21-bad.t
      Rename try() to zap() as we'll need try() for something else soon
      Freshen t/20-good.t
      Freshen t/22-bad-mixed.t
      Thanks Andrew and Florian
      Document the diagnostics
      Remove ck_leave()
      Cover ck_scope actually handling an OP_SCOPE
      This is 0.16
      Document and test constant folding
      Fix t/80-regressions.t failing on Win32
      Turn on CvCLONE for all anonymous subs passed as hooks
      Replace an occurrence of SvREFCNT_inc by SvREFCNT_inc_simple_NN
      Freshen t/30-scope.t
      Don't use 'Fatal' as a metasyntactic name
      Freshen t/10-args.t
      More freshness for t/21-bad.t
      Freshen t/40-threads.t
      This is 0.17
      Add the PERL_INDIRECT_PM_DISABLE environment variable
      Make sure PERL_INDIRECT_PM_DISABLE is not set before running tests
      This is 0.18
      SvIOK and SvROK imply SvOK
      Fix signedness of the pointer<->int conversions
      Use the hint pointer as the unique identifier for the %^H entry
      I_THREADSAFE ought to be true when PERL_INDIRECT_PM_DISABLE is set
      Only remap the hint through the pointer table for threaded perls
      c07766a234204d78a80b6d75534e03ae3465b6f3 is a better fix for RT #47902 than 4cb0e2cc26fbe39a7a040b090b1a1e6c2380af0f
      Make some room in t/4*.t
      Reset the callbacks when the root interpreter is destroyed
      We don't need this (wrong) macro wrapper for indirect_teardown()
      Introduce I_FORKSAFE
      This is 0.19
      Kwalitee test overhaul
      Only look up to the first require scope for the require propagation workaround
      Give an explicit value to dynamic_config
      Put prerequisites in their own separate hash
      Build the $name and the main $file from $dist
      Re-indent WriteMakefile() arguments
      Fix for {ENTER,LEAVE}_with_name
      Update ptable.h
      Add a link to Matt's blog post
      Bump copyright year
      {ENTER,LEAVE}_with_name are no longer needed starting perl 5.11.4
      Identify the require scope by the outmost cv outside of the current one
      Add a regression test for RT #50570
      Add a regression test for a bug that appeared while fixing RT #50570 but got re-fixed in a4ebcbe07de579d5877df84c114f162f49019724
      Protect $@ in the regression test for RT #50570
      Fix test failures with 5.12 on Windows where Strawberry Perl crashes because the SystemRoot environment variable is missing
      Remove an unused variable
      This is 0.20
      Disable taint mode for t/30-scope.t
      Add Tatsuhiko Miyagawa's regression test
      MY_CXT_INIT can be a declaration, so put it in a block to be strictly conformant to C89
      Use a PPCODE: section in CLONE
      Require perl 5.8.1
      Fix the thread destructor trick for 5.13.1
      This is 0.21
      Only check methods where the method name is a constant
      Allow /^:?fatal$/ to specify the lethal behaviour
      Typos
      Get rid of the linestr check
      More quotelike environments testing
      Yet more quotelike tests
      This is 0.22
      Properly handle method_named ops
      Store the offset from the source buffer beginning in the op info object
      Remove some dead code in indirect_ck_method
      Rename t/80-regressions.t to t/50-external.t
      Document and test that the "-e without semicolon" bug is fixed on 5.12
      Always get the hint directly from %^H
      Refine indirect_find
      Get rid of indirect_is_indirect
      Get rid of CALL_FPTR()
      Ignore MYMETA.yml
      Skip threads tests unless perl version is 5.13.4 or greater
      This is 0.23
      Update ptable.h
      Bump copyright year
      Switch to qw<>
      Clarify our policy towards C++ compilers
      Only declare dMY_CXT when really needed
      Revert "Always get the hint directly from %^H"
      Test interaction with eval STRING
      Test that indirect does not vivify entries in %^H
      Replace $] by "$]" in tests
      Add META.json
      Ignore MYMETA.json
      This is 0.24
      The last argument to hv_fetch is the lvalue flag, not the hash
      Remove remaining occurences of $^V
      Encode the README file in UTF-8
      Try to prevent t/50-external.t from crashing on Cygwin
      This is 0.25
      Skip the hint entry vivification test on 5.11
      Stop leaking the SVs stored in the hints structure.
      Add 'global' option to 'no indirect'
      Also forcefully reset the localize hint flag for use indirect
      A couple of cosmetic code changes
      Make sure code examples fit in a 80 columns terminal
      Freshen the documentation
      Forbid passing 'hook' and 'fatal' at the same time
      Port threads test to Test::Leaner and move boilerplate to an helper module
      Port t/46-stress.t to Test::Leaner
      Test "no indirect 'global'" with threads
      Revert "Try to prevent t/50-external.t from crashing on Cygwin"
      Preserve the PATH environment variable when running a sub-perl on cygwin
      This is 0.26
      Reorder the members of the indirect_op_info struct a bit
      Make sure the POD headings are linkable
      t/40-threads.t should always run at least one test
      Require Module::ExtractUse 0.24 for the Kwalitee test
      Update author tests, and add a POD spelling test
      Kill Debian_CPANTS.txt suffixes as well
      Quote version numbers in Makefile.PL
      Add BUILD_REQUIRES to WriteMakefile()
      Port module loading in tests to VPIT::TestHelpers
      Update VPIT::TestHelpers to e8344578
      Bump copyright year
      This is 0.27
      Silence an "unused result" compiler warning
      Properly set and check the line number of method and object tokens
      Clarify what indirect_find() does by renaming some variables
      Make indirect_find() more resilient to embedded nuls
      Replace check functions in a thread-safe way
      This is 0.28
      Force linking against the perl dll when using gcc 3.4 on Windows
      In string-like envs, take the position to the beginning of the string
      This is 0.29
      Update the bug tracker URL in META after the rt.perl.org upgrade
      This is 0.30
      Get rid of t/99-kwalitee.t
      Stop bundling author tests
      No tabs please
      Remove obsolete kwalitee workarounds
      Make Perl version numbers more readable
      Update bundled Test::Leaner to 0.05
      Handle __PACKAGE__ as object name correctly
      Remove profanity
      Add some tests for 'new { __PACKAGE__ }'
      This is 0.31
      Silence a sign-compare warning
      Use run_perl() from VPIT::TestHelpers
      Add support for PERL_OP_PARENT
      Don't segfault for indirect constructs inside the empty package
      Bump copyright year
      This is 0.32
      Don't cleanup thread local storage before END blocks are executed
      Nullify thread local storage entries when they are freed
      Use the new CLONE_PARAMS API with perl 5.13.2
      Do nothing after that the thread local storage has been freed
      This is 0.33

gregor herrmann (1):
      Imported Upstream version 0.33

-----------------------------------------------------------------------

No new revisions were added by this update.

-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-perl/packages/libindirect-perl.git



More information about the Pkg-perl-cvs-commits mailing list