karl williamson [Tue, 9 Dec 2008 04:59:05 +0000]
[perl #59908] \x, \0, and \N{} not ok in double-quotish when followed by > \x100
Attached is a patch for this problem.
The root cause was that S_scan_const() was not recoding to utf8 under
some circumstances when it should be.
I also changed it so that in all places, the flag that indicates the
output is in utf8 is changed from false to true if and only if the
destination is recoded to utf8. One place was skipping this, and then
setting it unconditionally later on.
In one place in the routine, the routine had code to do the recoding
itself. In the other places, it called sv_utf8_upgrade(). I changed it
to call the subroutine in all cases.
I fixed a bug that would appear only on EBCDIC machines where constants
of the form \N{U+....} would have been interpreted as EBCDIC.
And in inspecting the code, I realized there were problems with growing
the scalar value to fit the input. I cleaned those up.
I also added a number of comments to document things I found out, and
changed some existing ones to be more accurate.
Since no one responded to my request for where to put the test cases,
and I couldn't figure out a good place to put them, I added a new test
file, t/uni/lex_utf8.t.
David Golden [Wed, 7 Jan 2009 04:07:47 +0000]
make add-package git friendly and fix bugs (take 2)
(1) all work done in a new git branch by default
(2) "-n" flag ("no branch") added to get old rsync-friendly behavior
(3) looks in "scripts/" directory as well as "bin/" for executables
(4) patches utils/Makefile.SH instead of utils/Makefile
Steve Hay [Wed, 14 Jan 2009 18:14:24 +0000]
Note Math-Complex's test files in Porting/Maintainers.pl
Steve Hay [Wed, 14 Jan 2009 18:12:18 +0000]
Upgrade to Math-Complex-1.55
Steve Hay [Wed, 14 Jan 2009 18:08:16 +0000]
Upgrade to ExtUtils-Manifest-1.56
Steve Hay [Wed, 14 Jan 2009 18:02:38 +0000]
Upgrade to ExtUtils-Command-1.16
Steve Hay [Wed, 14 Jan 2009 17:46:43 +0000]
Upgrade to Module-Build-0.31012
We're now in sync with CPAN--no local changes remain in blead.
TODO: Various extra core changes are now required to handle the new bundle.pl script as per the existing config_data script.
Steve Hay [Wed, 14 Jan 2009 12:21:59 +0000]
Allow '\r' after "#line XXX" directives.
Fixes lib/feature.t and lib/warnings.t tests that have been failing on Win32 since commit 26b6dc3.
Aaron Crane [Wed, 14 Jan 2009 09:51:44 +0000]
applying patches and upstream
Nicholas Clark writes:
> $ perl Porting/Maintainers --opened
> (No files are modified)
I think that message would be even more helpful if it went to stderr;
(trivial) patch attached.
--
Aaron Crane ** http://aaroncrane.co.uk/
From
afa6ac1c2f6bde792be6ac065a87119e1dbdafee Mon Sep 17 00:00:00 2001
From: Aaron Crane <perl@aaroncrane.co.uk>
Date: Wed, 14 Jan 2009 09:46:33 +0000
Subject: [PATCH] Send informative output to stderr, not stdout
Nicholas Clark [Wed, 14 Jan 2009 10:34:23 +0000]
Actually commit the state infomation at the end of a test run.
Reduces my elapsed time by about 5% threaded or unthreaded on the second run.
Nicholas Clark [Wed, 14 Jan 2009 10:32:44 +0000]
Make make_ext.pl write out a shell script of its actions for make distclean etc
(In case the user runs make clean; make distclean; as clean removes miniperl)
Change the #! line to ./miniperl
Nicholas Clark [Tue, 13 Jan 2009 18:57:43 +0000]
If there is one argument, and it looks like a commit, list data from the files
that commit changed. For example, `perl Porting/Maintainers HEAD`.
Nicholas Clark [Tue, 13 Jan 2009 17:57:54 +0000]
Add the upstream status to the output of Porting/Maintainers for --opened.
Nicholas Clark [Tue, 13 Jan 2009 17:51:20 +0000]
chomp the output from git ls-files. Display a helpful message nothing is
modified, rather than failing back to the usage message.
Nicholas Clark [Tue, 13 Jan 2009 17:08:36 +0000]
ext/Hash/Util/Makefile.PL should not recurse to FieldHash's Makefile.PL
*nix, VMS and Win32 all know how to (and have to) call the latter directly.
As is, targets in ext/Hash/Util/FieldHash get called twice, which may result
in race conditions, and certainly messes up make clean; make distclean;
Jerry D. Hedden [Tue, 13 Jan 2009 15:59:28 +0000]
Silence time64.c build warning
Silence the following build warning:
time64.c:169: warning: 'check_tm' defined but not used
Message-ID: <
1ff86f510901081219q4f5f25a3od89c4e5ac03e8664@mail.gmail.com>
Steve Hay [Tue, 13 Jan 2009 15:56:00 +0000]
Silence Win32 compiler warnings.
Nicholas Clark [Tue, 13 Jan 2009 15:32:38 +0000]
Give the git file browser URL as the most up to date version of this file.
Steve Hay [Tue, 13 Jan 2009 14:57:07 +0000]
Restore Win32 build by creating git_version.h.
This solution is based on what is done for config.h (which isn't perfect because it leaves miniperl.exe out-of-sync with the final perl.exe build, but that doesn't really matter).
Also has a problem in that miniperl.exe is used in a couple of places before make_patchnum.pl has been run (to recreate git_version.h and create lib/Config_git.pl), causing the warning:
Warning: failed to load Config_git.pl, something strange about this perl...
The running of make_patchnum.pl also causes some warnings of its own if a ".git" directory is found but "git" isn't on the PATH:
'git' is not recognized as an internal or external command,
operable program or batch file.
These things will be hopefully ironed out later.
John Malmberg [Tue, 13 Jan 2009 13:11:58 +0000]
vms - unixspec refactor
Message-id: <
496B5458.10203@gmail.com>
Refactor of unixspec() to not use a thread context for internal routines.
Also fix unixspec() to better handle unescaping extended file
specifications.
Abigail [Tue, 13 Jan 2009 12:36:20 +0000]
Capitalize "SysV" correctly
John Malmberg [Mon, 12 Jan 2009 04:15:28 +0000]
vms - vmsspec refactor
Message-id: <
496AC3E0.2090207@gmail.com>
Refactor of vmsspec() to not use a thread context for internal routines.
Nicholas Clark [Mon, 12 Jan 2009 22:20:34 +0000]
Fix my refactoring typo in change 33001 that broke suidperl, only spotted 11
months later, after 5.8.9 was released.
Renée Bäcker [Mon, 12 Jan 2009 10:39:11 +0000]
show -E in error message when called with -E
Currently, perl shows -e if an error occurs even it is called with -E.
The attached patch changes this...
Current behaviour:
D:\Perl510\bin>perl.exe -E "say ->"
syntax error at -e line 1, at EOF
Execution of -e aborted due to compilation errors.
With patch:
D:\Perl510\bin>perl.exe -E "say ->"
syntax error at -E line 1, at EOF
Execution of -E aborted due to compilation errors.
Cheers,
Renee
Nicholas Clark [Mon, 12 Jan 2009 17:10:31 +0000]
Add $(EXE_EXT) where it was missing from $(RUN) ./miniperl
Nicholas Clark [Mon, 12 Jan 2009 16:45:58 +0000]
Correct the over-zealous addition of $(RUN) into test -f ./miniperl$(EXE_EXT)
(in change
cef6ea9dde586d23f429802e75d22875450d6b68)
Nicholas Clark [Mon, 12 Jan 2009 15:01:10 +0000]
make_patchnum requires miniperl$(EXE_EXT) to run, so it needs to depend on it.
Also, make_patchnum had not been part of the prerequisites of the common
targets 'perl' or 'test', only 'all', so often it was not getting run. Add it
as a prerequisite of perl$(EXT_EXE), so that it is run on 'make perl' or
'make test'.
Florian Ragwitz [Mon, 12 Jan 2009 11:42:45 +0000]
Ignore ext/DB_File/constants.xs.
Abigail [Mon, 12 Jan 2009 11:14:58 +0000]
Fix typo in pod/perlhack.pod
Rafael Garcia-Suarez [Mon, 12 Jan 2009 08:29:28 +0000]
Merge commit 'nothingmuch/p4_git_utils' into blead
Yuval Kogman [Fri, 5 Sep 2008 18:46:58 +0000]
add git-make-p4-refs script to Porting
Rafael Garcia-Suarez [Fri, 9 Jan 2009 16:09:33 +0000]
Test that overload::constant works in eval("")
John Malmberg [Sun, 11 Jan 2009 04:20:59 +0000]
more vms thread ctx fixes
Message-id: <
496973AB.8070809@gmail.com>
If Perl_my_trnlnm is called with a null implicit context, it would
access violate.
create_mbx does not need a implicit context.
Nicholas Clark [Sun, 11 Jan 2009 17:06:04 +0000]
Add -Ilib to the invocations of make_ext.pl. I guess Steffen had a copy of perl
already installed whilst testing.
Nicholas Clark [Sun, 11 Jan 2009 16:45:52 +0000]
Remove make_patchnum.com from MANIFEST following
c24e7b67b597eaa15fbaca34ebb57a7cc242d937h
Steffen Mueller [Sun, 11 Jan 2009 16:31:56 +0000]
Merge branch 'blead' into miniperl-make-ext
Craig A. Berry [Sun, 11 Jan 2009 15:55:02 +0000]
Remove make_patchnum.com (now using make_patchnum.pl)
Steffen Mueller [Sun, 11 Jan 2009 14:48:21 +0000]
Merge branch 'blead' into miniperl-make-ext
John Malmberg [Sat, 10 Jan 2009 18:40:40 +0000]
File::Copy fixes for VMS follow-up
Message-id: <
4968EBA8.8090207@gmail.com>
The eval block needs to capture local $SIG{__DIE__} to prevent modules
with signal trapping from failing when VMS::Feature is not present.
John Malmberg [Sat, 10 Jan 2009 18:12:30 +0000]
vms setup_cmddsc buffer fixes.
Message-id: <
4968E50E.6050601@gmail.com>
Craig A. Berry [Sat, 10 Jan 2009 17:17:58 +0000]
Steffen Mueller [Sat, 10 Jan 2009 18:34:57 +0000]
miniperl-ification of ext/utils/make_ext
Abigail [Sat, 10 Jan 2009 17:47:14 +0000]
Fix typo in pod/perldsc.pod
Steffen Mueller [Sat, 10 Jan 2009 16:06:22 +0000]
move ext/utils/make_ext to ext/utils/make_ext.pl
Abigail [Sat, 10 Jan 2009 11:15:05 +0000]
Typo fix in pod/perldiag.pod
John Malmberg [Fri, 9 Jan 2009 20:20:21 +0000]
'VMS support for Unix and extended file specifications' (PathTools RT #42153)
John Malmberg [Fri, 9 Jan 2009 20:09:36 +0000]
Merge changes from PathTools: 'Update to support VMS in Unix compatible mode and/or file names using extended character sets' (PathTools RT #42154)
Abigail [Fri, 9 Jan 2009 19:01:06 +0000]
Be consistent in writing deobfuscator
Abigail [Fri, 9 Jan 2009 18:57:27 +0000]
Fix typo
Abigail [Fri, 9 Jan 2009 18:38:15 +0000]
Capitalize "latin"
Craig A. Berry [Fri, 9 Jan 2009 18:55:45 +0000]
fake_localtime_r and fake_gmtime_r may need thread context.
They definitely need it on VMS.
Craig A. Berry [Fri, 9 Jan 2009 18:51:31 +0000]
simple_trnlnm needs an actual length as its last argument.
Plus the length of sys$posix_root was miscalculated.
Plus an argument reversal when translating the TMP logical.
Abigail [Fri, 9 Jan 2009 18:01:10 +0000]
Typo fix in pod/perl5110delta.pod
Abigail [Fri, 9 Jan 2009 16:38:06 +0000]
Fix typo in pod/perltodo.pod
Abigail [Fri, 9 Jan 2009 16:15:13 +0000]
-Be consistent with kb/kB.
-Mention Abigail in last section.
Abigail [Fri, 9 Jan 2009 16:01:38 +0000]
Merge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead
Abigail [Fri, 9 Jan 2009 16:00:37 +0000]
Fixed some spelling errors
Vincent Pit [Fri, 9 Jan 2009 15:45:06 +0000]
Prefer tabs over spaces in MANIFEST
Abigail [Fri, 9 Jan 2009 13:09:24 +0000]
Change the link to the sort pragma to be more clear.
This fixes #62136, reported by rrt@sc3d.org.
John Malmberg [Thu, 8 Jan 2009 22:37:19 +0000]
vms.c backslash efs and long name fixes
Message-id: <
496552AF.7040702@gmail.com>
This fixes the support for the backslash in a file name as it shows up
in the ExtUtils tests.
Also the flag for rename to work on efs file specifications was
incorrectly the bit number instead of the mask value.
John Malmberg [Sun, 4 Jan 2009 19:06:54 +0000]
ext/File/Gob/t/basic.t, ext/POSIX/t/posix.t vms fixes
Message-id: <
496108CE.1060704@gmail.com>
The tests ext/File/Glob/t/basic.t and ext/POSIX/t/posix.t need to know
if VMS is in the UNIX compatible mode.
John Malmberg [Sun, 4 Jan 2009 19:13:45 +0000]
lib/blib.t, lib/h2xs.t patches for VMS
Message-id: <
49610A69.3020908@gmail.com>
The tests lib/blib.t and lib/h2xs.t need fixes to know when VMS is in
the UNIX report mode.
John Malmberg [Sun, 4 Jan 2009 18:45:11 +0000]
mkppport using catfile for catdir
Message-id: <
496103B7.1000308@gmail.com>
John Malmberg [Sun, 4 Jan 2009 18:58:03 +0000]
t/op/chdir.t, t/op/defins.t on VMS
Message-id: <
496106BB.9020102@gmail.com>
The test t/op/chdir.t needs to expect that returned dir could be in
either case or in VMS or UNIX formst.
The test t/op/defins.t needs to know of VMS is dropping dots on
filenames when VMS is in the UNIX REPORT mode.
John Malmberg [Sun, 4 Jan 2009 20:48:08 +0000]
t/x2p/s2p.t fix for VMS in UNIX mode
Message-id: <
49612088.4010005@gmail.com>
When VMS is in UNIX mode, conversions from UNIX specifications to VMS
can result in a trailing '.' which needs to be removed if the intent is
to default the extension.
Abigail [Thu, 8 Jan 2009 16:31:17 +0000]
Fixed spelling of 'uncondtional', as reported by Ronald J Kimball
in
20090108160007.GA85010@penkwe.pair.com.
John Malmberg [Sun, 4 Jan 2009 19:45:24 +0000]
Make File::Copy detect Unix compatibility mode on VMS.
Message-id: <
496111D4.8030007@gmail.com>
This is needed as part of Perl support for VMS in UNIX or using the
extended character set.
Patch amended to only check for the VMS::Feature module on VMS.
Rafael Garcia-Suarez [Wed, 7 Jan 2009 17:49:19 +0000]
Use a slightly more paranoid regex
David Golden [Wed, 7 Jan 2009 20:48:43 +0000]
sort MANIFEST
Steffen Mueller [Wed, 7 Jan 2009 14:38:16 +0000]
CPAN.pm on CPAN is considered upstream
John Malmberg [Wed, 7 Jan 2009 04:52:49 +0000]
VMS feature logical name fixes
John Malmberg [Wed, 7 Jan 2009 04:21:17 +0000]
VMS thread context fixes
Abigail [Tue, 6 Jan 2009 23:04:31 +0000]
Merge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead
David Golden [Tue, 6 Jan 2009 22:45:37 +0000]
Add David Golden (DAGOLDEN) to AUTHORS
Abigail [Tue, 6 Jan 2009 16:51:08 +0000]
Merge branch 'blead' of ssh://perl5.git.perl.org/gitroot/perl into blead
Abigail [Tue, 6 Jan 2009 16:49:16 +0000]
Updated the 'some statistics for the other subdirectories' to include
data for the 5.6.0 to 5.10.0 releases.
Abigail [Tue, 6 Jan 2009 16:13:03 +0000]
Avoid wrapping long lines when displaying as man page
Steffen Mueller [Tue, 6 Jan 2009 12:06:58 +0000]
Sync with Filter::Simple 0.84 which has a CPAN-relevant fix in the Makefile.PL only
Abigail [Tue, 6 Jan 2009 15:07:05 +0000]
Calculated patch sizes for the modern era.
Abigail [Tue, 6 Jan 2009 14:31:35 +0000]
Update data for 5.7.3 release
Abigail [Tue, 6 Jan 2009 13:45:46 +0000]
Remove duplicate entry
Abigail [Tue, 6 Jan 2009 12:03:31 +0000]
Added 21st century data to 'SELECTED RELEASE SIZES'
Rafael Garcia-Suarez [Mon, 5 Jan 2009 17:14:01 +0000]
Require a space or a newline after a "#line XXX" directive
This fixes bug [perl #59170] Typo: bad regex for #line directive in perlsyn.
Rafael Garcia-Suarez [Tue, 6 Jan 2009 08:59:35 +0000]
Add a --checkout / -c flag to git-find-p4-change
Remove the now obsolete switch_to_perforce_id.pl
Rafael Garcia-Suarez [Tue, 6 Jan 2009 08:49:02 +0000]
Remove a couple of p4-specific utilities
Rafael Garcia-Suarez [Tue, 6 Jan 2009 08:44:55 +0000]
Delete the largely obsolete Porting/patching.pod
(but move some good advice in there to the perlrepository man page)
Rafael Garcia-Suarez [Tue, 6 Jan 2009 08:11:58 +0000]
Use git instead of p4 for the --opened option of Porting/Maintainers
Rafael Garcia-Suarez [Tue, 6 Jan 2009 08:01:37 +0000]
Remove notes about -v from perltodo
Rafael Garcia-Suarez [Tue, 6 Jan 2009 08:25:03 +0000]
Move .bat gitignores with their unix counterparts
Remove also redundant ignores for produced files in x2p
Rafael Garcia-Suarez [Tue, 6 Jan 2009 08:20:19 +0000]
Merge t/.gitignore in main git ignore file
Craig A. Berry [Tue, 6 Jan 2009 04:36:55 +0000]
On VMS, run make_patchnum.pl, not make_patchnum.com
Max Maischein [Mon, 5 Jan 2009 13:22:54 +0000]
Add lib/CORE/ to .gitignore (needed on Windows)
C:\Projekte\bleadperl-git\win32>dmake && git status
[...]
nothing to commit (working directory clean)
C:\Projekte\bleadperl-git\win32>
Steffen Mueller [Mon, 5 Jan 2009 16:04:48 +0000]
Update AutoLoader to 5.68.
Fixes by John Malmberg for running tests when VMS is a UNIX compatible mode.
Rafael Garcia-Suarez [Mon, 5 Jan 2009 13:36:12 +0000]
Emit less PERL_ARGS_ASSERT_* macros
Rafael Garcia-Suarez [Mon, 5 Jan 2009 13:11:00 +0000]
"rm -f" better in "make clean" targets
It always succeeds and does not warn.
Rafael Garcia-Suarez [Mon, 5 Jan 2009 09:53:40 +0000]
Upgrade to Time::HiRes 1.9719
Rafael Garcia-Suarez [Mon, 5 Jan 2009 09:47:45 +0000]
Bump two module versions after Haiku port
CPANPLUS and Module::Build
(see
df00ff3beeb297b9622f8acbed9c80d320c87580)
Ingo Weinhold [Mon, 5 Jan 2009 09:29:56 +0000]
Adjustments to POSIX for the Haiku port
Message-Id: <
20081029022544.413.1@knochen-vm.localdomain>
I re-introduced the use of the WMUNGE() macro, which was
(accidentally?) removed after 5.10.0. The macro is still a hack. As my
added comment explains the use of the OS's W*() macros in this context
is simply not correct and should probably better be fixed.
Nicholas Clark [Mon, 5 Jan 2009 07:18:54 +0000]
Update copyright year to 2009.
Rafael Garcia-Suarez [Mon, 5 Jan 2009 07:12:02 +0000]
Correct vim modeline
- ts=4 is evil
- use same settings as in everywhere else in the core sources