| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
| |
Note for future:
Looks like there is an x_serialization now, so if CPAN::Meta::YAML
and/or JSON::PP get updated then these META files will need regenerating
too.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
[DELTA]
2.150005 2015-06-09 19:08:44-06:00 America/Denver
[TESTING]
- Changed some test data from UTF-8 to ASCII
2.150004 2015-05-19 11:25:53-04:00 America/New_York (TRIAL RELEASE)
[DOCUMENTED]
- Noted explicitly that historical META spec files are licensed under
the same terms as Perl
[TESTING]
- Added test for 'x_deprecated' field in "provides"
[META]
- declared extra developer prereq
2.150003 2015-04-21 19:41:15-04:00 America/New_York (TRIAL RELEASE)
[CHANGED]
- Serialized CPAN::Meta objects now include a x_serialization_backend
entry
2.150002 2015-04-19 01:00:10+02:00 Europe/Berlin (TRIAL RELEASE)
[CHANGED]
- Metadata merging now does deep hash merging as long as keys
don't conflict
|
| |
|
|
|
|
|
| |
Commit 22ff313068 for [perl #123814] inadvertently changed the logic when
parsing a numeric parameter to the -C option, such that the successfully
parsed number was not saved as the option value if it parsed to the end
of the argument.
|
| |
|
|
|
|
|
|
|
| |
-previous usage proves getmagic is unnecessary or ignored or already called
S_do_smartmatch
Perl_sv_does_sv
-in pp_dbmopen dont call sv_isobject twice in a row on the same SV in 1
permutation
|
| |
|
|
|
|
|
| |
5.22.0 is out; development is now 5.23, leading towards 5.24.0 next May.
5.22.1 is expected sometime around August/September (around the same time
as 5.20.1 was last year).
5.20.3 is expected sometime before then, probably July.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
truncate($filename, $size) was using a simple PerlIO_open() to open
the file, which on Win32 defaults to a text mode open.
Unfortunately, on a text mode open(), MSVCRT attempts to seek to the
end of file using only 32-bit offsets, which fails.
For good measure, add in O_LARGEFILE if it's available, which may
prevent similar issues on other platforms.
Also, remove the erroneous SETERRNO() added by 375ed12a to the open
failure branch, PerlLIO_open() should already set errno on failure, so
we get sane error messages when the open fails.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously the aslice flags would end up as 255 in this case:
$ perl -MO=Concise,-debug -e '@nums[0..1] ~~ []' | grep OP_ASLICE -A 2
Smartmatch is experimental at -e line 1.
-e syntax OK
op_ppaddr PL_ppaddr[OP_ASLICE]
op_type 131
op_flags 255
Now:
$ perl -MO=Concise,-debug -e '@nums[0..1] ~~ []' | grep OP_ASLICE -A 2
Smartmatch is experimental at -e line 1.
-e syntax OK
op_ppaddr PL_ppaddr[OP_ASLICE]
op_type 139
op_flags 55
|
| |
|
|
|
|
| |
This issue was introduced in commit 269713a1
Ref https://rt.perl.org/Ticket/Display.html?id=125217
|
| |
|
|
|
|
| |
This ensures that authors like Ævar Arnfjörð Bjarmason sort with the As
in perldelta ACKNOWLEDGEMENTS sections. Previously Ævar appeared at the
end.
|
| | |
|
| | |
|
| |
|
|
| |
(observed at least in AIX 6.1)
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
INSTALL_BASE for example would cause modules to be installed in the
wrong place during a bisect.
|
| |
|
|
| |
Thanks to Andreas for the examples.
|
| |
|
|
|
|
|
| |
If you bisect with --module or --with-module, older perls need a
CPAN::MyConfig, otherwise they will stop to ask for user input. This
commit adds the ability to automatically create one if necessary (with
some user interaction required.)
|
| | |
|
| |
|
|
|
|
|
|
|
| |
For example:
.../Porting/bisect.pl --with-module=Moose -e 'use Moose; 1;'
This is different from --module in that --module only tests that a
cpan module can or cannot install properly (and pass its tests).
|
| |
|
|
|
|
|
|
|
|
| |
Original can be seen in bin/makeperl.pl of http://repo.or.cz/r/andk-cpan-tools.git
Now we can:
.../Porting/bisect.pl --module=autobox
And see when it broke!
|
| |
|
|
|
|
|
|
|
|
|
| |
The PadlistNAMES() macro was introduced with v5.17.3-69-g86d2498.
This macro happened to be lvalue-capable, although it wasn't documented
as such.
v5.21.6-163-g9b7476d as a side effect, broke the lvalueness, which broke
Coro.
This commit restores the lvalueness.
|
| | |
|
| |
|
|
| |
(the comments are whitespace and eaten away (yum) first)
|
| |\ |
|
| | | |
|
| |/ |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
If the added script changes, it may affect the output of mk_invlists.pl,
so needs to be added as a dependency.
|
| |
|
|
|
| |
If either of these two scripts change, it could affect this script, so
things need to be regenerated.
|
| |
|
|
|
|
|
| |
It turns out that the rational numbers furnished by the Unicode data are
not guaranteed to be in lowest terms. Version 8 is the first version
that has them this way. Add code to reduce them in preparation for that
release.
|
| | |
|
| | |
|
| |
|
|
|
|
| |
I mistakenly thought I saw corpus/ being missing from Perl, so IGNORED it
in Maintainers.pl, but actually the problem was simply that one file was
missing (it's new) -- so add it now.
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
