| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
| |
In spite of there being man pages for these, the #include file doesn't
define the mbstate_t type which is required for a parameter to these
functions.
Perhaps the Configure probe could be enhanced so it doesn't return
defined unless these can be successfully compiled, but for now use the
hints file.
|
| |
|
|
|
|
|
|
|
| |
On Darwin 15.6.0, mkostemp() was observed to be autodetected as present
but actually be unlinkable. It is unknown what other Darwin versions
are affected, so for the time being just override the autodetection on
all versions.
|
|
|
|
|
|
| |
FreeBSD's /usr/share/mk/sys.mk specifies -O2 for architectures other
than arm and mips. By default, compile perl with the same optimization
levels.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
More precisely, tailor the application of the workaround to the FreeBSD
userland version ranges specified in
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211743#c10. (Using
userland version ('uname -U') rather than kernel version ('uname -K') at
suggestion of Mathieu Arnold.
We probably don't need an 'else' branch; let Configure handle it.
|
|
|
|
|
|
|
|
|
|
|
|
| |
NetBSD 6 provides renameat() etc in it's libc, but in the cases where
we use them they fail with ENOSYS.
So I've modified the in-place edit clean up code to attempt to
fallback to the non-at versions of these functions, after checking
that the current directory is sane.
Once I was sure that worked, since the *at() functions don't work for
my use case on NetBSD 6, I've disabled them in hints.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I must have left the test in the wrong position while
working on getting it to work, and never tested it since.
My apologies.
With this simple trick the following failing tests start to work.
../cpan/Math-Complex/t/Complex.t
../cpan/Math-Complex/t/Trig.t
../cpan/Math-Complex/t/underbar.t
../cpan/Scalar-List-Utils/t/uniq.t
../ext/POSIX/t/math.t
../lib/warnings.t
op/sprintf2.t
op/time.t
Or rather, the broken part of the functionality, Infinity,
starts being avoided, as it should be. See the comment above
about the known brokenness of long doubles and infinity in AIX.
Note that even after this fix the following are still broken in
this AIX (the perl5 aix machine), but these are unrelated to infnan:
../cpan/ExtUtils-MakeMaker/t/03-xsstatic.t
../dist/Storable/t/store.t
../ext/XS-APItest/t/handy00.t
re/uniprops01.t
re/uniprops02.t
re/uniprops03.t
re/uniprops04.t
re/uniprops05.t
re/uniprops06.t
re/uniprops07.t
re/uniprops08.t
re/uniprops09.t
re/uniprops10.t
(as of blead 4faa3060)
|
|
|
|
| |
We rely on a C89 implementation anyway.
|
|
|
|
|
|
|
|
| |
The Configure changes here were generated using a version of metaconfig
that copies U/modified/i_stdlib.U from dist/U/vaproto.U, and changes it to
unconditionally define the i_stdlib Configure variable. That variable is
used by a large number of other Configure units, so it's not actually
practical to try and remove the relevant unit entirely.
|
| |
|
| |
|
|
|
|
|
|
|
| |
The Configure changes here were generated using a version of metaconfig
that makes U/perl/perlxv.U assume that the keyword exists, and prevents
U/modified/d_volatile.U from promising to define a "volatile" keyword;
otherwise, those units would bring in the relevant Configure probe anyway.
|
|
|
|
|
|
|
| |
"Sane" means that it works correctly on bytes with their high bit set, as
C89 also requires.
We therefore no longer need to probe for and/or use BSD bcmp().
|
|
|
|
|
| |
It is almost impossible to find any information about this platform on the
internet, which strongly suggests that it's as dead as dead can be.
|
|
|
|
| |
We can therefore also avoid probing for and/or using BSD bcopy().
|
|
|
|
| |
This means we also never need to consider using BSD bzero().
|
| |
|
| |
|
|
|
|
|
| |
note that we cannot use g*, as AIX/vac ships gxlc as xlc with gcc-like
option handling
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The lack of this caused several test failures on cygwin64, the one case
I tracked down involved memmem() which is a GNU extension that cygwin
supports.
Since the compiler couldn't see the memmem() prototype it treated it's
return value as int, which was then cast to (char *) preventing any
type-mismatch warning, but since int is 32-bits and (char *) on
cygwin64, the upper 32-bits of the pointer was cleared, resulting in a
crash.
After adding this a test cygwin64 build went from 30 or so test failures
to one.
|
|
|
|
| |
Including the various pieces of Module::CoreList.
|
|
|
|
|
|
|
|
| |
Note that at this stage, porting tests fail:
Failed 3 tests out of 2390, 99.87% okay.
../dist/Module-CoreList/t/is_core.t
../dist/Module-CoreList/t/maintainer.t
porting/corelist.t
|
|
|
|
|
|
| |
The Configure scan fails to find dlopen() with g++. Explicitly making
it availble allows Configure to default to using dynamic loading, but
still allows the user to override and use static loading.
|
|
|
|
|
| |
For: https://rt.perl.org/Ticket/Display.html?id=131337
Signed-off-by: James E Keenan <jkeenan@cpan.org>
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
usedtrace builds add references to libelf symbols, causing link
failures without it.
at hints time we don't know if the user will interactively select
dtrace and there's no CBU, so it's added unconditionally on 10.x
|
| |
|
| |
|
|
|
|
|
| |
but instead require an extra option -Ddarwin_distribution to produce
the same results.
|
|
|
|
| |
Linux has used ELF for over twenty years.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
(a) This enables the malloc wrap also on Hurd, as used on Linux
already; Perl's own test suite passed with it, and I see no reason to
diverge from the behaviour on Linux and kFreeBSD. I'm not sure whether
it affects the ABI though, so it might be safe only for the upcoming
perl 5.24 in experimental (since that breaks the ABI anyway)
(b) This improves the reporting of the GNU libc used, so it's shown in
`perl -V` (as libc value, instead of the currently empty string).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The main failure appears to be a bug in freebsd. Jim Keenan and I have
created a stand-alone C program, not involving Perl, that reproduces it,
which I will attach to the ticket. I have searched their bug db and not
found this reported, so will create a ticket against them.
Several of the failures are bugs in some of the locale definitions for
freebsd, like not all lowercase letters also being alphas. I will
report these as well, and adjust the allowable failure percentage for
this platform, if necessary, to get these to not fail the test at large.
The bug is that newlocale() and/or uselocale() are not working properly.
These are from POSIX 2008, and perl has not used them previously.
I sort of expected some platforms to have not implemented them properly;
this is the first one we've encountered that does so.
This changes the hints file so that it appears that uselocale() is not
on the system.
|
|
|
|
| |
(including regen/opcode.pl)
|