Vincent Pit [Tue, 6 Aug 2013 03:26:38 +0000 (00:26 -0300)]
Get rid of t/99-kwalitee.t
New versions started to throw warnings if specific environment variables
are not set. This may be appropriate for dist manager users, but not really
for the rest of the world.
Since kwalitee metrics are either obvious or useless, there's no big loss
in removing this test anyway.
Vincent Pit [Fri, 2 Nov 2012 00:27:25 +0000 (22:27 -0200)]
Don't run the POD spelling test in taint mode
This causes the test to crash on perl 5.12 and below with sensitive
architectures like FreeBSD. Ultimately the root of the issue seems to have
been https://rt.perl.org/rt3/Public/Bug/Display.html?id=60378.
Vincent Pit [Fri, 11 Nov 2011 18:41:17 +0000 (19:41 +0100)]
Threads tests may not be able to spawn all the threads
To work around this :
- move all the threads boilerplate to a new helper test module ;
- capture the warnings/errors that threads->create may throw and reprint
them properly as diagnostics ;
- use Test::Leaner instead of Test::More, as older Test::More sometimes
cause out of sequence bugs ;
- last but not least, do not hardcode the plan and use done_testing with
the actual number of spawned threads instead.
Vincent Pit [Tue, 4 Jan 2011 18:10:07 +0000 (19:10 +0100)]
Always skip seen ops in our peep replacement
Sometimes, there are loops in the next chain as e.g. with infinite loops
like "for (;;) { ... }". We can't reuse the op_opt member because it's
almost always set after the first pass.
Vincent Pit [Fri, 31 Dec 2010 14:05:15 +0000 (15:05 +0100)]
Fix rare segfaults and bugs with threads
There was a race in the "PL_ppaddr[OP_PADSV] save/restore dance" in that
perl might be compiling ops inside and outside the scope of the pragma at
the same time. We solve this by setting the op_ppaddr member of PADSV ops
to a_pp_deref during the peephole optimization phase.
Vincent Pit [Wed, 29 Dec 2010 23:25:59 +0000 (00:25 +0100)]
Force linking against the perl dll when using gcc 3.4 on Windows
When a Windows perl links an XS shared object to an import library using
gcc/g++ version 3.4.x, strange breakage happens at load time because the
mutex-related symbols seems not to be reachable. This situation happens
especially for AS perl and the default compiler suite it installs.
We fix this by forcing g++ to link against the dll instead, which should
be in the same directory as the perl executable (at least for AS perl).
This change is the integration of commit 14f8af2 from Variable::Magic's
repository.
Vincent Pit [Sun, 3 Oct 2010 00:31:48 +0000 (02:31 +0200)]
Skip threads tests unless perl version is 5.13.4 or greater
There was a long standing bug in the handling of GV <-> CV double linkage
that could (and explicitely did with a poisonous perl) cause segfaults at
thread destruction. It got fixed by Dave in commit 803f274 which went in
5.13.3, but the fix was amended for 5.13.4 in commit 09aad8f. Since it's
not really fair for the user to not be able to install the module because
of this, we skip the threads tests unless perl is at least 5.13.4.
Vincent Pit [Tue, 1 Jun 2010 16:20:35 +0000 (18:20 +0200)]
Improve the thread destructor trick
It's currently only used before 5.10.1, so it wasn't really broken by the
recent changes to blead. But the new technique is cleaner and easier to
maintain.
Vincent Pit [Thu, 18 Jun 2009 17:54:35 +0000 (19:54 +0200)]
Classify aliasing constructs in the "store" category
There's no way to guess at compile time if the scalar will be modified, such as in "eval '$_ = 1' for $x->{key}". Practically, there's no way to distinguish the xelem op from the one you get for "$x->{key} = $val". I was going against the interpreter's will and this is my defeat.