Jarkko Hietaniemi [Sun, 28 Oct 2001 13:31:54 +0000 (13:31 +0000)]
[ 12733]
Use fd refcounts to solve the problem highlighted by new Embed.t, at
least for PERLIO=perlio unthreaded case
[ 12732]
Typo in exe name
[ 12730]
Fixup Embed.t for Win32/VC++
[ 12729]
Adapt testcase from Doug Maceachern <dougm@covalent.net> for perlio_destruct
bug as lib/ExtUtils/Embed.t which tests that and some aspects of
ExtUtils::Embed.
[ 12711]
Use ref count scheme rather than PerlLIO_dup() to do fp_dup().
Restores op/fork.t on Win32 (still segfault on exit of ok 2).
[ 12557]
All tests pass (legitimately) on ithreads
[ 12555]
Restore non-ithreads build. Interestingly binmode test passes
non-ithreads - so crlf layer is not damaged and ithreads fail
is a symptom ...
[ 12552]
Win32 PerlIO_cleanup special cases should no longer be required.
[ 12551]
Multiple win32io.o is fatal for GCC (and noise with VC++)
[ 12547]
Convert rest of PerlIO's memory tables to per-interp and add clone functions
for them. Call explicit cleanup during destruct process.
- one binmode test is failing
- also ext/threads/t/basic.t fails under make test, and is noisy under
harness. (Threads results are intermingled and don't match order expected.)
[ 12544]
PerlIO layer table as PL_perlio (per-interpreter)
[ 12541]
Fix typos in new locking on MemShared
Make buffers in PerlIOBuf_* in per-thread heap (they are cloned after all...)
- seems to make segfaults more deterministic - suspect they are
attempt to free() after Interp and Host have gone.
projects / p5sagit/p5-mst-13.2.git / commitdiff