Front page | perl.perl5.porters |
Postings from August 2011
Re: Perl 5.14 is dead
Thread Previous
|
Thread Next
From:
David Golden
Date:
August 7, 2011 19:58
Subject:
Re: Perl 5.14 is dead
Message ID:
CAOeq1c9G6ipa1qm3RC0x4rsDt1rXoiDPvm1FwWu1q7FopO654g@mail.gmail.com
On Sun, Aug 7, 2011 at 4:52 PM, Father Chrysostomos <sprout@cpan.org> wrote:
> In addition, this new commit fixes a regression from 5.12:
>
> f8b0429 [perl #95748] IPC::Open3::open3(..., '-') broken
+1
>> Except for 043d1e6, these are all regressions of one sort or another.
>>
>> c533395 [perl #90160] U* gives ‘U0 mode on an empty string’
>> b2b7346 2nd try: [perl #91834] utf8::decode does not respect copy-on-write
>> a0aa607 Allow lvalue subs to return COWs in list context
>> 043d1e6 move "garbage collection" glossary entry where it belongs
>> a130272 Allow lvalue subs to return COWs in reference context
>> e5accad Allow COW values to be deleted from restricted hashes
>> fb2352e Allow restricted hashes containing COWs to be cleared
>> b2b95e4 Stop localised ties from becoming ro when COW
>> 47d6f3d Get PerlIO::scalar to write to COWs
>> 4ab5bd5 Make Devel::Peek::fill_mstats work on COWs
>> 71edc89 [perl #92258] <$fh> hangs on a glob copy
>> f5d1ed1 Fix my + attrs + list assignment
>> 4dbb339 Allow ‘sub x :lvalue’ to apply to XSUBs and stubs
>> 21690b7 Make it possible to have read-only glob copies
My lack of response to these previously is due to many of them
sounding like new capabilities. I don't understand what they are a
regression against. Did all of these things work in some previous
Perl?
Otherwise, only 71edc89 might rise to the "crashing bugs" standard for
backporting. How many of the rest are prerequisites for fixing
71edc89?
>> These two go together, as the first one only partially fixed the bug:
>>
>> d34a666 [perl #91880] $_ refcounting problems in @INC filters
>> 339c6c6 Follow-up to d34a6664
+1
>> These four need some explanation:
>>
>> a991bd3 Make Storable freeze COWs properly
>> d5c6c69 Fix a Storable test to work in 5.8.9-
>> fa819c1 Make SvIsCOW honest about globs
>> 3e89ba1 Make Internals::SvREADONLY smarter
>>
>> a991bd3 includes a test, but that test triggers the bug fixed by 3e89ba1. So if both are backported, either the tests from a991bd3 should be left out, or d5c6c69 should be backported as well. 3e89ba1 depends on fa819c1.
Again, are these regressions from a prior Perl? The policy on
dual-life modules backports is stated as "fix crashing or security
issues only". Do these meet that standard? Is there anything that
precludes a CPAN-only release from blead that could fix this problem
for 5.14.1 and prior?
-- David
Thread Previous
|
Thread Next