Front page | perl.perl5.porters |
Postings from May 2002
May 29, 2002 17:13
Message ID: MqW98gzkgmaP092yn@efn.org
>On Wed, May 29, 2002 at 03:27:18PM -0700, Yitzchak Scott-Thoennes wrote:
>> In article <20020529175206.B30595@alpha.hut.fi>,
>> Jarkko Hietaniemi <firstname.lastname@example.org> wrote:
>> > http://email@example.com
>> > http://firstname.lastname@example.org
>> > ftp://email@example.com
>> > ftp://firstname.lastname@example.org
>> > ftp.funet.fi::perlsnap
>> >Reminder: 5.8.0 RC1 Saturday the 1st of June.
>> >- MM 6 syncup [Michael]
>> >- Test::Harness syncup [Michael]
>> >- VMS PERLIO=perlio [Craig, NI-S]
>> >- VMS socket trouble README.vms addition [Peter]
>> >- Math::Big*, big* bugs [Tels] (ETA Thursday)
>> >- ithreads XS docs [Arthur]
>> >- sporadic lib/Memoize/t/expfile.t failures [mjd]
>> You've let self-ties fall off the list!
>I lost faith after both you and NI-S seemed rather, err, lost.
"All that is gold does not glitter,
not all those who wander are lost"
>Remember, at some point WE MUST STOP PATCHING.
>> correctly croaks on aggregate self-ties
>> restores sv_magic call to 5.6.1 state to fix scalar/IO self-ties
>> restores the tie.t test to before the Great Retraction.
>> (plus spelling bee, plus TODO failing test).
>> Only run through minitest so far.
>> Left to do: fix self-tied globs :(
>> My current thought on this is to 1) move the tie magic back onto the GV
>> or 2) require self-tied globs to use weakrefs, croaking if they don't.
>If I don't have working tested patches by Thursday evening, no go.
>(I said earlier Friday evening, but NI-S talked some sense to me.)
>Core *.c patches must be in the Friday morning snapshot.
Actually, things should be pretty good now, except that self-tied
globs won't get destroyed when they should. If that bothers anyone,
we can always say to use self-tied IO's instead.