develooper Front page | perl.perl5.porters | Postings from April 2016

Perl 5.22.2-RC1 on Sunday 10th

Thread Next
From:
Steve Hay via perl5-porters
Date:
April 7, 2016 16:31
Subject:
Perl 5.22.2-RC1 on Sunday 10th
Message ID:
CADED=K4XmHrcGw7AxSG+0Z=Q40coEoyK20LUSmpjbSaFdXGxHg@mail.gmail.com
I'm planning on releasing 5.22.2-RC1 on Sunday 10th. Hopefully the
release candidate phase will be smoother than was the case for 5.22.1
and I can get the release out before the end of the month, thus not
overlapping much with the release candidate phase of 5.24.0.

(Ideally I should have got 5.22.2 out sooner to avoid any overlap;
apologies again for being a little late on this.)

The voting file is updated with the latest blead changes (currently
just a Cygwin test fix and documentation relating to it). I won't be
looking to pull anything else from blead after RC1 is out.

The maint-5.22 branch itself is nearly fit for RC1 apart from some
further perldelta work that I have to do, so if anyone feels like
getting a head-start on the testing then feel free :-)

I would like to take this opportunity to double-check one thing that
has come to my attention in reviewing the changes made so far:

Commit 971809a ([perl #126410] keep the DESTROY cache in mro_meta)
added a new member to the mro_meta struct in hv.h. As is too often the
case it isn't entirely clear whether this is a "public API" that has
been changed. It is mentioned in perlmroapi.pod, but only as a "black
box", with no details of its actual structure given. The only other
documentation that mentions it is the perldeltas for 5.10.1 and
5.12.0, which say it is a private structure, so we should be ok in
that regard.

However, is the change safe from a binary-compatibility point of view?
The structure is now larger than it was, and the new member isn't even
added to the end of it so the member after it is now at a different
offset.

I guess the answer is a question of exactly how/where the struct is
used, but unless we're confident that it's ok then I think it would be
safer to revert that commit from the maint branch, especially since
users won't have long to wait for it in 5.24.0 anyway.

Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About