develooper Front page | perl.perl5.porters | Postings from May 2010

Re: 5.12.1 prep

Thread Previous | Thread Next
Craig A. Berry
May 7, 2010 06:29
Re: 5.12.1 prep
Message ID:
On Thu, May 6, 2010 at 11:28 PM, karl williamson
<> wrote:
> Jesse Vincent wrote:
>> I'm planning on cutting an RC0 on Friday. What do folks want merged that
>> hasn't yet been merged down?
>> -jesse
> I'd like to lobby for some patches to go into 5.12.1.
> First, I'm surprised that no one mentioned
>  7dcb3b25fc4113f0eeb68d0d3c47ccedd5ff3f2a
> which fixes [perl #72998] regex looping.  I thought a dot release would
> include all available sane fixes for perl crashing or hanging.

This one has been marked as rejected.  I don't know why or by whom --
that's an area where the selection process has started to see some
improvements but needs more.  I'm not against the patch, but I don't
understand it well enough to be for it either.  But you're right to
bring it up again if it fixes a hang.

> Second, I don't understand the reluctance to include all doc patches that
> are applicable.

I think the second part of the sentence answers the first.
Determining whether they are applicable takes time and attention.  The
very first statement in Jesse's maint-5.12.x policy was, "New releases
of maint should contain as few changes as possible."[1]  So the
default setting is that not much should happen, and I suspect we've
actually done a bit more than Jesse expected us to.  For a change to
go in, it needs specific advocacy and approval.  Doc patches are
acceptable but not guaranteed, and something not going in is not an
indication of reluctance.

> I can't speak about changes that others have made in particular, but it
> seems to me that the impetus should be that doc patches should go in a dot
> release unless otherwise contraindicated.

The maint releases are now based on a schedule, not on a particular
category of changes.  I think security patches and fixes for
regressions are about the only things that *might* make the schedule
take a back seat.  For everything else, it's just a matter of whether
it's made it through the process by the time the train leaves the
station.  The combination of time-bound releases and the absence of a
maint pumpking really means it can't be any other way.

We are doing a lot of doc patches for 5.12.1, but, as you point out,
not everything.  I think a lot of this is the luck of the draw.  If
someone with a particular passion for improving core docs had been
available to put in a lot of time the last couple of weeks, there
might be more doc patches in.  Then again, there might be fewer
because there is no guarantee such a person would accept as many as
the rest of us have.

Keep the Unicode patches coming.  I'm sorry we can't assimilate them
into production releases any faster.

[1] <>

Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About