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

Re: revert MG consting (Coro breakage) for 5.24?

Thread Previous | Thread Next
Dave Mitchell
May 4, 2016 08:04
Re: revert MG consting (Coro breakage) for 5.24?
Message ID:
On Tue, May 03, 2016 at 06:09:52PM -0400, Ricardo Signes wrote:
> * Dave Mitchell <> [2016-05-03T11:56:31]
> > I propose that we revert this commit for 5.24 (we need another RC anyway
> > for the SvGROW) issue.
> I have reverted this in a local branch.
> It is not clear to me that there is anything much to gain, here.  As per my
> message of a few minutes ago, Coro doesn't seem to need this change, if code is
> (conditionally?) removed from Coro.  Why did no one see this earlier?

I had always vaguely assumed that Coro had been patched up to work with
5.22 - I knew that there were efforts to so do - and it was only when
someone pointed out to me Aristotle's blog entry from yesterday that
I became aware that it (may) still be an issue.

I don't know enough about the technical issues with Coro to understand
whether the vtable consting is actually a constraint on getting Coro to
run under 5.24.

> This problem has generated much heat, but little light, and now there's an 11th
> hour rush to delay a release to make a possibly unneeded change that could've
> been made months ago.  Will we ever find out whether this change is really
> needed?  I have no idea.  I imagine, though, that no matter what, there are
> going to be hard feelings on *some* part for making *or* not making this
> change.
> Of course, if that's lose-lose, what's to be lost by applying this patch?
> Probably nothing.

I proposed making the change at this very late stage because there is
negligible risk of a downside, there's a probable upside, and a new RC has
to be released anyway for the SvGROW stuff.

I before E. Except when it isn't.

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