Re: 5.20.1 status update #3

On 18 August 2014 23:05, Karl Williamson <pub...@khwilliamson.com> wrote:

> On 08/18/2014 01:48 AM, Steve Hay wrote:>>> Currently remaining:>>>> Needing one vote:>> 1d8d63d Remove compiler warning spam on dos-djgpp [+1 hmb]>>>> +1


Thanks, now cherry-picked.


>>>>> Needing two votes:>> version-0.9909, which contains f57000b PATCH: [perl #121930] Bleadperl>> breaks MDOOTSON/Wx>>>> The version.pm update fixes #121930 except that there is an>> outstanding question over what to do about adding a new XS-callable>> function (747c467, see>> ../perl5-porters/210612/) and I believe>> Karl is also kindly writing some documentation describing the issues>> raised by this bug.>>>> I would argue for bending perlpolicy far enough to accomodate adding>> the new XS-callable function on the grounds that it's a part of the>> fix for the regression issue in question ;-) Votes?...>>>> All this is dependent on there being votes for the first fix.  So let me> lobby for why something should go in.>>
[...]
The argument sounds good to me, but we still need votes if this is to go
in.


As I was writing the documentation to give advice to XS writers about how
> to deal with locales, I realized that having this function in 5.20 would> allow people to start using it now, without having to wait until 5.22.> 5.20 is really the first release that has relatively sane locale handling> that would benefit from the use of sync_locale().  Since it does create> compatibility issues, how about making it a macro in maint v5.20?  It is> just a wrapper around calls to several existing functions.  In 5.22 it> becomes a function.>

The macro suggestion sounds like a good way to avoid any difficulties with
adding a new function, yet still without the hassle of ppport.h changes.
Please could you make a patch against maint-5.20 showing what this would
look like so that we can hopefully get votes on that too?



>> The bottom line is that I think something needs to be done.  If the> proposed patch is defective in some way, tell me what needs to be done to> fix it or allay your concerns, so that we can get something in.>

The ony objection that I've heard so far was that the patch touched vutil.c
so perhaps wasn't appropriate, but that concern has been removed by the
release of a new version.pm to CPAN: Updating CPAN modules in core is
subject to the same conditions as any other changes proposed for inclusion,
so that aspect is no longer an issue.

There has been plenty of time now for any other objections to be raised, so
given that none have been forthcoming can we get some votes to get this in
now?