Re: Raw std::auto_ptr use

On Mon, 28 Jun 2010 19:46:20 +0200, Kinkie <gkinkie_at_gmail.com> wrote:
> Hi Alex,
>
>> How should we proceed?
>
> We do have a mean to test on non-gcc compilers, namely the build farm.
> SunStudio is quite picky for instance.
> So what I'd suggest we do is to fix the brokenness recently creeped in
> for non-gcc builds, and then commit to trunk and see what happens. If
> it builds and works, we may have a winner. If not, we can revert and
> reassess our options.
>
> Do you think this might work?

We also have a means of globally protecting any broken OS from desired
code (compat/os/*.h definitions).

So, I'm for #1 in trunk to see what happens in Hudson.

 * I hope to be done with the NTLM helper breakages in a day of two.
 * Kinkie; the ongoing Solaris breakage is SASL library detection. Started
when you added config.test extensions to locate the libsasl2. So configure
says it exists to build the helper, but make can't find it.
  - the config.test checks need to be mirrored in configure.in to set LIBS
properly.

Re Boost; We now have a few solid things which could benefit. It may be
worth putting it on the RoadMap for 3.3.

PS. Per the testers request for support in 3.1.5; since 3.1.5 is due out
this coming week I'd say no, but maybe a chance for 3.1.6 after these
portability tests are completed.

Amos
Received on Mon Jun 28 2010 - 23:16:36 MDT