Re: Remaining Configuration & Build bugs

[Rik]

On 10/27/2013 09:07 AM, address@hidden wrote:
> Some questions: Do we need to fix MSVC issues now?
No.
>  Whats the procedure for
> closing out bugs from old versions of octave?
I think the best thing will be to produce the 3.8 release, and then go back
and tag the issues that we think are fixed with a request to the original
reporter to verify the issue with the new release.  After a certain amount
of time we can assume that any non-response means the bug was actually
fixed.  If it isn't, and they just didn't respond, they can always re-file
a new bug report showing the behavior reproduced on the newest release.
>  Do we care about feature
> requests for this release?
No, that phase has passed.  Also, the 3.8 release is intended to be a
rather short-lived release.  It shouldn't be the case that we need to go
nearly 2 years before the 4.0 release which means that any outstanding
feature requests implemented on the development branch won't languish for
too long.

>
> A quick look through the list for me (looking only at the Windows issues):
> 40321 (mkoctfile) - Now fixed (in test)
> 40273 (SED changes): complete - left open for upsteam changes to gnulib
> 40252 (Qt5): Not going to change yet ?
> 39856 (Compiling Oct files in Win7): Issues with MSVC and octave 3.6.4
> 39465 (Mingw Java support in configure):
> 39446 (Native build of octave with MSVC installed): Effects oly those who
> want to create their own native build octave in Windows.
> 38351 (mkoctfile flags): Works in mingw built version for me, and was
> written against 3.6.2.
> 34301, 34169 (-f flag in mkoctfile): Feature requests.
>  
I think the support for unknown compilation options, like -f, should be
done.  There is a patch attached to #34169, but it isn't the way we want to
go.  Jordi suggested in that report the correct strategy.

--Rik