[Pkg-d-devel] Attempt to build BioD blocked by undeaD and missing module string (Was: How to build D source)

[Matthias Klumpp]

2017-03-03 14:40 GMT+01:00 Andreas Tille <tille at debian.org>:
> Hi,
>
> On Fri, Mar 03, 2017 at 10:53:03AM +0100, Matthias Klumpp wrote:
>> > I have no idea what might be wrong here.  In testing we have ldc
>> > 1.1.0+b6-1 and unstable has 1.1.0+r-1 - may be there is some difference.
>> > I've attached the full build log.
>>
>> Your Meson version is too old, make sure you have one >> 0.38 -
>> previous versions had a bug making them treat warnings as fatal when
>> vompiling with LDC.
>
> Hmmm, so are we releasing a broken version with Squeeze?  I admit I
> finally need to backport the packages to Jessie (at least in the
> foreseable future).  So having a working version of meson in Squeeze
> (=testing) where to backport from would be really convenient.  Is this a
> good reason to file a bug severity grave and is there any chance to
> backport the "treat warnings as non-fatal" feature to 0.37.1?

I would love to have 0.38 in Stretch, but even without the full
version, the patch is pretty trivial. Grave severity is probably
warranted, since there is absolutely no way to disable this behavior.
For D specifically, the following patches would be neat:

SONAME fix: https://github.com/mesonbuild/meson/commit/c6fb5aec58af87fcb26943bae57b4ba2abd24b74
No linker flag breakage:
https://github.com/mesonbuild/meson/commit/a9c09e6004ff1b5344782e698549e9fba041fca9
Proper warning flags (!):
https://github.com/mesonbuild/meson/commit/fc72aa1e65cbe6b36af16c480b8a892bdef7638a
GDC makedeps: https://github.com/mesonbuild/meson/commit/8ee1c9a07a3a35e3ed262fbc358fd86c257a966e

On a second look, it actually might make sense to just have all of
0.38 in Stretch... But who knows what the release team would like
better - in any case, one can always backport a more recent meson to
Stretch.

Cheers,
    Matthias

-- 
I welcome VSRE emails. See http://vsre.info/