[Python-apps-team] Bug#682652: Bug#682652: Bug#682652: cython: Cython might cause problematic code when trying to build using gcc-4.7

[Yaroslav Halchenko]

Andreas Tille <tille at debian.org> wrote:

>On Fri, Jul 27, 2012 at 03:43:57PM -0400, Yaroslav Halchenko wrote:
>> not after the release I guess...
>
>Not *after* means *before* the release. ;-))
>OK, understood (and expected) the motivation.
>
>> if I were you I would have just
>> pregenerated .c files and shipped those for the systems which lack up
>to
>> date cython (ftpmasters might be of different opinion... but there
>are
>> pkgs in archive which just rely on pregenerated .c altogether)
>> 
>> see e.g. pandas where I do the same for easy backporting
>
>I admit I considered the same but I hesitated because I have no idea
>whether the generated C files might be arch independant.  If the
>resulting C++ files would be different per architecture this could
>case problems - if you would confirm that this is not the case I
>would go this path.
>
>Kind regards
>
>       Andreas.

Afaik generated code is architecture agnostic and relies on preprocessor to provide architecture specific code
So it should be safe to include it but great to make it so it would get regenerated if a newer/same version of cython is available so later in case of cython bug fixes they could get in effect via binNMU   
-- 
Sent from a phone which beats iPhone.