[arch-multilib] feature request: lib32-libwebkit

[Oon-Ee Ng]

On Sat, Feb 4, 2012 at 7:11 AM, Oon-Ee Ng <ngoonee.talk at gmail.com> wrote:
> On Sat, Feb 4, 2012 at 6:23 AM, rafael ff1 <rafael.f.f1 at gmail.com> wrote:
>> The binary (/usr/bin/g-ir-scanner) and libraries (/usr/lib/*) are part
>> of the package gobject-introspection. The package lib32-webkit
>> requires a lib32- of gobject-introspection. But I never had success
>> with this package, as I always end up this error
>> (http://pastebin.com/kh9zYS7A), which looks like a need for a 32 bit
>> libs/modules of python.
>> Python, on the other hand, is not much multilib friendly (if I'm not
>> wrong, this expression was used by someone for python support in a
>> feature request for multilib). You set libdir to /usr/lib32, but
>> "${prefix}/lib/python2.7/site-packages" were hardcoded in it,
>> therefore ignoring /usr/lib32... Well, don't know. I gave up, hoping
>> some TU from [multilib] would adopt this cause. :)
>>
>> About the dead package lib32-python from AUR, I was its maintainer.
>> The package was a copy-and-paste of libs of i686 package, but it
>> didn't make sense as target source codes (that you would compile)
>> would search /usr/bin/python2, which was not patched to use /usr/lib32
>> folder... So this package was useless.
>>
> Thanks for the explanation. I built a lib32-python in /opt/python32
> but yes, it seems to hardcode references to /usr/lib/python in any
> case.
>
> I have trouble believing that lib32-libwebkit (which AFAIK must exist
> in other linux distros if realstudio is to run properly on them at
> all) requires something unbuildable though. Rushing off to a sports
> game, may take a look at debs/rpms later on. I know how to extract
> from them, but do they have build instructions embedded like our
> PKGBUILDs?

After some more messing around, I realized using a 32-bit chroot would
be easier.... if anyone would be kind enough to build lib32-libwebkit
for [multilib] I'd be grateful though =).