On 22/09/2011 23:20, Steve Langasek wrote:
> On Thu, Sep 22, 2011 at 09:24:24AM +0200, Vincent Danjean wrote:
>> For my part, I think this is really error prone.
>
> Well, we'll have to agree to disagree then and let the maintainer decide.
Ok ;-)
>> Lots of native libraries will be wrongly det
On Thu, Sep 22, 2011 at 09:24:24AM +0200, Vincent Danjean wrote:
> > Keeping /usr/lib/pkgconfig on the path was quite deliberate on my part.
> > Certainly any library shipping in /usr/lib/pkgconfig is not a multiarch
> > library, but that doesn't actually mean it has to be a *native* library: it
>
On 22/09/2011 07:36, Steve Langasek wrote:
> So my suggestion would be to make the script special-case i386, instead of
> adding a runtime dependency on dpkg-dev for dpkg-architecture.
We can use dpkg-architecture and fail back on hard-coded (with or
without the handling of i386) if dpkg-architect
Hi,
I like your proposal. I have one little remark.
pkg-config does not behave the same if PKG_CONFIG_LIBDIR is unset or
if it is empty (in this case, it does not try system directories). You
can keep this behavior by replacing:
if [ -z "$PKG_CONFIG_LIBDIR" ];
by:
if [ x"${PKG_CONFIG_LIBDIR+se
4 matches
Mail list logo
