[OE-core] locale data broken in gstreamer

Martin Jansa martin.jansa at gmail.com
Wed Nov 28 08:57:26 UTC 2012


On Tue, Nov 27, 2012 at 12:44:21PM -0600, Mark Hatle wrote:
> On 11/27/12 12:23 PM, Martin Jansa wrote:
> > In last rebuild from scratch I've noticed couple of weird locale data:
> 
> Paul E and I just observed this as well...  My build machine is an -old- FC13 
> machine..  system we found it on was x86.

OK, the issue is triggered by:

commit 841ea3c1c18e50e77fccbd5f44d6a79a50913b67
Author: Richard Purdie <richard.purdie at linuxfoundation.org>
Date:   Thu Oct 11 08:43:01 2012 +0000

    autotools.bbclass: Fix gettext macro versions issues

Especially this part:
-if [ -d ${S}/po/ -a ! -e ${S}/po/Makefile.in.in ]; then
-     cp ${STAGING_DATADIR}/gettext/po/Makefile.in.in ${S}/po/
+if [ -d ${S}/po/ ]; then
+     cp ${STAGING_DATADIR}/gettext/po/Makefile.in.in ${S}/po/

And it influences gst* because they are hacking their po/Makefile.in.in in autogen.sh by this patch
http://cgit.freedesktop.org/gstreamer/common/commit/gettext.patch?id=c81ad072c76522175cbddead96d6f3c448068d67

RP: do you remember why you removed "-e ${S}/po/Makefile.in.in" check?
I guess it was because it can be from older gettext version (in gstreamer 
case 0.17 instead of our 0.18)

Is there some way to inject Makefile.in.in modification between this 
gettext modifications and running autoreconf/configure? I don't see any 
in run.do_configure, but I guess there is cleaner solution for upstream 
to use.

As work around we can replace DOMAIN setting in po/Makevars in all gst* packages.

Cheers,
> 
> --Mark
> 
> > Collected errors:
> >   * check_data_file_clashes: Package gst-plugins-base-locale-en-gb wants
> >   * to install file /usr/share/locale/en_GB/LC_MESSAGES/.mo
> >          But that file is already provided by package  * gstreamer-locale-en-gb
> >   * check_data_file_clashes: Package gst-plugins-base-locale-cs wants to
> >   * install file /usr/share/locale/cs/LC_MESSAGES/.mo
> >          But that file is already provided by package  * gstreamer-locale-cs
> >
> > All gstreamer and gst-plugins-base locales package just .mo file.
> >
> > from log.do_install:
> > make[1]: Entering directory `/OE/shr-core/tmp-eglibc/work/armv4t-oe-linux-gnueabi/gstreamer/0.10.36-r1/gstreamer-0.10.36/po'
> > installing af.gmo as /OE/shr-core/tmp-eglibc/work/armv4t-oe-linux-gnueabi/gstreamer/0.10.36-r1/image/usr/share/locale/af/LC_MESSAGES/.mo
> > installing az.gmo as /OE/shr-core/tmp-eglibc/work/armv4t-oe-linux-gnueabi/gstreamer/0.10.36-r1/image/usr/share/locale/az/LC_MESSAGES/.mo
> > installing be.gmo as /OE/shr-core/tmp-eglibc/work/armv4t-oe-linux-gnueabi/gstreamer/0.10.36-r1/image/usr/share/locale/be/LC_MESSAGES/.mo
> >
> > They are installed to $(DOMAIN).mo
> > gstreamer-0.10.36/po/Makefile.in.in:$(INSTALL_DATA) $$realcat $(DESTDIR)$$dir/$(DOMAIN).mo; \
> > gstreamer-0.10.36/po/Makefile.in.in:echo "installing $$realcat as $(DESTDIR)$$dir/$(DOMAIN).mo"; \
> >
> > and DOMAIN should be set to GETTEXT_PACKAGE (gstreamer-0.10)
> > ./gstreamer-0.10.36/po/Makevars:DOMAIN = $(GETTEXT_PACKAGE)
> > ./gstreamer-0.10.36/po/Makefile:DOMAIN = $(GETTEXT_PACKAGE)
> >
> > Anyone else seen this, any idea why it fails now? gst* recipes weren't changed lately IIRC,
> > so maybe some change in gettext which can break more recipes?
> >
> > Cheers,
> >
> >
> >
> > _______________________________________________
> > Openembedded-core mailing list
> > Openembedded-core at lists.openembedded.org
> > http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
> >
> 
> 
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core at lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core

-- 
Martin 'JaMa' Jansa     jabber: Martin.Jansa at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.openembedded.org/pipermail/openembedded-core/attachments/20121128/2c231781/attachment-0002.sig>


More information about the Openembedded-core mailing list