[OE-core] [PATCH][v2 00/10] Make MIME work on desktops

Richard Purdie richard.purdie at linuxfoundation.org
Sat Jan 18 11:13:17 UTC 2020


On Sat, 2020-01-18 at 12:08 +0100, Andreas Müller wrote:
> On Sat, Jan 18, 2020 at 9:40 AM Richard Purdie
> <richard.purdie at linuxfoundation.org> wrote:
> > > V1 -> V2:
> > > * Shelve global path export thingy
> > > * Add me as maintainer of itstool (let's consider 'ü' in my name
> > > as
> > > test case :)
> > > * Change "to to" to "to" in commit message
> > > * Build shared-mime-info from git to allow automated upgrades
> > > * mime-xdg.bbclass: Be prepared for packages as libreoffice:
> > > Desktop
> > > files
> > >   installed are absolut symlinks broken for us. In case other
> > > projects do same:
> > >   Offer asolution and guide recipe writers how to handle by
> > > generating warnings
> > >   with instructions how to handle.
> > 
> > Thanks, this looks good to me. It did generate a warning on master-
> > next
> > builds in many places:
> > 
> > WARNING: libfm-1.3.1-r0 do_package_qa: QA Issue: package contains desktop file with key 'MimeType' but does not inhert mime-xdg: libfm-gtk path '/work/armv7vet2hf-neon-oe-linux-gnueabi/libfm/1.3.1-r0/packages-split/libfm-gtk/usr/share/applications/lxshortcut.desktop' [mime-xdg]
> > WARNING: libfm-1.3.1-r0 do_package_qa: QA Issue: package contains mime types but does not inhert mime: libfm-mime path '/work/armv7vet2hf-neon-oe-linux-gnueabi/libfm/1.3.1-r0/packages-split/libfm-mime/usr/share/mime/packages/libfm.xml' [mime]
> Slipped through - will send a follow up

Thanks, should be simple at least! :)

> > It also generated:
> > 
> > https://autobuilder.yoctoproject.org/typhoon/#/builders/75/builds/1458
> > 
> > which partly is the python2 removal that was in with the test and
> > partly this series since itstool is GPLv3 :(.
> Is there any activity required from my side?

Well, as things stand this series breaks meta-gplv2. This is currently
tested on the autobuilder and turns it red. I'm therefore blocked on
merging.

> > I have to admit I'm not sure how reasonable it is to keep hacking the
> > gplv2 layer to work...

The question now is really whether we want to continue supporting meta-
gplv2 with master?

I don't like the layer, I think there are other approaches we're going
to need to take going forward and the project may need to send that
message.

If someone steps up and fixes it, great, we can move forward and I'll
defer that discussion.

If nobody does, quickly, then I will need to start a discussion about
dropping it from the core YP testing. This patch is blocked until
something changes though :(

Cheers,

Richard




More information about the Openembedded-core mailing list