Bugzilla Bug 233968

Please arches, stabilize net-print/hplip-2.8.6b due to bug #233959. It's
a bit fresh but the ebuild and the package itself fix a lot of bugs and add a
lot of printers. Also upstream is of above average quality. If it's OK with
everybody I'm ready to pick up the pieces in case something breaks.

Please do so quickly. Thanks in advance.

Denis.

ppc64 stable

Denis, sorry to say but I had serious problems with the new ebuild for 2.8.6b,
upon emerging it, hp devce manager complained about not being able to
communicate with the device and also complained about not being able to start
sistray. I renamed the ebuild for 2.8.5 and emerged it, Hplip 2.8.7 with the
2.8.5 ebuild works perfectly. I don't know what you changed in that ebuild but
it caused me major problems!!!!!!!

(In reply to comment #2)
> Denis, sorry to say but I had serious problems with the new ebuild for 2.8.6b,
> upon emerging it, hp devce manager complained about not being able to
> communicate with the device and also complained about not being able to start
> sistray. I renamed the ebuild for 2.8.5 and emerged it, Hplip 2.8.7 with the
> 2.8.5 ebuild works perfectly. I don't know what you changed in that ebuild but
> it caused me major problems!!!!!!!

Check your USE flags, try with qt3 alone and qt4 alone, and if it doesn't work
open a bug for it with all the usual info (emerge --info, hp-check, emerge -vp
hplip). Don't forget to add the actual output when it fails because "major
problems!!!!!!!" is kinda hard to trace.

Let's not have this block the stabilization (i.e. suspected PEBKAC).

Denis.

Please make sure that all dependencies are stable/keyworded too. I have now
created all necessary bugs (see blockers). If you see more dependencies that
need to be stabilized/keyworded please open a bug and add to blockers here.

Re-CCing ppc64.

Denis.

amd64/x86 stable

ppc64 already stable

I've also hit the problem referred to in comment #2.

The problem is with this line in the ebuild:

sed -i -e "s/--force-startup/--force-startup --qt${QT_VER}/" base/device.py ||
die "sed failed"

Which creates the invalid line:
os.spawnlp(os.P_NOWAIT, path, 'hp-systray', '--force-startup --qt3')

This works properly:
os.spawnlp(os.P_NOWAIT, path, 'hp-systray', '--force-startup', '--qt3')

Thanks a lot Gordon. This issue doesn't occur when you don't disable hp-systray
starting automatically at logon, which is my case.

Billy, can you please sync and emerge the fixed ebuild and then confirm it's
the same issue you were having ?

Denis.

I just emerged the fixed ebuild and it now appears to work properly. But please
remember that this needs to be bumped immediately to 2.8.7, thats the newest
version.

(In reply to comment #9)
> remember that this needs to be bumped immediately to 2.8.7

"immediately" ? Who are you to speak like this ? There's a bug for this
already.

By the way, you could at least thank Gordon for tracing and fixing an issue for
you that you didn't even bother reporting properly (i.e. in its own bug and
with a valid description).

Denis.

Denis, you are taking my comment out of context. I simply meant that if we are
trying to stabilize a version right now with this bug, it should get bumped to
the newest version. Please excuse my manner of making my point, no disrespect
was intended. If I am correct, the 2.8.6b ebuild was created at the same time
2.8.7 was released. It would be great to apply these efforts towards the newest
version. I am very glad Gordon has tracked down the problem and that it is
fixed. I use HPLIP for a long time and amazingly, my officejet pro l7780 works
better on Gentoo that it does on Windows XP. On XP I am unable to scan to my
computer, On Linux it works perfectly.
My Apologies for any misunderstandings!!

(In reply to comment #11)
> the newest version. Please excuse my manner of making my point, no disrespect
> was intended.

All is right then.

> If I am correct, the 2.8.6b ebuild was created at the same time
> 2.8.7 was released. It would be great to apply these efforts towards the
> newest version.

Yes and no. I always test an ebuild such as for hplip for a few days before I
commit it. The damage that a stupid mistake can do for such a widely used
package makes it worth it. Plus, the 2.8.6b ebuild had been in use for a few
days by Gentoo users and about 10 days by me when the security issue appeared.
Compare that with not yet existing and completely untested on Gentoo for 2.8.7.
So since we usually require one bug-free month before stabilizing and we had
less than a week in that particular case, I decided it was bad enough like this
without having to be even more unreasonable.

I'm currently testing 2.8.7 and should be committing it soon. But it won't be
stable before at least a month.

Denis.

ppc stable

All are stable.

(In reply to comment #14)
> All are stable.

Thanks.

Denis.