On Tue, Sep 15, 2020 at 7:45 AM Richard Purdie
<[email protected]> wrote:
>
> On Tue, 2020-09-15 at 07:33 -0400, Bruce Ashfield wrote:
> > On Tue, Sep 15, 2020 at 1:44 AM Jens Rehsack <[email protected]>
> > wrote:
> > >
> > >
> > > > Am 15.09.2020 um 00:34 schrieb Bruce Ashfield <
> > > > [email protected]>:
> > > >
> > > > From: Jens Rehsack <[email protected]>
> > > >
> > > > Backporting the 10 patches since the lttng 2.12.2 release. We'll
> > > > drop
> > > > them once .3 is released, but for now, we need the fixes to build
> > >
> > > Yet another question: there is a stable-2.12 branch ... wouldn't it
> > > be reasonable to use this one as upstream source instead of the
> > > tarball?
> > >
> >
> > We used to do that .. yes. But then the lttng recipe was changed away
> > from git to using the tgz.
> >
> > I complained at the time (I'm sure it is in the archives for all to
> > see), since I'm constantly doing new kernels and have to do this sort
> > of exercise.
> >
> > The devupstream bbclass allows me to keep moving on my new kernels
> > when I need to pick up changes for -dev, and luckily -stable doesn't
> > normally need this sort of thing.
> >
> > IIRC it was switched to tgz since things like AUH work better with the
> > tarballs, but I can't be sure.
> >
> > But let's just say that I agree that this recipe is better as a branch
> > based, git recipe and not a tgz based recipe. I just haven't waded
> > into the discussion to switch it back :)
>
> Perhaps we have a need for both?

I suppose we could, but once we've switched back to the git based
recipe, I don't see the need for the tgz based one at all (maybe I'm
missing something though, see below for my rambling).

And by that, I don't see enough pull to a stable/released version of
lttng versus just using the stable/released branches. If we march the
branches forward like this, they are good for all users. i.e. I don't
see anyone carrying a preferred version to pickup and older released
tgz variant, since it wouldn't be the one getting build testing, etc.

devupstream can continue to point at the development/master branch,
and that should cover all the use cases, no ?

Bruce

>
> Cheers,
>
> Richard
>


-- 
- Thou shalt not follow the NULL pointer, for chaos and madness await
thee at its end
- "Use the force Harry" - Gandalf, Star Trek II
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.

View/Reply Online (#142589): 
https://lists.openembedded.org/g/openembedded-core/message/142589
Mute This Topic: https://lists.openembedded.org/mt/76853782/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub  
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to