On Fri, 2023-08-25 at 19:16 +0200, Peter Kjellerstedt wrote:
> After commit a8e7b0f932 (base/package: Move source revision information
> from PV to PKGV) was integrated, having a recipe with a SRCREV that
> currently cannot be fetched would lead to an exception during parsing.
> Catch that exception and instead raise bb.parse.SkipRecipe. That way
> the parsing continues as it should. Instead you now get a meaningful
> error if you try build a recipe with a SRCREV that cannot be fetched,
> e.g.:
> 
>   ERROR: Nothing PROVIDES 'psplash'
>   psplash was skipped: Fetcher failure: Unable to resolve 'unknown-ref'
>   in upstream git repository in git ls-remote output for
>   git.yoctoproject.org/psplash

Something doesn't sound quite right in this description. You've said "a
SRCREV that currently cannot be fetched" but "unknown-ref" will never
be fetchable and is completely invalid as a revision. I'd guess bitbake
is assuming it is a tag and trying to resolve it.

I really don't like complicating the core code when it doesn't make
sense to, particularly when it relates to cornercases most people don't
hit. I think SkipRecipe is also the incorrect thing to do here,
particularly when no reason is being passed back to bitbake to give
back to the user.

I suspect if you set the SRCREV to something that looks like a
revision, like "badbadbadbadbadbadbadbadbadbadbadbadbadb" it will avoid
the problems. Alternatively, just add the SkipRecipe locally along with
your unknown-ref.

Cheers,

Richard

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#186874): 
https://lists.openembedded.org/g/openembedded-core/message/186874
Mute This Topic: https://lists.openembedded.org/mt/100960059/21656
Group Owner: openembedded-core+ow...@lists.openembedded.org
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to