On Fri, 2022-05-27 at 09:29 +0100, Jose Quaresma wrote:
> The archiver doesn't support using multiconfig sharing the same TMPDIR
> so rise an error when we have something that shouldn't exist.
>
> When in the multiconfig and using the same TMPDIR: the two machines can make
> the same copy at the same time to the same destination, which will not work.
>
> Signed-off-by: Jose Quaresma <jose.quare...@foundries.io>
> ---
> meta/classes/archiver.bbclass | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/meta/classes/archiver.bbclass b/meta/classes/archiver.bbclass
> index c19c770d11..5514e44a78 100644
> --- a/meta/classes/archiver.bbclass
> +++ b/meta/classes/archiver.bbclass
> @@ -486,6 +486,8 @@ python do_unpack_and_patch() {
> if d.getVarFlag('ARCHIVER_MODE', 'diff') == '1':
> src = d.getVar('S').rstrip('/')
> src_orig = '%s.orig' % src
> + if os.path.isdir(src_orig):
> + bb.error("A previous copy of the original source already exist
> '%s'" % src_orig)
> oe.path.copytree(src, src_orig)
>
> if bb.data.inherits_class('dos2unix', d):
I suspect we need to make archiver hard error if multiconfig is being
used in that case. This can work around some subset of the races
but could also still result in broken trees of files :(
I'd also note that the patch will still break since bb.error doesn't
actually exit, it just prints an error and causes the task to fail. You
probably meant bb.fatal().
Cheers,
Richard
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#166205):
https://lists.openembedded.org/g/openembedded-core/message/166205
Mute This Topic: https://lists.openembedded.org/mt/91372555/21656
Group Owner: openembedded-core+ow...@lists.openembedded.org
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-
