* uri_replace not working prooperly
@ 2021-11-13 18:10 Mark Hatle
0 siblings, 0 replies; only message in thread
From: Mark Hatle @ 2021-11-13 18:10 UTC (permalink / raw)
To: bitbake-devel; +Cc: Mark Hatle
I ran into a problem yesterday (which prompted the previous email I sent about
lockfile and ENAMETOOLONG) that was triggered by the uri_replace code.
Specifically patches:
96c30007dc0b32eee2b15771daec7948bc9bfd97
fetch2: fix downloadfilename issue with premirror
and
8a3ff9f3eaf19d4258eb070c5dc230dface269b2
bitbake: fetch2: fix premirror URI when downloadfilename defined
(The former works to resolve an issue with the later.)
What is happening is the uri_mirror can (and should) adjust it's own mirror urls
in a pattern. However, something is going catastrophically wrong and ends up
generating this long entry:
> DEBUG: For url ['http', 'petalinux.xilinx.com', '/sswreleases/rel-v2022.1/downloads/git2_git.qemu.org.git.seabios.git..tar.gz', '', '', OrderedDict()] comparing ['svn', '.*'
> , '/.*', '', '', OrderedDict()] to ['https', 'downloads.yoctoproject.org', '/mirror/sources/', '', '', OrderedDict()]
> DEBUG: For url ['https', 'downloads.yoctoproject.orggit2_git.qemu.org.git.seabios.git..tar.gz', '/git2_git.qemu.org.git.seabios.git..tar.gzmgit2_git.qemu.org.git.seabios.git
> ..tar.gzigit2_git.qemu.org.git.seabios.git..tar.gzrgit2_git.qemu.org.git.seabios.git..tar.gzrgit2_git.qemu.org.git.seabios.git..tar.gzogit2_git.qemu.org.git.seabios.git..tar
> .gzrgit2_git.qemu.org.git.seabios.git..tar.gz/git2_git.qemu.org.git.seabios.git..tar.gzsgit2_git.qemu.org.git.seabios.git..tar.gzogit2_git.qemu.org.git.seabios.git..tar.gzug
> it2_git.qemu.org.git.seabios.git..tar.gzrgit2_git.qemu.org.git.seabios.git..tar.gzcgit2_git.qemu.org.git.seabios.git..tar.gzegit2_git.qemu.org.git.seabios.git..tar.gzsgit2_g
> it.qemu.org.git.seabios.git..tar.gz/git2_git.qemu.org.git.seabios.git..tar.gz', '', '', OrderedDict()] comparing ['ftp', '.*', '/.*', '', '', OrderedDict()] to ['http', 'pet
> alinux.xilinx.com', '/sswreleases/rel-v2022.1/downloads', '', '', OrderedDict()]
(later line repeats for each type..)
The end result of this is that the locking code attemtps to lock a file with the
name:
git2_git.qemu.org.git.seabios.git..tar.gzmgit2_git.qemu.org.git.seabios.git..tar.gzigit2_git.qemu.org.git.seabios.git..tar.gzrgit2_git.qemu.org.git.seabios.git..tar.gzrgit2_git.qemu.org.git.seabios.git..tar.gzogit2_git.qemu.org.git.seabios.git..tar.gzrgit2_git.qemu.org.git.seabios.git..tar.gz/git2_git.qemu.org.git.seabios.git..tar.gzsgit2_git.qemu.org.git.seabios.git..tar.gzogit2_git.qemu.org.git.seabios.git..tar.gzugit2_git.qemu.org.git.seabios.git..tar.gzrgit2_git.qemu.org.git.seabios.git..tar.gzcgit2_git.qemu.org.git.seabios.git..tar.gzegit2_git.qemu.org.git.seabios.git..tar.gzsgit2_git.qemu.org.git.seabios.git..tar.gz_git2_git.qemu.org.git.seabios.git..tar.gz
Which needless to say is way way way too long.
I'm working to find a way to reproduce this issue outside of our internal
configuration. (as you can see what is being referenced is all public urls, so
it's just a matter of me building a reproducer.)
The odd thing, if I do a -c cleanall on the recipe, and retry it then works
fine. So it may also have some issue with other tasks running at the same time.
If I can figure out a solid reproducer for this (even if it's complex), I'll
reply to this with the details.
--Mark
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2021-11-13 18:10 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-11-13 18:10 uri_replace not working prooperly Mark Hatle
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.