On Tue, Dec 17, 2024 at 02:20:16PM +0100, Chris Hofstaedtler wrote: > josch@ wrote on salsa: > > I would bet a lot on this breaking something. The problem are > > settings which have a GET function which is allowed to perform > > some magic.
> > Can you try this patch and see what happens? I have to say that results in some "interesting" breakage. For one, autopkgtest is unhappy. Running sbuild -s --no-clean-source -d unstable hello gives: | autopkgtest | ----------- | | autopkgtest [14:56:35]: starting date and time: 2024-12-17 14:56:35+0100 | autopkgtest [14:56:35]: version 5.42 | autopkgtest [14:56:35]: host tiksta; command line: /usr/bin/autopkgtest /home/ch/Debian/tryout/hello_2.10-3_arm64.changes --apt-upgrade -- unshare --release unstable --arch arm64 | autopkgtest [14:56:36]: testbed dpkg architecture: arm64 | autopkgtest [14:56:36]: testbed apt version: 2.9.17 | autopkgtest [14:56:36]: @@@@@@@@@@@@@@@@@@@@ test bed setup | autopkgtest [14:56:36]: testbed release detected to be: None | autopkgtest [14:56:36]: updating testbed package index (apt update) | Get:1 http://deb.debian.org/debian unstable InRelease [202 kB] | Get:2 http://deb.debian.org/debian unstable/main arm64 Packages [9977 kB] | Get:3 http://deb.debian.org/debian unstable/main Translation-en [7349 kB] | Fetched 17.5 MB in 2s (11.6 MB/s) | Reading package lists... | autopkgtest [14:56:38]: upgrading testbed (apt dist-upgrade and autopurge) | Reading package lists... | Building dependency tree... | Reading state information... | Calculating upgrade...Starting pkgProblemResolver with broken count: 0 | Starting 2 pkgProblemResolver with broken count: 0 | Done | Entering ResolveByKeep | | 0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded. | Reading package lists... | Building dependency tree... | Reading state information... | Starting pkgProblemResolver with broken count: 0 | Starting 2 pkgProblemResolver with broken count: 0 | Done | 0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded. | autopkgtest [14:56:39]: testbed running kernel: Linux 6.11.10-arm64 #1 SMP Debian 6.11.10-1 (2024-11-23) | autopkgtest [14:56:39]: @@@@@@@@@@@@@@@@@@@@ source /home/ch/Debian/tryout/hello_2.10-3.dsc | Unexpected error: | Traceback (most recent call last): | File "/usr/share/autopkgtest/lib/VirtSubproc.py", line 833, in mainloop | command() | File "/usr/share/autopkgtest/lib/VirtSubproc.py", line 762, in command | r = f(c, ce) | ^^^^^^^^ | File "/usr/share/autopkgtest/lib/VirtSubproc.py", line 696, in cmd_copydown | copyupdown(c, ce, False) | File "/usr/share/autopkgtest/lib/VirtSubproc.py", line 584, in copyupdown | copyupdown_internal(ce[0], c[1:], upp) | File "/usr/share/autopkgtest/lib/VirtSubproc.py", line 611, in copyupdown_internal | copydown_shareddir(sd[0], sd[1], dirsp, downtmp_host) | File "/usr/share/autopkgtest/lib/VirtSubproc.py", line 566, in copydown_shareddir | shutil.copy(host, host_tmp) | File "/usr/lib/python3.12/shutil.py", line 435, in copy | copyfile(src, dst, follow_symlinks=follow_symlinks) | File "/usr/lib/python3.12/shutil.py", line 260, in copyfile | with open(src, 'rb') as fsrc: | ^^^^^^^^^^^^^^^ | FileNotFoundError: [Errno 2] No such file or directory: '/home/ch/Debian/tryout/hello_2.10.orig.tar.gz' | autopkgtest [14:56:39]: ERROR: testbed failure: unexpected eof from the testbed | | E: Autopkgtest run failed. Which is certainly weird. I think this might be some interaction between -s and $source_only_changes=1. I don't have any more details yet, need to dig into testing/diffing logs more. Chris
