syncpackage crashed with urllib.error.URLError in do_open(): <urlopen error [SSL: BAD_SIGNATURE] bad signature (_ssl.c:1129)>

[Erich Eickmeyer]

Happened while trying to sync a package. Looks like it definitely should not fail like this.

ProblemType: Crash
DistroRelease: Ubuntu 21.10
Package: ubuntu-dev-tools 0.184
ProcVersionSignature: Ubuntu 5.13.0-14.14-lowlatency 5.13.1
Uname: Linux 5.13.0-14-lowlatency x86_64
NonfreeKernelModules: zfs zunicode zavl icp zcommon znvpair nvidia_modeset nvidia
ApportVersion: 2.20.11-0ubuntu68
Architecture: amd64
CasperMD5CheckResult: unknown
CurrentDesktop: KDE
Date: Fri Sep 10 13:56:12 2021
ExecutablePath: /usr/bin/syncpackage
InstallationDate: Installed on 2021-03-20 (173 days ago)
InstallationMedia: Ubuntu-Studio 21.04 "Hirsute Hippo" - Alpha amd64 (20210320)
InterpreterPath: /usr/bin/python3.9
PackageArchitecture: all
ProcCmdline: /usr/bin/python3 /usr/bin/syncpackage qxgedit
Python3Details: /usr/bin/python3.9, Python 3.9.7, python3-minimal, 3.9.4-1
PythonArgs: ['/usr/bin/syncpackage', 'qxgedit']
PythonDetails: N/A
SourcePackage: ubuntu-dev-tools
Title: syncpackage crashed with urllib.error.URLError in do_open(): <urlopen error [SSL: BAD_SIGNATURE] bad signature (_ssl.c:1129)>
UpgradeStatus: Upgraded to impish on 2021-06-13 (89 days ago)
UserGroups: adm audio cdrom dip libvirt lpadmin plugdev sambashare sudo

[Erich Eickmeyer (eeickmeyer)]

The issue resolved itself with the apache ssl update shortly after I filed this bug, but it still shouldn't fail as hard as it did.

[Dan Streetman (ddstreet)]

i don't see what else syncpackage should do; if urlopen() fails it should just pass the error up; doing anything else is just hiding the problem.

[Erich Eickmeyer (eeickmeyer)]

Instead of crashing hard, I would think it would output a standard error as opposed to a complete python crash. Something like, "Unable to verify SSL signature. Try again later" or something of that respect. The implementation as it exists causes Apport to think there's a bug in the code when, in fact, there is not. It's just failing, but the fail shouldn't invoke a crash report.