Well I noticed the crash because upower uses idevice_new in up-device-idevice.c, I guess to monitor idevice's batteries. In my system upower was endlessly rebooting itself, and when checking I saw the crash was in idevice_new itself. Further investigation lead me to confirm that idevice utils crash themselves, thus the observations in my previous email.
So I guess the short story is that I noticed the problem because it crashed upower, but the bug was in libimobiledevice all along. My apologies if my initial report wasn't very complete, I'll blame myself for being rusty from years of non desktop work :P. On Fri, Oct 4, 2019 at 4:11 AM Yves-Alexis Perez <cor...@debian.org> wrote: > > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA256 > > On Fri, 2019-10-04 at 03:18 -0500, Diego Escalante Urrelo wrote: > > While trying to play around with this, I rebuild libimobiledevice > > locally from the sources available in unstable and the crash went > > away. > > If I reinstall the .deb from the repositories I can get the crash > > back, reinstalling the locally built package fixes the problem (from > > apt source + dpkg-buildpackage -us -uc). > > > > I guess this is still a libimobiledevice bug since I can reproduce the > > stack smash by running ideviceinfo and a few other of the idevice > > utilities (idevicename, idevicesyslog): > > It seems there's a tight link between libimobiledevice and libusbmuxd > versions, but I fail to see how UPower is related to that, to be honest. > - -- > Yves-Alexis > -----BEGIN PGP SIGNATURE----- > > iQEzBAEBCAAdFiEE8vi34Qgfo83x35gF3rYcyPpXRFsFAl2XDJkACgkQ3rYcyPpX > RFufCggA4bLQqtUGNHdROTEZt3iOcKivQIwsu7t1JM47RdgXqwWhtSz2OqDwclpF > KoL24Rep1NbseLkcUeACMpCnJupmwfV8nbaEv8JFymdHX75MrvYLuyaxk7NYgpbW > sznhOa/8CUDIV2Z9z/CIA9vC3SoY2pbkjXJDABgH1gkd+A2nKkI3r7cwSuq0lwR9 > cIpfo7ItTHvq5F977JKcmsGDr/N9a3P7Ra1gbw98e7zOVbnrDBRD/9TISR8oF9dA > Ni+Ic1x6Ipiz+RNo/6dXBF5c7e7wlFvvJJ/LEFkriChWQu9VsvH8842rlCwch6Gb > xMInKU3OaOqUMK9nRS6G0M1mJpYsnw== > =2ETL > -----END PGP SIGNATURE-----
