Red Hat Bugzilla – Bug 1288046
Evolution gets some stuck "... (cancelling)" tasks after connectivity breaks (e.g. laptop undock)
Last modified: 2017-03-09 13:00:07 EST
Created attachment 1101782 [details] backtrace for "Reconnecting to '<account_name>' (cancelling)" Description of problem: When using evolution with hard connection breaks (e.g. pull laptop out of dock, suspend, resume elsewhere or at the same location after a couple of hours), I'm getting some stuck network operations, some still active, some "cancelling". The attached backtrace is an example of evo in such a state, the particular message is: "Reconnecting to 'acount_name@example.org' (cancelling)". When in such a state, evo can't be cleanly shut down, it has to be killed with 9. Stop button (per task or global) has no effect either. Otherwise, the program works mostly normally, I noticed some IMAP sync issues but I can't tell for sure that they are caused by this bug. Version-Release number of selected component (if applicable): evolution-3.12.11-15.el7.x86_64 evolution-data-server-3.12.11-24.el7.x86_64 How reproducible: frequent Steps to Reproduce: 1. use Evo on location-changing over some time, do hard connectivity breaks 2. 3. Actual results: Evo develops 'cancelling' tasks that are cancelling forever Expected results: Evo handles sudden connectivity breaks gracefully Additional info:
Thanks for a bug report. The backtrace shows just that, the IMAPx account is connecting to the server. The thing is that there is missing another thread, which is supposed to do the actual connection. Either it didn't start yet, or the connect failed in some way (could be cancelled) and it didn't tell its initiator about it.