Closed
Bug 954748
Opened 10 years ago
Closed 10 years ago
Input/output error on quit when closing Instantbird
Categories
(Chat Core :: IRC, defect)
Tracking
(Not tracked)
RESOLVED
FIXED
1.2
People
(Reporter: aleth, Unassigned)
Details
Attachments
(1 file, 1 obsolete file)
543 bytes,
patch
|
clokep
:
review+
|
Details | Diff | Splinter Review |
*** Original post on bio 1316 at 2012-03-03 22:51:00 UTC *** Quite often one sees the JS-IRC quit message "Quit: Input/output error" instead of the default quit message. Circumstantial evidence suggests this occurs when IB quits (as opposed to just the account). Flo mentioned that probably indicates that we are sending crap to the server at the time we disconnect in some cases.
Reporter | ||
Comment 1•10 years ago
|
||
*** Original post on bio 1316 as attmnt 1219 at 2012-03-04 22:50:00 UTC *** This fixes the problem. I'm not sure whether calling this.disconnect instead would be better though.
Attachment #8352972 -
Flags: review?(clokep)
Comment 2•10 years ago
|
||
Comment on attachment 8352972 [details] [diff] [review] Patch *** Original change on bio 1316 attmnt 1219 at 2012-03-04 22:57:21 UTC *** I'm OK calling this.quit() here instead of this.disconnect() (I think it works better actually). I agree with flo that we should move the comment that method (and the socket disconnection code) together. (and delete the imAccount below that).
Attachment #8352972 -
Flags: review?(clokep) → review-
Reporter | ||
Comment 3•10 years ago
|
||
*** Original post on bio 1316 at 2012-03-04 23:11:32 UTC *** The patch doesn't always solve the problem, though it improves things a bit. It seems the quit message sent to the socket doesn't reach the server as IB can quit before it is transmitted.
Reporter | ||
Comment 4•10 years ago
|
||
*** Original post on bio 1316 as attmnt 1220 at 2012-03-04 23:26:00 UTC *** As discussed on IRC, a complete solution will additionally require IB to wait until sockets have transmitted all data still in the pipe before quitting.
Attachment #8352973 -
Flags: review?(clokep)
Reporter | ||
Comment 5•10 years ago
|
||
Comment on attachment 8352972 [details] [diff] [review] Patch *** Original change on bio 1316 attmnt 1219 at 2012-03-04 23:26:43 UTC was without comment, so any subsequent comment numbers will be shifted ***
Attachment #8352972 -
Attachment is obsolete: true
Comment 6•10 years ago
|
||
Comment on attachment 8352973 [details] [diff] [review] Patch *** Original change on bio 1316 attmnt 1220 at 2012-03-04 23:50:24 UTC *** This one looks good! There seems to be a flush method [1] that maybe we should call on the socket in the quit method? [1] https://developer.mozilla.org/en/XPCOM_Interface_Reference/nsIOutputStream#flush%28%29
Attachment #8352973 -
Flags: review?(clokep) → review+
Reporter | ||
Comment 7•10 years ago
|
||
*** Original post on bio 1316 at 2012-03-05 00:10:15 UTC *** (In reply to comment #5) > There seems to be a flush method [1] that maybe we should > call on the socket in the quit method? close() in http://lxr.instantbird.org/instantbird/source/chat/modules/socket.jsm#198 already calls flush().
Comment 8•10 years ago
|
||
*** Original post on bio 1316 at 2012-03-06 00:41:34 UTC *** (In reply to comment #4) > Created attachment 8352973 [details] [diff] [review] (bio-attmnt 1220) [details] > Patch > > As discussed on IRC, a complete solution will additionally require IB to wait > until sockets have transmitted all data still in the pipe before quitting. This is checked in as http://hg.instantbird.org/instantbird/rev/f893c587df9e but I'm going to leave this open for now.
Reporter | ||
Comment 9•10 years ago
|
||
*** Original post on bio 1316 at 2013-08-26 10:42:16 UTC *** We haven't seen this in a while, so resolving as fixed.
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Updated•10 years ago
|
Target Milestone: --- → 1.2
You need to log in
before you can comment on or make changes to this bug.
Description
•