|
From: | Jan D. |
Subject: | bug#19660: Acknowledgement (24.4.51; Crash in draw_row_fringe_bitmaps) |
Date: | Tue, 27 Jan 2015 10:16:48 +0100 |
User-agent: | Mozilla/5.0 (Macintosh; Intel Mac OS X 10.10; rv:31.0) Gecko/20100101 Thunderbird/31.3.0 |
Eli Zaretskii skrev den 2015-01-26 10:31:
Date: Mon, 26 Jan 2015 10:12:35 +0100 From: "Jan D." <jan.h.d@swipnet.se> CC: 19660@debbugs.gnu.org, "Kim F. Storm" <storm@cua.dk>If no other information emerges, my suggestion would be to remove the assertion on master, and see if we get any bug reports.No need for that, the NS version crashes later on if I remove that assert.Where does it crash, and why? Another similar assertion or something else?
Entering the GUI event loop, while in the event loop already. There is an assertion for that on NS.
Better block input too much than to seldom I guess.Right. But understanding why this happens is better yet ;-) Seriously: I thought we aren't supposed to be hit by this stuff, since we now never run code out of an interrupt handler. I guess I missed something important.
I suspect doing unblock_input can enter the GUI event loop unconditionally if there is input pending. This might be wrong in some situations. Do we actually need block/unblock_input anymore?
Unfortunately I managed to loose my .desktop that consistently triggered the error, so I can't investigate further.Don't you have a backup of that somewhere?
Yes, but it does not crash anymore. It was probably timing related also as GUI events seems to be involved.
Jan D.
[Prev in Thread] | Current Thread | [Next in Thread] |