GNOME Bugzilla – Bug 764299
Strange cursor artifacts
Last modified: 2016-04-04 04:04:33 UTC
Created attachment 324917 [details] screencast As of version 0.44, I experience a strange cursor glitch. Not sure how to describe it. I have attached a screencast to demonstrate. It happens consistently in vim but now and then in neovim and zsh prompt. I have tested multiple vte based terminal emulators, multiple configurations and a fresh install. All the same. This does not occur with 0.42 installed. It still exists with the current master.
Could you please record a "script" session of this bug occurring, along with timing? E.g. "script -t 2>timing", then vim..., then exit from vim, exit from the shell, and attach both "typescript" and "timing" to this bug? Please also take note of the exact window size (as reported e.g. by "stty size"). Does replaying the script with "scriptreplay -t timing" (in a window of the same size) also reproduce the glitch for you? Also, what's the exact 0.42 version? Could you please try vte-0-42 from git, or the 0.42.5 tarball? There was a cursor related change between 0.42.4 and 0.42.5, I'm curious if it brought any regression (although I find it very unlikely).
Very useful utilities I didn't know of! The versions I tested against were from the corresponding git branches: vte-0-42 - 0.42.6.3493.e62b98c vte-0-44 - 0.44.1.3932.0af9012 master - 0.45.0.3936.8bc51bb Replaying the script indeed reproduces the glitches.
Created attachment 324919 [details] script recording @ 157 x 54
Created attachment 324920 [details] script recording - timing
Thanks, I can also reproduce the problem by replaying the recording.
423063840d1304fdd7ff2d87f2041e70a07f599a is the first bad commit commit 423063840d1304fdd7ff2d87f2041e70a07f599a Author: Christian Persch <chpe@gnome.org> Date: Fri Nov 20 20:58:31 2015 +0100 widget: Move some methods to VteTerminalPrivate :040000 040000 0a8b8e809b3634061a0a39bed406076647b54df6 22b9b5c1d64748c3cb2b5f5a89077d1593e6e449 Msrc
Created attachment 324934 [details] [review] Fix Could you please test this patch? Note: The first FIXME was indeed valid and this needed to be fixed. I don't see the point in the second FIXME, since it's in the branch where saved_cursor and cursor are the same position, so the code is just correct there.
Fixed on master and vte-0-44 (0.44.1). Older branches were not affected.
Hi, Egmont. Sorry for not responding sooner, timezones and all. I can confirm that the recent change fixes the problem. A very sincere thank you for all effort involved. All the best!
*** Bug 764560 has been marked as a duplicate of this bug. ***