XCB Assert failure again
Posted: Sun Jul 30 2023 3:20 pm
All,
Everything is merged from tip of tip in LS-CS. RedDiamond, and basically anything written with CopperSpice and using the clipboard gets the following assertion.
This happens when applications are left up and "idle" while other applications (editors/browsers/etc) are being used.
Did not dig into code, but I would hazard a guess that the clipboard code was written more from a "there can be only one" perspective than reality permits.
https://www.uninformativ.de/blog/postings/2017-04-02/0/POSTING-en.html
I will further a guess that we get here as part of some "timed validation" for a <CTRL>-C copy done in the application using CopperSpice. One of the numerous other applications running as since done multiple <CTRL>-C copies to clipboard and the one CopperSpice wishes to validate simply no longer exists.
This should be taken in stride rather than throwing an Assert.
Everything is merged from tip of tip in LS-CS. RedDiamond, and basically anything written with CopperSpice and using the clipboard gets the following assertion.
Code: Select all
ASSERT: "m_timestamp[mode] == XCB_CURRENT_TIME || getSelectionOwner(atomForMode(mode)) == m_owner" in file /home/developer/github_projects/copperspice/src/plugins/platforms/xcb/xcb_support/qxcb_clipboard.cpp, line 466
Aborted (core dumped)
Did not dig into code, but I would hazard a guess that the clipboard code was written more from a "there can be only one" perspective than reality permits.
https://www.uninformativ.de/blog/postings/2017-04-02/0/POSTING-en.html
I will further a guess that we get here as part of some "timed validation" for a <CTRL>-C copy done in the application using CopperSpice. One of the numerous other applications running as since done multiple <CTRL>-C copies to clipboard and the one CopperSpice wishes to validate simply no longer exists.
This should be taken in stride rather than throwing an Assert.