Replace system_clock with steady_clock #2559
Conversation
This fixes #857 and #1932. Also replaces PR #1725, since we don't need our own implementation of condition_variable anymore since people who update their polybar should have GCC 10 by now. The m_updated_at field of the mpd module was removed instead of having its clock change because it became unused in commit 645a314.
|
Thanks for refreshing the PR :) I'm not totally up-to-date on the bug. But what happens if someone builds polybar with gcc < 10. Do they just get the old behavior of the clock not being truly monotonic? If so, I guess that's fine, we don't lose anything. But if this makes polybar not work properly with older gcc versions, we may have to rethink this. We aim to support any currently supported Ubuntu version (at least standard support), so we should try to make polybar work also on Ubuntu 18.04 with gcc 7.4 |
There was a problem hiding this comment.
Digging deeper into the original but report, it seems older GCC versions will still have the issue. That's fine.
I'll wait another day or two for you to comment, if you maybe have a different insight.
But otherwise, I think this is ready to merge. Thanks a lot 😃
|
Also did the digging (on gcc 7.5), and I came to the same conclusion as you did. Thanks for reviewing :) |
|
Alright, thanks! |
What type of PR is this? (check all applicable)
Description
Where used for timing, replace system_clock with steady_clock so that the updating of modules isn't affected by system clock changes.
The m_updated_at field of the mpd module was removed instead of having
its clock changed because it became unused in commit 645a314.
Related Issues & Documents
This fixes #857 and #1932. Also replaces PR #1725, since we don't need
our own implementation of condition_variable anymore since people who
update their polybar should have GCC 10 by now (which was released in may 2020).
Fixes #857
Fixes #1932
Closes #1725
Documentation (check all applicable)