Commit Graph

16 Commits

Author SHA1 Message Date
Carsten Haitzler f95551b50b ecore-x - use current time not vblank timestamp for animator time
sometimes on some systems and drviers time here is not what we
want/expect. it can be in the future. there are reasons so, work
around this and use the timestamp of "now" when we get the event as
opposed to the timestamp given to work around odd weirdnesses and
complaints time went backwards.

@fix
2022-09-16 14:10:15 +01:00
Carsten Haitzler 1d57aa2a4c ecore - loop time - do not allow setting into the future - not intended
setting a loop time timestamp in the future will lead to all sorts of
bad things. the idea is it was meant to go back a little in time AFTER
some sync/animation etc. event to pretend to be at the time when that
event happend (it just took some time ot arrive at the process) and so
animation and other timelines all agree to be at this time a little
bit in the past. going forwards leads to bad things so disallow it and
complain. this fixes weston in a window problems when it sends
timestamps in the future from weston...

@fix
2021-12-04 09:18:47 +00:00
Vincent Torri de296b13c8 Eina: use eina_fnmatch() instead of fnmatch()
Reviewers: raster

Reviewed By: raster

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D12262
2021-05-03 18:48:24 +01:00
Carsten Haitzler 90e5fd831d ecore-x - vsync - also log the direct device timestamp if it goes back 2021-02-03 02:26:54 +00:00
Carsten Haitzler 4013c98af1 ecore_x - vsync ... this handles time going backwards
yes - time went backwards. we get time from the device driver and
vsync events... this is so incredibly wrong ... it should not have
ever happened... but it did and that caused all sorts of bad things to
happen to animators. this guards against that and tries to get the
system clock time and if that doesnt work it just takes last time +
0.901.

@fix
2021-02-03 00:43:20 +00:00
Carsten Haitzler 6f803ad233 ecore x - allow vsync animator to delay by some fraction of a frame
@feat
2021-02-01 18:06:43 +00:00
Carsten Haitzler 8cb57ecc1b ecore x - fix typo in internal backdoor api to get vsync debug 2020-09-04 14:42:07 +01:00
Carsten Haitzler cef058c48b ecore x - go back to vsync thread. this will be more reliable
while it requires an extra wakeup from vsync thread -> e and this adds
a bit of jitter, it's mroe reliable in the face of "missed a frame"
timings and doesnt degrade as badly.
2020-08-20 08:52:32 +01:00
Carsten Haitzler 7308011cff ecore-x vysnc - env var - fix typo 2020-06-23 14:01:11 +01:00
Carsten Haitzler 4a88067b2e ecore-x : move x vsync device back to mainloop from thread
after i noticed some jitters, it seesms one thread to listen for vsync
events then to wake up the mainloop can suffer from irregular
scheduling jitters. it seems to be highly depenedent on both gpu
driver and cpu but it seemed the vsync thread itself was more reliably
woken than the mainlooop it then signalled, so merging this back is
just batter. it's configurable via an environment variable so we can
try either right now and see, but default is now to be in main loop.
2020-05-23 09:34:57 +01:00
Carsten Haitzler 859faad17b ecore-x expose some debug times like the exact time the vsync woke
this is for finding and eliminating jank in the vsync scheduling etc.
- as its underscore it's clearly a private api.
2020-05-23 09:34:57 +01:00
Carsten Haitzler 9768a5ad98 ecore x - fix debug printf format to print time 2020-02-14 09:00:48 +00:00
Carsten Haitzler d67e9875c8 ecore_x - add nouveau into the whitelist and clean up so it's a list
it was a lot of if cases before now it's an array with min version
parameters and globs for matching drivers etc. - much cleaner and
neater to afdd things to the whtielist now.
2018-10-11 16:25:05 +01:00
Carsten Haitzler 18cccd8798 ecore x - fix vsync to not block amdgpu drivers
amdgpu vsync works so don't block is so those drivers get proper
vsynced rendering as well.

@fix
2018-07-24 13:09:55 +09:00
Gustavo Sverzut Barbieri 010b14bc37 ecore_x_vsync: not using Ecore_Con.h anymore. 2016-11-21 19:47:35 -02:00
Carsten Haitzler 4ed2e01591 remove xcb support in ecore_x and evas engines as per mailing list
as per mailing list discussion about dropping xcb support now. it
hasn't been complete for a long time, thus not recommented for being
turned on. as we are moving to a wayland world xcbmakes even less
sense. as agreed, time to clean up a bit and remove a distraction as
well as not well tested code. this also updates po's too.

@feature
2016-11-03 22:22:54 +09:00