Opened 13 years ago
Closed 13 years ago
#9139 closed defect (Fixed)
video rendered tilted 45 degrees when running mythfrontend with -w and an odd number of horizontal pixels.
Reported by: | Owned by: | markk | |
---|---|---|---|
Priority: | minor | Milestone: | 0.24.1 |
Component: | MythTV - Video Playback | Version: | 0.24-fixes |
Severity: | medium | Keywords: | |
Cc: | danielk@… | Ticket locked: | no |
Description
Not a big deal, but running mythfrontent -w -geometry 711x400 (or any other odd numbered width, height can be odd w/o causing problems) causes the video playback to be rendered tilted 45 degrees to the right, so you get a split in the video starting top-right going down into the screen at a 45 degree angle). I'm seeing this on a test system where I only have software rendering (vnc server), so not a big deal, but I figured this may come up in other situations where it might actually matter.
This used to work in 0.23, broken in 0.24.
Change History (4)
comment:1 Changed 13 years ago by
Cc: | danielk@… added |
---|---|
Component: | MythTV - General → MythTV - Video Playback |
Owner: | set to markk |
Status: | new → assigned |
comment:2 Changed 13 years ago by
I'm seeing this with both the Qt and the OpenGL paint engines (just tested), but I typically use the Qt painter here (as there is no graphics hardware at all in this case, just a dumb vnc X server). I have not seen any segfaults due to this that I can recall.
Which video render is this happening with?
This can occur when some element of the video pipeline assumes a direct relationship between the width and the stride or when a buffer allocation is not rounded up to the nearest dimension as required by the codec or the buffer's color model. Generally when such an error is made we end up painting past the end of the buffer which can cause segfaults, but we add protective padding at the end of our buffer allocations which should avoid such a segfault in most cases.