While finding it incredibly annoying with my NexImage 5 (2.2 micron pixels) I was looking around and found that instead of reading something from v4l2 (as it doesn't seem possible) INDI defaults to 5.6 micron pixels with no mention of size.
I added and tested both a Celestron NexImage 5 and Logitech Webcam 9000 Pro which I use via v4l2 with a check and assigning the correct pixel size. I'm going to submit a pull request (with reference to this thread), but if anyone has better ideas (as a database or other store would probably be better) I'd love to hear them. Idea is that we'll probably need to do it manually, unless someone has a good way to find the pixel sizes on chips, so it'll print out the value, and people can look it up and post it here for now. (Again, not super clean, but it'll certainly help lessen annoyance for people who do it once it's in.)
If the pull request/idea is rejected, then that's fine!
Example of messages that people will get:
[INFO] For future autodetection of pixel size, please report the following: Reported Name: 'NexImage5, Common Name (Eg: NexImage 10), Pixel Size to the following thread: www.indilib.org/forum/ccds-dslrs/4392-in...detection.html#33485
[INFO] Setting pixel size to default of 5.6
Example of reports (use commas for non-square pixels):
Reported Name: 'NexImage5', Common Name: NexImage 5, Pixel Size: 2.2
Reported Name: 'UVC Camera (046d:0809)', Common Name: Logitech Webcam Pro 9000, Pixel Size 3.3
When ever i go to adjust any of the properties in the indi control panel i get the following errors, but this only started to happen the last couple of days, before that it was working perfectly.
2019-01-15T19:59:03: Driver indi_v4l2_ccd: indi_v4l2_ccd dispatch error: Property Image Adjustments is not defined in V4L2 CCD.
2019-01-15T20:00:18: Driver indi_v4l2_ccd: indi_v4l2_ccd dispatch error: Property CCD_FRAME is not defined in V4L2 CCD.
2019-01-15T20:00:23: Driver indi_v4l2_ccd: indi_v4l2_ccd dispatch error: Property CCD_BINNING is not defined in V4L2 CCD.
2019-01-15T20:00:28: Driver indi_v4l2_ccd: indi_v4l2_ccd dispatch error: Property CCD_COMPRESSION is not defined in V4L2 CCD.
2019-01-15T20:00:32: Driver indi_v4l2_ccd: indi_v4l2_ccd dispatch error: Property CCD_FRAME_TYPE is not defined in V4L2 CCD.
2019-01-15T20:00:33: Driver indi_v4l2_ccd: indi_v4l2_ccd dispatch error: Property CCD_FRAME_RESET is not defined in V4L2 CCD.
2019-01-15T20:00:35: Driver indi_v4l2_ccd: indi_v4l2_ccd dispatch error: Property CCD_COLOR_SPACE is not defined in V4L2 CCD.
I just tested the current head GIT V4L2 driver with the web-camera of my laptop and it seemed to work fine so hopefully it's just a dodgy build in the PPA, the log looks a bit like the driver crashed, there might be more info in the indiserver output.
Thanks, yes it crashes and is restarted. If you can run it under gdb (instructions are somewhere on this forum, can't browse with this phone unfortunately) that would give where the crash happens. It does get pretty far in initialization though.
I've had some instability with V4L2 sometimes. (Before and after this patch) I haven't tracked it down.
Nothing the patch does should have caused that unless there's an assumption about pixel size. (Which wouldn't be the problem in the logs above, since it went with the default. So as far as the code, the only difference is the message. )
I appreciate your work, and just pushed a few more, that I don't know if they work, and added a tested option, so that things can be added, and if they work, they can be marked as tested.
INDI as far as I know doesn't have an easy way to get that information, or any surveys. I know I am ambivalent on them, but maybe add an option in Kstars to generate a report of data that could be posted somewhere could be something looked at? Because sometimes it's hard to tell if people are using something or if it was working years ago, but got broken. (Any way it's done, it's imperfect.)