Stream: brlcad

Topic: subprocess issue


view this post on Zulip Sean (Dec 27 2020 at 06:55):

@starseeker ran into a new bug on mac. did an rt, closed the window, and then went to quit mged... got:
Screen-Shot-2020-12-27-at-1.01.30-AM.png

view this post on Zulip Sean (Dec 27 2020 at 06:55):

Details: Screen-Shot-2020-12-27-at-1.01.41-AM.png

view this post on Zulip Sean (Dec 27 2020 at 06:56):

investigating on command line: Screen-Shot-2020-12-27-at-1.02.10-AM.png

view this post on Zulip Sean (Dec 27 2020 at 06:57):

I didn't check if the process exited, but the window had been closed, so it certainly should have exited. I ran "rtabort" and then was able to exit.

view this post on Zulip Sean (Dec 27 2020 at 07:00):

So couple critical issues there. 1) the obvious one, that rt probably didn't exit cleanly when the framebuffer was closed, and 2) that that prevented exiting mged (imho nothing should prevent exiting as mged doesn't have a concept of unsaved changes).

view this post on Zulip starseeker (Dec 27 2020 at 14:11):

OK. The refusal to quit is r77990, and it actually functioned exactly as intended. I'm OK to revert ahead of release while we discuss, but that behavior was/is intended to avoid users accidentally aborting long running background jobs (say, for example, we get facetize running as a subprocess.)

view this post on Zulip starseeker (Dec 27 2020 at 14:12):

Or even the MGED rt being run with output directed to a .png file - if launched from MGED the user might forget it is running, and may accidentally quit and lose the render.

view this post on Zulip starseeker (Dec 27 2020 at 14:14):

I added the -f option to allow users to force quitting anyway, but at least that way they're knowingly wiping out anything running unseen in the background.


Last updated: Oct 09 2024 at 00:44 UTC