webui: Saving webui-desktop log to anaconda.log #5734
Conversation
adamkankovsky
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
62ccfff
to
cc4b7af
Compare
adamkankovsky
had a problem deploying
to
gh-cockpituous
GitHub Actions
Failure
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
cc4b7af
to
e818a9c
Compare
adamkankovsky
had a problem deploying
to
gh-cockpituous
GitHub Actions
Failure
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
e818a9c
to
f7359bc
Compare
adamkankovsky
had a problem deploying
to
gh-cockpituous
GitHub Actions
Failure
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
f7359bc
to
0121d18
Compare
adamkankovsky
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
0121d18
to
b23a75b
Compare
adamkankovsky
had a problem deploying
to
gh-cockpituous
GitHub Actions
Failure
|
Removing my review after discussion with @adamkankovsky. He pointed me to a fact that |
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
b23a75b
to
ff106f2
Compare
adamkankovsky
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
adamkankovsky
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
adamkankovsky
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
8d52554
to
0dfb4f4
Compare
adamkankovsky
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
|
/build-image --live |
|
Images built based on commit 0dfb4f4:
Download the images from the bottom of the job status page. |
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
0dfb4f4
to
ad37262
Compare
adamkankovsky
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
|
@jkonecny12 I tested this on boot.iso and it works well. |
| f.write(repr(proc.pid)) | ||
| proc.wait() | ||
| try: | ||
| proc = startProgram( |
There was a problem hiding this comment.
Why not use stdout and stderr parameters of the startProgram?
There was a problem hiding this comment.
Because its easier to use the subprocess.PIPE as default and proc.comunicate() for me. But i can change it to IO string buffer as well.
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
ad37262
to
b146ee5
Compare
adamkankovsky
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
b146ee5
to
ce2830a
Compare
adamkankovsky
had a problem deploying
to
gh-cockpituous
GitHub Actions
Failure
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
ce2830a
to
847454e
Compare
adamkankovsky
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
847454e
to
29a0379
Compare
adamkankovsky
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
29a0379
to
4111757
Compare
adamkankovsky
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
# Conflicts: # pyanaconda/ui/webui/__init__.py
adamkankovsky
force-pushed
the
webui-desktop-log-update
branch
from
4111757
to
1ad5a5d
Compare
adamkankovsky
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
| with open(self._viewer_pid_file, "w") as f: | ||
| f.write(repr(proc.pid)) | ||
|
|
||
| (output_string, err_string) = proc.communicate() |
There was a problem hiding this comment.
The issue I'm seeing (if I read it correctly) is that you are blocking completely this thread. If this thread is the main thread (which I guess it is) it will also block event loop so no events are processed. That could have a problematic outcomes like not being able to react on exceptions and react on DBus events.
Please verify if my assumptions are correct or wrong.
There was a problem hiding this comment.
There shouldn't be any difference from the original code, there was a wait in the same place, which blocked the kernel until the webui was finished.
The BOSS who takes care of this should have been running in a parallel process long ago and this should not affect him in any way.
There was a problem hiding this comment.
I somehow remember that I said even back then that this is not correct way to handle it. But if it was already part of the original code I guess it should be fine here....
|
/kickstart-test --testtype smoke |
No description provided.