Closed fornellas closed 1 year ago
The Travis build failure in Debian is more of the same old
Compilation failed: Failed to close file descriptor for child process (Operation not permitted)
that we've had before. I don't know enough about docker or Travis to interact with that
@lukefromdc it was an oversight on my side, when cutting the PR, sorry: just fixed the *
to /
.
I just tested that but got no change in setting window-scale by autodetection on my monitor
Check the code, there are various exceptions to set things to 100%. I read through all of it, and checked what was failing for my case, and 2 conditions were failing (thus this PR). There may be some other left over corner case for you. As this PR is now, it is working for my case (re-tested).
The Travis build failure in Debian is more of the same old
Can't see how these could be related to this PR TBH, it seems to be happeing at other cases as well.
Not worried about the travis failure on Debian, that's a problem w the current Debian builds on all packages due to a problem w the base image as reported elsewhere
Not worried about the travis failure on Debian, that's a problem w the current Debian builds on all packages due to a problem w the base image as reported elsewhere
When using debian=latest
tag (debian 11) a build from master runs fine https://app.travis-ci.com/github/mate-desktop/mate-settings-daemon/builds/255422658, which is good.
When using fedora=rawhide tag (fedora >= 37) building fedora fails too, which is bad. https://app.travis-ci.com/github/mate-desktop/mate-settings-daemon/builds/255422798
Build errors are caused by a glibc update
https://gitlab.gnome.org/GNOME/glib/-/merge_requests/2435
and an update to docker policy seccomp
https://github.com/moby/moby/pull/42681/commits/9f6b562dd12ef7b1f9e2f8e6f2ab6477790a6594
More details you will find here:
https://gist.github.com/nathabonfim59/b088db8752673e1e7acace8806390242
I did a PR with a workaround to fix this build pain. Please review. https://github.com/mate-desktop/mate-dev-scripts/pull/36 This PR can be tested with this test branch https://github.com/mate-desktop/mate-settings-daemon/pull/392
As a side note, both commits should be squashed.
Having a fix for a fix in git history isn't nice.
You can do that with git rebase -i <last-commit-hash-before-PR>
Calling author of HIDPI in Mate for a review.
@vkareh
@raveit65 squashed. Not sure if desirable, but I usually configure my repositories to force all PRs to be squashed on merge, so there's no chance of putting "lots of intermediate commits" on master, and it leaves each PR author free to have their history.
Works for me but discussion needs to be solved.
Ditto. Closes #390.