00:14 <gbruno> [github] kintel closed pull request #5189 (Stop installing qt5-default on Debian based OS's) https://github.com/openscad/openscad/pull/5189

00:14 <gbruno> [github] kintel pushed 2 modifications (Stop installing qt5-default on Debian based OS's (#5189) Only still available on Ubuntu 20.04 which has neither Qt 4 or Qt 6.) https://github.com/openscad/openscad/commit/b3070a2aa98a2ba97d945631c55b9471e525aa2c

11:31 <mohammad1722> Hey Everyone! So, last week I added the submission parser script and a GitHub Action for the library-manager repo.

11:31 <mohammad1722> Basically these are the automated handlers of people's pull requests that add/remove libraries from the list.

11:31 <mohammad1722> The expected user behavior is:

11:31 <mohammad1722> 1. Fork the library-manager repo

11:31 <mohammad1722> 2. Update the list file (output_files/list_file.txt) by adding or removing library links (links to the library's repository).

11:31 <mohammad1722> 3. Commit the changes, push them to the fork, and open a pull request (PR).

11:31 <mohammad1722> The PR triggers the GitHub Action, which parses the diff to identify changes. If the diff involves only the list file, the Action evaluates the changes (addition or deletion) and automatically accepts or rejects the PR.

11:31 <mohammad1722> After a successful merge, the bot automatically updates the output_files/accepted_repositories.txt file to reflect the newly added libraries. This step is necessary to determine the library name per each library repo, as each library must be checked later that each release has the same name.

11:31 <mohammad1722> Currently, I've tested adding libraries and the parsing seems working well. However, the automatic merge of the PR fails. Maybe this is a permission issue?

11:31 <mohammad1722> Future improvement here could be:

11:31 <mohammad1722> - When the PR is rejected, the bot should comment with the rejection reason and maybe instructions to fix the submission.

11:31 <mohammad1722> - Don't allow anyone to remove any library and have a human review that first.

11:31 <mohammad1722> My next steps:

11:31 <mohammad1722> - Check why exactly the merge step fails

11:31 <mohammad1722> - Develop the other automated part which is a periodic GitHub Action that check the accepted repositories file regularly to see if any of the repos has a new release and then produce an index_file which contains links to each version of each library available.

12:53 <kintel> mohammad1722: In terms of automatically merging PRs, are you using GitHub's auto merge feature?

12:53 <kintel> https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/incorporating-changes-from-a-pull-request/automatically-merging-a-pull-request

12:53 <kintel> If yes, that needs to be enabled on the Repo first

12:54 <kintel> https://github.com/openscad/openscad-library-manager/settings\

12:54 <kintel> (I've never used it myself so I'm not familiar with how it works)

13:37 <mohammad1722> kintel: I am using the GitHub API (specifically octokit/request-action@v2.x) to do the merging

13:37 <mohammad1722> IDK about automerge too, but at first glance I don't think it does what I want

16:31 <gbruno> [github] cjmayo opened pull request #5190 (Fix installed version numbers reported by check-dependencies.sh) https://github.com/openscad/openscad/pull/5190

16:34 <t4nk_fn> *sigh* .. I thought "let's see if there's some bugs in openscad I could look into" .. so I saw some bug about not being able to return to preview mode with f9 or something,

16:34 <t4nk_fn> then decided to rebuild my openscad

16:35 <t4nk_fn> ... now I'm facing error while loading shared libraries: libpolygon.so

16:35 <t4nk_fn> does that have something to do with clipper2?

16:36 <teepee> no, I think that's a recent change in manifold, compiling parts as separate library

16:36 <t4nk_fn> he

16:36 <t4nk_fn> heh, I was just compiling manifold.

16:36 <teepee> it may depend itself on clipper2 though

16:37 <teepee> not sure about that

16:37 <t4nk_fn> this was exactly one of the dangers of me installing them as separate packages... you need to take care of updating them yourself :|

16:37 <t4nk_fn> .. yeah, it's fixed now

17:01 <gbruno> [github] kintel pushed 1 modifications (Fix installed version numbers reported by check-dependencies.sh (#5190) Was reporting: opencsg 1.4.2 1.6. OK fontconfig 2.10 2.15. OK Now: opencsg 1.4.2 1.6.0 OK fontconfig 2.10 2.15.0 OK) https://github.com/openscad/openscad/commit/9eb020255f5563b31047bd367abdf611f8d3a951

17:01 <gbruno> [github] kintel closed pull request #5190 (Fix installed version numbers reported by check-dependencies.sh) https://github.com/openscad/openscad/pull/5190

17:41 <kintel> t4nk_fn Lots of things that can be picked up. Are you mostly interested in Qt-related topics or are you open to everything?

17:42 <kintel> Qt/Gui stuff is "harder" in that there are no automated tests, so manual testing effort across platforms is more work

17:43 <kintel> ..but otoh, non-gui stuff means establishing automated tests ;)

17:47 <t4nk_fn> I'm not much at anything, really, though I'd say I'm most familiar with Qt stuff, like that's why I picked up on the signal thing last time

17:52 <teepee> there is certainly lots of issues and feature requests for the GUI, both big and small

19:53 <t4nk_fn> heh

19:55 <t4nk_fn> yeah, MainWindow.cc calls qglview->hasOpenCSGSupport() before it had a chance to determine that, since it is done in GLView::enable_opencsg_shaders

19:57 <t4nk_fn> and later on in MainWindow::viewModePreview() => viewActionPreview->setChecked(true); ... but it was not enabled in the first place

20:08 <t4nk_fn> seems pretty messy all in all ;)

20:09 <t4nk_fn> I'm no programmer, but I'd say it oughtta be pulled out of enable_opencsg_shaders perhaps, to solve it in a more decent way

20:10 <t4nk_fn> another way would be to just enable it before you setChecked(true)

20:11 <t4nk_fn> I never used theres F9-F12 keys before, but I think there are some other issues there aswell

20:18 <t4nk_fn> did I just write 'theres'? ;)

20:18 <t4nk_fn> them there!

20:38 <teepee> I was never clear how the combination of those view options work anyway, so maybe it's worth stepping back a bit more and see what makes sense :)

21:03 <t4nk_fn> wireframe mode would seem nice, albeit as a rather useless plaything

23:02 <kintel> t4nk_fn Once you move from GUI into visualization, things also gets trickier as there is a lot of legacy code there, which should be cleaned up first. ..so wireframe mode done right is likely a good chunk of work, and wireframe mode as a hack might do more damage than good as the next cleanup just becomes harder.

23:06 <kintel> Also, in terms of OpenCSG, we require OpenGL >=2 now, so OpenCSG should always be available

23:06 <kintel> See https://github.com/openscad/openscad/pull/4550

23:07 <kintel> ..so a bunch of the checks you mention could simply be ripped out