add headless flag for CLI execution (2)#62
add headless flag for CLI execution (2)#62sarumaj wants to merge 1 commit intopixelomer:mainfrom sarumaj:main
Conversation
|
It is failing due to changes in submodules |
There was a problem hiding this comment.
There was no reason to update this dependency. There are breaking changes and the changes are completely unrelated to this pull request.
https://github.com/pixelomer/Shijima-Qt/commits/d904504183581512c859a368d9b260053fda0373/
There was a problem hiding this comment.
No reason to update this dependency either, at least not within the context of this pull request.
There was a problem hiding this comment.
I revoked the dependency updates.
| auto manager = ShijimaManager::defaultManager(); | ||
| if (!headless) { | ||
| ShijimaManager::defaultManager()->show(); | ||
| manager->show(); | ||
| } |
There was a problem hiding this comment.
This change does make the HTTP server run, but also causes the shimeji windows to have no parent. This causes all of them to show up as separate icons in the bottom bar, at least on KDE Plasma.
There was a problem hiding this comment.
Ok, I give up then. The design would have to change to separate concerns (GUI/backend operability). Would be really cool though if you happen to restructure Shijima a bit and make the headless execution possible. Thanks!
|
Have you tried running or even just compiling your changes before submitting this pull request or the previous one? |
Not yet. Adapting the pipelines to run in my forked repository scope would be too much effort (you hard-coded references to you own repository there), I will have to build it locally. |
|
Feature request handed over to @pixelomer. |
2 participants
Hello, I would like be able to run Shijima-Qt in headless mode without the need to display the manager window. This is my personal proposal how we could achieve it. It might as well solve the issue #48, which in particular requests just this behavior. What did I change?
Since, the HTTP server is started by ShijimaManager, this time we allow the manager to be created, we just obstruct its visibility (opposingly to PR #61).