cancel @cancel@merveilles.town

@hamcha custom QStyle engine will let you change pretty much anything (sizing, shapes, colors, whatever), but it's quite a bit of work to make one

@hamcha you can make custom color schemes (if you didn't already know) https://dev.cancel.fm/wiki?name=Custom_Themes

If you want to make an entirely custom skin/theme you'll have to implement a Qt QStyle. here's the one I wrote for Ripcord https://github.com/randrew/phantomstyle you can use it as a starting point

Or you can look around on the internet for existing QStyle engines (though most of them only work on Linux)

You'll need to link it against Qt 5.9 and drop it into the plugins dir, then RIPCORD_STYLE_ENGINE=myengine

@neauoire ah sorry, i woke up just a few minutes ago, haha. my sleep schedule is fubar right now.

@neauoire @npisanti *shakes fist*

@npisanti @neauoire i don't have a strong opinion about this one, though. if you guys think it's better to default to a different seed every time, that's fine.

@npisanti my personal preference would be to not have any randomness in orca itself at all, and just default the seed to 1 unless the user specifies it

@npisanti ah i thought you meant you would just run orca like

orca --seed `myrandom`

or something

@npisanti yes that would be an even better solution

@npisanti there's no reason at all to use it... seconds is fine.

@npisanti PID is a bad source of randomness. on some systems it will be the same number if the user quits and runs the process again. depends on kernel and how it's configured. if it's run in a vm container or sandbox it may be 0 every time, which would also make the seconds useless.

@npisanti why PID?

@npisanti @neauoire (though you'd init the random seed from the current time and only use the call to random once, so it's kind of an unnecessary step)

@npisanti @neauoire you could also just use the random functions from libc

@npisanti @neauoire surely you don't need better precision than one second for the pseudorandom number? the PID and other stuff just adds unnecessary complication... https://godbolt.org/z/5Tx_2M

@npisanti @neauoire well if i'm out-voted on it being enabled by default, at least it shouldn't be locked behind editing a source file to re-enable it. i'd rather not be patronizing with regards to how to use the software.

@npisanti @neauoire i don't know what that function was supposed to do... i have no idea how to fix it

@npisanti @neauoire i couldn't even run orca-c in debug mode because it aborted immediately due to overflow

@npisanti @neauoire it should be on by default. i don't see a reason to disable a feature that was already working and people may have been using. i would be pissed off if someone removed a feature from software i was using with no explanation.

@npisanti @neauoire I reverted it because there was no way to re-enable it without editing a source file. Can you change it to a run-time parameter or a `tool` parameter? I thought the JS version had no mouse support simply because it wasn't implemented.

@npisanti @neauoire why is the mouse disabled now in orca-c?