#python

Ryan around the Web

• This is likely the same (ish) issue as #122 - probably a thing that was being properly retained before is not being retained now.

  I have to find time to dig in and poke at this stuff, but I've been fairly busy as of late - but the last non-beta version should theoretically be fine if you're finding this a blocker. Appreciate the heads up otherwise! on todos_list Example Program Crashes in Release Mode 2 weeks ago

• …what thing are you packaging that depends on this?

  I guess it could be removed if it’s a formality - if it’s there it’s probably a decade or so old and I’d have no context beyond that lol

  On Sun, Aug 11, 2024 at 22:27, Paul Pfeister @.***(mailto:On Sun, Aug 11, 2024 at 22:27, Paul Pfeister < wrote:

  Hey @.(https://github.com/ryanmcgrath)> on Remove nose-cov dependency?

  3 weeks ago

• which unless I am mistaken

  I do not believe you are. :)

  A PR is welcome, otherwise I'll get to it in the next batch of fixes once I find the time. Thanks for the callout! on AttributedString::set_text_color uses range end as a length a month ago

• Huh, it's likely a simple-ish fix of just changing the crate name to use underscores for consistency. It's not like it gets uploaded or distributed outside the project.

  I'll see about poking at this over the weekend since the way Windows finds the lib is a bit custom.

  (I don't suppose there's a way to force the Cargo 1.78 behavior anywhere, is there?) on Build breaks after Cargo 1.79 a month ago

• I don't know of any particular way to do that and I'm not really inclined to support it, since it's an ultra-niche use case.

  Someone is free to PR it, but with DriverKit being the way it is I feel like it'd be tricky to develop outside of this org - but perhaps with a properly disabled SIP + systemextensionctl setup it's possible.

  You might be able to get away with using just gcusbadapter.kext from the hjelmn project (with SIP kext checks disabled only, don't disable all of SIP necessarily), since GCAdapterDriver handles the part that gcusbadapter_raw.kext does. kexts are not guaranteed to work forever though, so... YMMV.

  on Expose GC Controllers to system as Gamepads? 2 months ago

• Well then.

  Thanks for catching that, heh. Merged! on fix: typo in window’s set_subtitle fn 2 months ago

• Pushed 1 commit to @ryanmcgrath/cacao 2 months ago
• Closed a pull request in @ryanmcgrath/cacao:

  fix: typo in window’s set_subtitle fn 2 months ago

• Hmmm, might be down to a difference in obj/objc2. @madsmtm might know more, otherwise I'd have to find time to dig in - and I'll be busy + traveling the next two weeks.

  If you wanted to tinker and debug, a webview not loading sounds like something not getting retained somewhere - either the webview itself, or the webview's delegate. It's where I'd start looking at least. on Browser Example Broken 2 months ago

• Hmmm, are you saying that the webview isn't loading anything, or that back doesn't work?

  And what version of macOS are you testing against...? on Browser Example Broken 2 months ago

• Do you have more info on what you're trying to accomplish? e.g is this for a bare CALayer you own or a CALayer backing an NSView?

  Since the objc property of CALayer is exposed publicly, you can message it directly with objc calls (e.g let _: () = msg_send![my_view.layer.id, surface];). You of course need to handle the lifetimes and safety yourself, but this is one of those territories where it's the preferred way of doing it at the moment.

  If you wind up finding a more comfortable API and/or you want to buff the CALayer wrapper to be more comprehensive, a pull request is always welcome. I wouldn't pull IOSurface into this framework necessarily, similar to how Metal isn't really in here either (though other crates exist for it). In general I just try to make it possible for crates to interop via the objc layer if possible.

  I considered making a channel on Matrix at one point, but the level of activity here isn't so significant that it's needed. on missing CALayer.contents -> Any? 2 months ago

• Opened a pull request in @vitaly-m/postgis-diesel:

  Remove default features, no 32-col tables by default 2 months ago