#openssl

Ryan around the Web

• This is like… a decade old, so no, I unfortunately have no clue lol. It’s like three laptops ago which is effectively three lifetimes ago.

  On Tue, Sep 19, 2023 at 08:22, NeverEnoughRakia @.***(mailto:On Tue, Sep 19, 2023 at 08:22, NeverEnoughRakia < wrote:

  That said, this is a 'fork' of the most recent (haha) version found on the internets, with a ton of patches from various people around the internet cobbled together. Hello, In case you had bookmarked or you still remember from where you took the patches, could you please post a list of them? I guess some of their authors even commented them, for e.g. they might have made them for specific reasons you might have omitted.

  — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***> on Consider posting from where you took the patches?

  a week ago
• Hmm… I’ll try to find time to take a look at this over the weekend. This wound up being more invasive than I thought (which isn’t your fault, I think I sent you down that path) so want to regroup and figure out the best option. It may end up being your original take, not sure yet.

  And yeah ownership of the ObjC object is something I want to try and hide from this all, since it can get thorny quickly - it’s what led me to the handler approach originally.

  Def appreciate the efforts tho and we’ll figure out the best way to do this!

  On Fri, Sep 15, 2023 at 20:51, Isaac-Leonard @.***(mailto:On Fri, Sep 15, 2023 at 20:51, Isaac-Leonard < wrote:

  Okay, so I think I've implemented everything now that's needed to use Layout as a dye trait. I had to create a new trait for the set_frame method as I didn't really know how that could work without generics, I updated the frame example with the change. I had to change how ObjcAccess works too as the get_from_backing_obj method couldn't work without the generic return type. Instead I've made a new method that just returns the reference to the inner Id that callers can then just dereference and do what they were doing before. I'm not sure if I've be putting Shared or Owned for the ownership generic of Id though, for now I've just got Owned. I've also not had any experience with objective C so I'm not sure if I'm creating layout anchors correctly,. I'm creating them the same way its done in the constructor for view but not sure if this creates non-allowed duplicates or anything. I believe all of these changes may allow me to implement a more reactive framework on top of cacao too now.

  — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***> on Add trait to get layout anchors for a view

  2 weeks ago
• Offhand I dunno - but if that's deprecated the replacement does the same thing, we can certainly just swap the key out. on Add trait to get layout anchors for a view 2 weeks ago
• Hmm, interesting.

  FYI this might be a thing you'll need to rebase since we merged the objc2 PR recently. Apologies for the hassle but it should overall be better for the library going forward. :( on Add trait to get layout anchors for a view 2 weeks ago

• Note the change in API for Image::draw, which is unfortunate. It's not a strictly a needed change, but I wanted to avoid the direct dependency on core-graphics since it locks the public API to the specific version cacao depends on, and is used so little.

  Hmm, you're specifically referring to not needing CGContextRef? I see what you mean, though it is definitely unfortunate... but then again, if there were to be other spots that use CGContextRef, it might make sense to bite the bullet and at the very least feature-flag it.

  How often does core-graphics change anyway? It might not be a big deal to have it as a dependency. on Replace core-foundation and core-graphics 2 weeks ago

• Yup, no rush! on Initial conversion to objc2 2 weeks ago
• Pushed 1 commit to @ryanmcgrath/cacao 2 weeks ago
• Closed a pull request in @ryanmcgrath/cacao:

  Bump actions/checkout from 3 to 4 2 weeks ago

• So I think this is a gray area for me... it's definitely getting outside of what AppKit/UIKit offer by default, but cacao's taken the opinionated position before so I could be convinced to support this.

  I would say a few changes come to mind tho:

  • The layout anchor getters could probably be safely thrown onto the Layout trait without much issue, to reduce trait importing (mental) overhead.

  • top_to_bottom would be improved (IMO) with the following api:

  pub struct ColumnLayout {
      pub spacing: usize // psuedocode, do whatever type here
  }
  
  pub fn column<I, C>(parent: &Layout, layout: ColumnLayout, children: I)
  where
      I:  IntoIterator<Item = &C>,
      C: Layout,
  {
      // Manage adding views, setting spacing, etc
  }
  

  This feels more declarative in nature to me, e.g I could then write:

  impl MyGenericView {
      pub fn build(&self) {
          column(&self.view_handle, ColumnLayout {
              spacing: 8
          }, [
              &self.child_view_1,
              &self.child_view_2,
              // etc
          ]);
      }
  }
  

  The above is all pseudocode and not tested but hopefully showcases the idea.

  fill_safe_area I'm less enthused about and might want that in a contrib or helpers module, but I'll muse on that one for a bit. on Add trait to get layout anchors for a view 2 weeks ago

• Merged. :)

  Thank you again for the sheer amount of work here, as well as for the wider ObjC work you're doing in the Rust community. It's really great that someone's picked up the torch and focused on making it all safer and more reliable overall.

  I'll let this sit in trunk for a bit to see how people find it - whether we need to fix any bugs or anything - but then if all seems fine I'll eye pushing out the next (-beta, maybe) release that uses this. on Initial conversion to objc2 2 weeks ago

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

  Initial conversion to objc2 2 weeks ago