ndk: Add bindings for AFont

[paxbun]

This PR provides bindings for AFont, AFontMatcher, and ASystemFontIterator and some helper structs for the better abstraction of some parts of the OpenType specification.

Example:

for font in ndk::font::SystemFontIterator::new().unwrap() {
  log::debug!("Font: {}", font.path().display());
  for idx in 0..font.axis_count() {
    log::debug!("  Axis {idx} ({}): {}", font.axis_tag_at(idx), font.axis_value_at(idx));
  }
  log::debug!("  Collection Idx: {}", font.collection_index());
  log::debug!("  Locale: {:?}", font.locale());
  log::debug!("  Weight: {}", font.weight());
  log::debug!("  Is Italic: {}", font.is_italic());
}

Result (Tested on Galaxy Tab S8):

[paxbun]

TODOs

• [x] Reflect PR number to CHANGELOG.md
• [x] Reflect NDK version (I used 25.1.8937393) to CHANGELOG.md and Cargo.toml or re-generate ndk-sys bindings

[MarijnS95]

Nice! I'll pull in the bindings regeneration via #370, and then have a look at this.

[MarijnS95]

@paxbun just a quick heads-up that I haven't been able to catch up to this before leaving for a while. I hope that's not a problem and I'll do my very best to pick this up in about 2 weekes, apologies for all the delays :crossed_fingers:

[MarijnS95]

By the way, a friendly request to not reply Fixed! to every review comment in what is hopefully the last round of changes so that we can merge this in. My notification inbox is filled with those messages while lacking context, and GitHub uses the "Resolve conversation" button for that.

[MarijnS95]

@paxbun any updates or should I apply the remaining comments/remarks/suggestions and merge this? I'd like to get the "breaking" changes imposed by the ndk-sys regen in and publish all these crates soon-ish™.

[paxbun]

@MarijnS95 I think I mostly applied your suggestion, and I also checked whether there are any ffi::* functions not wrapped. I also added constructors that constly panic to AxisTag.

[MarijnS95]

Note that I reported the header issues at https://github.com/android/ndk/issues/1920, because surface_control.h (which I need and am writing a wrapper for) has the same (and many more C++-like) issues.

Hopefully we can drop the workarounds when upstream picks them up.

[paxbun]

I applied all the suggestions you mentioned! (except for the axis tag one)