Today work was focused on more PR reviews. Testing attempts on Fuchsia have also resumed.
The Windows PRs seem solved. The PR concerning linking the right routines is done. That one needed some reorganizing. I just cherry picked a custom patch over one of the commits and applied it to the other Windows PR. This other PR also seems done. It needed simplifying some code in the build script.
The Fuchsia PR also seems done. Small fixes were needed. Some polyfills for C macros were exposing implementation details. That’s fixed. Then a bunch of fields were public when they probably shouldn’t. You can’t really tell in C without opaque types so we’re taking a guess here. The only other thing left is to test the changes out. That’s what I’ve started doing.
Testing on Fuchsia seems straightforward with FEMU. FEMU is their fork of QEMU. The Rust Platform Support page has instructions on the expected package format. It should hopefully be simple. This is a WIP.
The newlib PR got some modifications. The aarch64 module got removed. There’s just no supported Rust target with that triple combination. Other changes include being more stringent in the type definitions. Conditionally defined file offset types now give a compilation error if the target is not supported.
Work on the aix module is pending.
None at present. Testing on MIPS targets with uClibc is pending. These are all tier 3 targets. This is low priority. Input would be welcome.
Today the report has been submitted earlier than usual. I’ve got some stuff to deal with later on.
The expected timeline is being met. The unix module is almost done. All PRs concerned with this are also almost done. The Windows PRs are right around the corner. Reviews are mostly done. The one pain point is the Linux uClibc PR. Hopefully that gets resolved soon. The aix module and other child modules to the unix module should be done by next week. The hard cap in the proposal was week 7. Things look well. But something’s surely bound to come up.