Today we worked on finishing up work on constant deprecation. I’ve already covered all of the constants I had in mind; All *MAX constants, as well as all *NUM constants.
Yesterday’s efforts on the former group has finally culminated in a final PR (meaning it stopped being a draft PR.) This meant a bunch more research into both kernel and standard library implementation source code, though I won’t go into any details in this report. See rust-lang/libc#5122 for notes on each and every constant (or set of constants). Of note is that I couldn’t get my hands on the QNX Neutrino SDK so I could not have a look into the header files for that one target. I did get to deprecate the constants that I realized matched those values POSIX defines as non-runtime invariant.
The latter group of constants was fairly easy to go through. I did short work of it, and another PR with those batched changes is ready.
At this point, I would like to ask for some input by Trevor Gross, as I believe I have went through most if not all constants with an immediately obvious SemVer-breakage issue.
None at present.
This week’s goal has been met, and I’ve even gotten to finish up what I expected would take me all of next week. This does not mean that I can just move on to the next part of the proposal, but it does mean I can start work on it. I believe next week I will get some response back from other libc maintainers on comments I made across the different PRs I submitted concerned with deprecation work. Either way, I will start working on the bit-width transition issues again starting tomorrow (following from the changes I already made prior to GSoC.)