Today work was focused mostly on reviewing and answering to the feedback I had gotten in the last few days to pending PRs and issues.
I reviewed feedback and tweaked the FreeBSD PR. This one still needs more work.
I reviewed the bot results from running tests on the MCP implementation PR. This was simple.
I solved some conflicts on one of the AIX PRs.
I worked some more on the first draft to my GSoC work product.
I looked some more into the L4Re PR.
I got the emscripten and Android PRs ready to merge after some final touches needed from recent feedback.
The FreeBSD PR concerning the netlink/netlink.h interfaces received some feedback on the changes I had made to potentially solve the issue without there being need for extending ctest.
The comments were mostly concerned with simplifying some logic in the build script. A recent patch got merged that should allow making shallow copies of the test generator instances produced by ctest.
That was used to avoid multiple test suite runs with differently configured test generator instances. The one thing that remains to be solved are the SemVer tests, which don’t pass just yet and may very well require extending.
Those are implemented inline in the build script itself, so it should be fairly straightforward to change. I also made sure to keep those changes in the initial patch I submitted to the PR.
If we do end up extending that small test of infrastructure, that part of the old patch should come in handy.
I then moved on to solving the UI tests in the MCP implementation. This was fairly simple to do but it did take quite a while because I thought I needed first to read through the relevant section in the rustc dev guide.
I did eventually end up back on the command line, where a simple --bless to x test adding the new openbsd7.9 value to the set of possible target_env values was enough to solve the test failures.
I then moved on to solving one of the AIX PRs merge conflicts. After recently splitting the original AIX PR into three separate patchsets, one of them got merged, and another one started having merge conflicts.
Then I worked on the GSoC work product report, on which I’ve continued making progress. Out of the three pages of GitHub issues and PRs I have to document, I’m now down to only two pages. There’s been a lot of small-time PRs.
Then I addressed some long-standing (3 days old) feedback I got from the target maintainer for L4Re in the corresponding PR. They confirmed that they don’t set _GNU_SOURCE in their makefiles but it’s still odd.
Them not setting it up in the L4Re build system doesn’t warrant that we do set it in the ctest run that compiles our C tests. Of course, I’m assuming here L4Re works like Linux in that respect, but I could very well be wrong.
Maybe the headers are pre-compiled in that target and setting up feature test macros just doesn’t work out the same way. I’ve got to look some more into this, but I’ve tagged the maintainer on the PR thread asking about this.
The last two things I did were to replace old deprecation attribute annotations in the Android and emscripten PRs for FIXME comments to actually add those deprecations once we reach that stage of the LFS plan.
Those two PRs should now be ready for merging.
None.
The MCP implementation should be ready now. The FCP for it is also nearing its end, and there does not seem to be any objections to it. Hopefully we can start using this soon and get the old OpenBSD issue closed.
In terms of specific tasks for this week, I know tomorrow I’ll start off by looking into the other AIX PRs, which were the ones that received the latest feedback. I should probably have time to get something else done tomorrow.
I’ll keep looking through the pending issues in the 1.0 milestone, and see into solving/closing those.