This report got out of hand quite badly. It has been left untouched from when it was first written, but note that by the time I wrote it (at the end of the work day, as usual) I was more tired than I should’ve been to appropriately structure my thoughts without exposing an overly pathetic perspective about myself in public.
Today work was quite miserable; Solving a ton of Git conflicts on stuff that I had finished working on between yesterday and this morning.
I started off by finishing up the changes on musl and uClibc concerning siginfo_t. This was not the hard part, though. I then started working on changes to the Git history, which is where I took the longest time.
As I went through the changes, new stuff starting popping out; Chiefly, impl blocks that I had forgotten to include in some files. I then decided it would be best if all commits provided a working revision that actually built.
That took me along another fairly large session of Git rebase conflicts to get a single commit all the way down to the start of the patchset. The details are not really worth it.
The last thing I did was to separate the patches that handled actually adding the bindings to the new module from those that remove them from the old but still prevalent module structure. This took the longest, by far.
I first started off by, just in case, creating new branches off of the original patchset (pre-reorganizing stuff.) Then I dropped those patches from the PR’s branch, and conflicts starting popping in that rebase session.
This took a while to solve and affected all other patches stacked on top of the now dropped initial patchset. Then I had more conflicts from the one commit I decided on putting at the start of the patchset to make all revisions build.
After solving all of that, I started on actually splitting the changes from the patch that had the new organization and the reexports at new/mod.rs. That had a few more conflicts, but those were expected.
All along this, my editor starts experiencing bugs and I have to cherry pick unmerged PRs from upstream to get a custom build that lets me keep working. This wasn’t that time-consuming (10 minutes + 5 minute build) but it was a waste.
Next thing I do is notice the alarm; It’s time for the daily report, and I only got one PR done today (not even a new PR.)
Beyond dangerously useless days like today, none.
Today was catastrophic, and I’m not sure anymore that I can finish tomorrow the small-to-medium work I mentioned yesterday. If I do, then the plan remains unchanged because I’ll then start looking into the MCP implementation.
If I don’t (and this is now very likely) then I will have to extend that period for one more day. This gets us to Friday. It leaves two days for the MCP implementation.
I honestly think that’s doable, but days like this make me doubt.