This took me quite a bit longer to write up than I thought it would: https://github.com/nikomatsakis/project-ffi-unwind/pull/17
I don't know whether this is a good model for future discussion-summaries. I had planned to write something like this fairly regularly (and writing it weekly would certainly make it easier, since the discussions would be fresher in my mind). But it's a lot of work, and it's unclear what the level of detail would be.
There were two additional issues: first, I don't understand everything that was discussed. For instance, I see that there's some concern about how
pthread error-handling interacts with
catch_unwind, but I didn't have time to fully investigate the details. Second, I don't know how successful I was in avoiding bias; for instance, I don't really understand @centril's concerns about letting
extern "C" functions unwind, so it's possible that that section in my summary isn't fairly presented.
I'm looking now
btw, that is always a challenge with summaries --
this is why, for contentious things, I generally prefer to write the sumamry in a hackmd / paper doc, and then send it to the other party
anyway, I did find it helpful to read that, but I can believe that it's too much work
I think going topic-by-topic rather than chronologically would help somewhat. I only went chronologically because I wanted to make sure I covered everything going back to around the time you told me that there was a lot of discussion you hadn't yet caught up on :laughing: