Post B4FihlWHT7dUn46CSO by diegovsky@bolha.us
 (DIR) More posts by diegovsky@bolha.us
 (DIR) Post #B4Fhwjcqh2FnJ0kGqu by thephd@pony.social
       2 likes, 2 repeats
       
       Second half of the February-March C meeting is over.Surprisingly, a few things I expected to get spiked into the dirt made it through, and a lot of it was inversely proportional to the amount of effort I put in compared to the first half of the meeting:defer TS approval to finally publish (after I fix all the godawful typos/thinkos)const integers are really Constant Expressions (e.g. usable in static assert and shit)Progress was made on the _Any_func* paper, but in order for it to proceed I have to remove everything related to "representation affinities" and leave the actual bit representation -- and whether they convert cleanly with memcpy or otherwise -- as a black box instead. This readily takes my most important use case and shoots it in the back of the head but, whatever. Something something compromise. Varying levels of Tentative Support for everything else, though apparently there's serious objections to printing an _Any_func* which seems... like a very bizarre part of the thing to be against. Maybe i can do a better job of convincing people of that next time.Thread Attributes made forward progress. People keep trying to kill off the wide string and UTF-8 string versions of the "set name". I don't know why they keep doing that, because that just means certain platforms (not just Windows, but CERTAINLY Windows) will be utterly fucked. We did agree to keep THRD_NAME_*_MAX, where * is either nothing or SIZE.
       
 (DIR) Post #B4FiGeQ7kutGxBI2ym by thephd@pony.social
       2 likes, 1 repeats
       
       For the papers I wasn't responsible for, most saw reasonable forward progress except for Dependent Type Improvements (Variably-Modified Types -- pointers to VLAs, basically -- inside of structs with additional syntax for runtime sizes in those pointed-to VLAs). _Maxof and _Minof got spiked since the important part of the feature was actually _Widthof and that was removed for a variety of reasons (apparently it's controversial to expect the return type to be like everything else that's an operator like this and just return size_t? I think someone wants it to be int. I don't know why we want to start carving out those sorts of sheaningans).I might need to pick up a paper for doing Unicode printing.
       
 (DIR) Post #B4FihlWHT7dUn46CSO by diegovsky@bolha.us
       0 likes, 0 repeats
       
       @thephd what about the closure paper? not sure if that is the any func pointer thing
       
 (DIR) Post #B4FiqykXeJK0QeA9Sa by thephd@pony.social
       0 likes, 0 repeats
       
       @diegovsky It isn't. The closure paper was seen during the first half of the meeting, and none of the solutions have solidified as a clear winner.
       
 (DIR) Post #B4Fj0fSEFAmDTtDVr6 by diegovsky@bolha.us
       0 likes, 0 repeats
       
       @thephd what? the final ones were so good, especially the replacement for gnu's named inner functions