@jarredsumner@saltyAom In that case this PR does way too much. It practically doesn’t mention LSAN or fixing any of the things reported by ASAN/LSAN.
This should isolated introduce the proper ASAN allocator selection as failing.
The follow ups should fix the issues.
Make ASAN builds/tests separate.
@jarredsumner@saltyAom Huh? Are you shipping asan builds? Why would this have any impact on shipping builds?!
Unless of course this changes way more, in which case, why?
@SheriefFYI This might be controversial
For applications I write and are just to produce a single executable I have found using zig as C/C++ build system to be surprisingly nice.
For libraries the same just make it output a file to link to and a directory of header files.
@SheriefFYI Perfect, looking forward to what comes out of your work.
You saying “is a must” makes me laugh because most people commenting on this have no clue what that even means
@SheriefFYI OMG. I actually skimmed most of the connected threads. The early parts of “just add interpolation” damn, I have done my fair share of adding multiplayer to single player games and no, that’s just not the answer, even coming from that point.
@SheriefFYI Is this pure peer to peer or is there are server authoritative thing in the middle that just simulates inputs (which is also not as easy as it sounds as you’d want to “predict” client side and then send corrections etc?
Wanted to give @useblacksmith a try on some private projects, got greeted with abuse prevention, I'm out.
I logged in with my private GitHub account that has existed for 10+ years.
I am in no mood to send an email to use TRY product, too much friction.
@rflaherty71 Big focus was making sure it runs great instead of chasing the most graphical fidelity!
I knew many people who did amazing things to optimize it, and we aren’t even done yet
@theo I get it, it might be pretty decent for TypeScript/React stuff, but for anything C, C++, Rust, Zig etc. I have had nothing but terrible experiences.
@theo Honestly, I've tried; many times. It's proven to be entirely useless to the problems I deal with.
Makes sense, they generally aren't trained on those problems.
I don't expect it to be smarter, but I do expect more from it, it's not junior level for MY problems.
I have recently discovered that you can detect “/permissive-“ at compile time in the preprocessor on MSVC.
Yes, it’s annoying when you have to support permissive on and permissive off.
But hey, now I can conditionally compile things 😅
@NotNordgaren@Jonathan_Blow I’m missing the context here to say 100%, but I have seen similar things like that with varargs before and that also matches how varargs are implemented.