| 15 Mar 2026 |
 dramforever | seems precarious | 05:52:55 |
| dramforever | i have no idea why on earth would this suddenly be a problem now | 05:54:31 |
 emily | I think we enabled hardening for strict flex arrays at some point | 05:58:18 |
| emily | though I don't think "this cycle" recent | 05:58:24 |
| emily | https://github.com/NixOS/nixpkgs/pull/473648 | 05:58:47 |
| dramforever | clang update? | 06:00:06 |
| dramforever | oooh wait | 06:00:31 |
| dramforever | it is this cycle recent | 06:00:35 |
| dramforever | well i think that just settles it? | 06:01:23 |
| emily | how are you meant to strict flex arrays in C++ if it doesn't have them | 06:02:17 |
| dramforever | "gnu extension" | 06:02:28 |
| emily | or did they add them | 06:02:36 |
| dramforever | actually i have no idea why it seems to work in both gcc and clang | 06:05:01 |
 Vladimír Čunát | What do you mean? | 06:07:01 |
| Vladimír Čunát | clang implements basically all "gnu extensions" | 06:07:23 |
| Vladimír Čunát | IIRC they're very compatible with gcc. | 06:07:38 |
| Vladimír Čunát | Anyway, the best workaround is to disable this particular hardening in the chromium build? | 06:09:48 |
| dramforever | sorry i meant it doesn't seem to be documented anywhere for gcc the FAM is accepted in c++ | 06:10:49 |
| dramforever | * sorry i meant it doesn't seem to be documented anywhere for gcc that FAM is accepted in c++ | 06:10:53 |
| dramforever | also yes i think we just disable that for chromium | 06:11:05 |
| dramforever | 
Download Screenshot_20260315_141012.png | 06:11:15 |
| dramforever | i know this is windows but chromium is definitely not -fsfa clean | 06:11:26 |
| Vladimír Čunát | [1] should be fine | 06:11:27 |
| Vladimír Čunát | * [1] should be fine | 06:11:34 |
| dramforever | oh wait | 06:11:35 |
| dramforever | good point | 06:11:36 |
| Vladimír Čunát | For strictflexarrays1 which is the default now. | 06:11:48 |
| Vladimír Čunát | And [1] is surely fine without any C++ extensions. | 06:12:12 |
| dramforever | that is a good point but i don't want to audit the entire chromium source code to see if that 8 really was the only place it explodes | 06:12:15 |
| dramforever | * that is a good point but i don't want to audit the entire chromium source code to see if that single char[sizeof(uint64_t)] really was the only place it explodes | 06:12:30 |
