3 ms·
Fedora made the switch in F40. Arch has it as a todo. https://fedoraproject.org/wiki/Changes/ZlibNGTransition https://fedoraproject.org/wiki/Changes/ZlibNGTran
by notanote 2y ago
Fedora made the switch in F40. Arch has it as a todo.
https://fedoraproject.org/wiki/Changes/ZlibNGTransition https://fedoraproject.org/wiki/Changes/ZlibNGTransition
https://archlinux.org/todo/zlib-ng-migration/ https://archlinux.org/todo/zlib-ng-migration/
- masklinn 2y agoThe just released git 2.49 also has experimental support for -ng.
- bravetraveler 2y agoSmall example of the user experience https://github.com/zlib-ng/zlib-ng/issues/1708 https://github.com/zlib-ng/zlib-ng/issues/1708
- ahepp 2y agoRust wouldn't prevent this would it? Since the root cause is a caller violating an invariant?
- cwillu 2y agoRust caused it. More charitably, unnecessary churn caused it.
- bravetraveler 2y ago100% churn, IMO. Language didn't matter at all. In the end, the failure was procedural. This lasted well into the next Fedora release. That means... this endured for at least six months. The reporter caught it in beta, but Fedora and 'zlib-ng' marched on. Not saved until Mono/Unity caught up. This didn't have to be hamfisted, but... t'was. Performance gains are make-believe when consumers don't work. Fedora should have waited for 41 after finding issues on 40. Distributions should do more to curate the software they distribute. This technical aspect was reported - and ignored - in the early stages during the 'sentiment gauging' phase on their own forum. Not just the 'zlib-ng' repository. https://discussion.fedoraproject.org/t/f40-change-proposal-transitioning-to-zlib-ng-as-a-compatible-replacement-for-zlib-system-wide/95807 https://discussion.fedoraproject.org/t/f40-change-proposal-t... It's more important to be seen changing things than doing it well, apparently. The proposal wasn't required to go as it did.
- ahepp 2y agowe can always blame it on closed source software!