IRC channel logs

<kpcyrd>sam_: how come people care tho? if I made my own proprietary unix I wouldn't expect to be able to sway anything

<sam_>the guy has been around in the git community a pretty long time and isn't just a rando

<sam_>i guess that's why

<sam_>also just that git sort of relies on this ubiquity and so anything which affects that matters

<kpcyrd>I guess 🤷 other projects approach this by having optional rust implementations of certain features, like "use the basic C implementation or opt into some fancy optimized Rust implementation" (that happens to be memory safe too)

<kpcyrd>git is something you really want to be memory safe, but the gix alternative is kinda going overboard with the amount of libraries they split their code into

<sam_>right, that part was discussed in the article

<unmush>I am pleased to report that I have bootstrapped my way to mono-6.12.0.206, which is the last tagged release I could find from https://gitlab.winehq.org/mono/mono.git

<unmush>it helped a LOT to just cherry pick commits touching mcs/mcs and apply them to an older version

<unmush>the full path I used: pnet-git@0.8.0-0.3baf947 -> mono@1.2.6 -> mono@1.9.1 -> mono@2.4.2.3 -> mono@2.6.4 -> mono@2.11.4 -> mono@3.0.12 -> mono@3.12.1 -> mono@4.9.0 -> mono@5.0.1 -> mono@5.1.0 -> mono@5.2.0.224 -> mono@5.4.0.212 -> mono@5.4.0.201-0.d0f51b4 -> mono@5.8.0.129 -> mono@5.8.0.129-0.3e9d7d6 -> mono@5.10.0.179 -> mono@6.12.0.206

<unmush>it seems like standard practice in mono was to start depending on a feature in their standard library, then several months later get around to implementing it in their compiler

<unmush>this laziness being possible because after the microsoft acquisition they added a git submodule called roslyn-binaries and copied those as-is in the install step

IRC channel logs

2024-11-25.log