miniblog.

← Back to all posts
Oct 17, 2021 at 00:35
Interesting idea from a discussion around limiting rustc error output: why don't compilers use a pager when there are a lot of errors? Git uses this for large output and it works well there. (origin: https://github.com/rust-lang/rust/issues/27189#issuecomment-291420407)
Add `rustc --fatal` which stops at the first build error · Issue #27189 · rust-lang/rust
Add `rustc --fatal` which stops at the first build error · Issue #27189 · rust-lang/rust
(Should this be an RFC?) Motivation When refactoring, one may have dozens or hundreds of build errors which need to be fixed one by one. Sometimes, a build error is only a consequence of an earlier...

Related Posts

Aug 27, 2023 at 03:53
Implementing interactive languages, and the tradeoffs of interpreters, JIT compilers, and AOT compilers:
Implementing interactive languages
Jan 13, 2023 at 01:26
Interpreters versus compilers as a spectrum, and a fun worked example of optimising Brainfuck:
Laurence Tratt: Compiled and Interpreted Languages: Two Ways of Saying Tomato
May 22, 2022 at 23:14
It's strange how AOT compilers usually have different modes for optimising time or space usage, but JITs rarely have modes like "prefer fast startup". The vast majority of JITs seem to focus on runtime of hot code after warming up.