Public
- Public
- Network
- Groups
- Popular
- People

Conversation

Notices

Masaki Hara (qnighy@qnmd.info)'s status on Tuesday, 04-Dec-2018 20:23:36 JST Masaki Hara

https://internals.rust-lang.org/t/homu-queue-woes-and-suggestions-on-how-to-fix-them/8954
RustコンパイラはPRをmasterにマージする直前に、マージ結果のコミットに対してテストをしている。そのためこの本番テストは(投機的に実行しない限り)直列に実行せざるを得ない。このCIは現在約3時間かかり、qnighyもビルドキューで1週間待たされたことがある。
In conversation Tuesday, 04-Dec-2018 20:23:36 JST from qnmd.info permalink
Attachments
1. Domain not in remote thumbnail source whitelist: internals.rust-lang.org
  
  Homu queue woes, and suggestions on how to fix them
  
  from Rust Internals
  
  A successful PR landing cycle for the compiler now takes over 3 hours. This means that roughly 7.5 PRs can land per 24 hours, which often isn’t enough to keep up with current demand. Also, intermittent failures are common enough that this maximum of 7.5 often isn’t reached. As a result, PR landing times have been very long and unpredictable lately, and roll-ups have been used more, which has risks. (Have you ever tracked a performance regression down to a roll-up with 20+ PRs in it? It’s frustr...

Feeds