Performance regression after rendering refactor #361
Comments
p8
changed the title
jonathanhefner
added a commit
to jonathanhefner/sdoc
that referenced
this issue
Jan 17, 2024
In d08ca4d, an isolated scope was created for each rendered file. That inadvertently reset memoized values from `git` calls for each file, causing a performance regression. This commit fixes the regression by memoizing the values globally. With d08ca4d~1 (444c4a4): ```console $ time bundle exec rake test:rails real 0m31.567s user 0m31.159s sys 0m0.394s ``` With d08ca4d: ```console $ time bundle exec rake test:rails real 1m11.195s user 1m2.671s sys 0m8.811s ``` With d08ca4d + this commit (retrofitted): ```console $ time bundle exec rake test:rails real 0m26.155s user 0m25.796s sys 0m0.331s ``` With `main`: ```console $ time bundle exec rake test:rails real 1m16.000s user 1m7.478s sys 0m8.818s ``` With `main` + this commit: ```console $ time bundle exec rake test:rails real 0m29.964s user 0m29.562s sys 0m0.399s ``` Fixes rails#361.
|
Thank you for finding and bisecting this! ❤️ |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Generating the docs is multiple factors slower on main.
I've bisected it to the following commit:
d08ca4d758f2633b9a2a49a045d56d4856c2e5e4.#312
The text was updated successfully, but these errors were encountered: