[RFC] Enable usage of rust-sccache for cargo builds #41617
Conversation
| x86_64*|i686*|arm*|aarch64*) hostmakedepends+=" rust-sccache" ;; | ||
| *) ;; | ||
| esac | ||
| fi |
There was a problem hiding this comment.
I really don't like implicit build-style dependencies, especially with logic like this.
There was a problem hiding this comment.
I agree it's not the cleanest. But isn't it explicit since XBPS_SCCACHE is opt-in?
There was a problem hiding this comment.
With ppc support being dropped from void, don't need the arch checking logic - so this is as "clean" as it's going to be.
|
Pull Requests become stale 90 days after last activity and are closed 14 days after that. If this pull request is still relevant bump it or assign it. |
|
Pull Requests become stale 90 days after last activity and are closed 14 days after that. If this pull request is still relevant bump it or assign it. |
|
Pull Requests become stale 90 days after last activity and are closed 14 days after that. If this pull request is still relevant bump it or assign it. |
icp1994
force-pushed
the
enable-sccache
branch
from
ca7095b
to
cf39655
Compare
|
Pull Requests become stale 90 days after last activity and are closed 14 days after that. If this pull request is still relevant bump it or assign it. |
icp1994
force-pushed
the
enable-sccache
branch
from
cf39655
to
ab97e5a
Compare
|
Pull Requests become stale 90 days after last activity and are closed 14 days after that. If this pull request is still relevant bump it or assign it. |
|
Pull Requests become stale 90 days after last activity and are closed 14 days after that. If this pull request is still relevant bump it or assign it. |
icp1994
force-pushed
the
enable-sccache
branch
from
ddee717
to
39fd2ae
Compare
Reviving #30702 - I tested the changes in this PR: YES
I have been using this for almost a week and it has reduced build times for rust packages quite considerably in my laptop. I have addressed some of the questions/suggestions from the reviews of the previous PR.
Although not the most scientific benchmark, I noted the build times from the cargo output while packaging 4 templates across iterations. In the first two iterations, all of them ran without sccache with an older version of the template to get a baseline without the interference of the crates being downloaded. Third iteration was with same old versions but with sccache enabled which created the cache of all 4 packages combined. Finally, last iteration with the full cache and the latest version of the templates - a typical scenario while updating a package. The difference between the second and the fourth iterations are as follows:
I will add docs for README later if there is consensus to merge this eventually.