-
Notifications
You must be signed in to change notification settings - Fork 12.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
fix sysroot issue which appears for ci downloaded rustc #104076
Conversation
r? @jyn514 (rustbot has picked a reviewer for you, use r? to override) |
b0b8196
to
1cf5fe6
Compare
I don't think this is the correct fix. It breaks I think a more fundamental issue is that we're using sysroots as both an input and output; ci-rustc-sysroot is an input for building rustdoc and an output for placing the new rustdoc binary. We should be using |
I played around with this a bit more - this new code may or may not be on the right track, but I don't have more time to dedicate to it.
|
Will cover it, thanks |
Can you clarify |
@ozkanonur I don't remember off the top of my head, but I think it was trying to copy a file to a sysroot directory that didn't exist, or something like that. I'd expect you to be able to replicate the error locally. |
Somehow it doesn't happen on my side. configprofile = "tools"
changelog-seen = 2
[build]
compiler-docs = true
[rust]
deny-warnings = false
download-rustc = true terminal output with printing some logs
|
5e51704
to
241bab9
Compare
src/bootstrap/compile.rs
Outdated
for stage in 0..=builder.top_stage { | ||
if stage != compiler.stage { | ||
let dir = sysroot_dir(stage); | ||
if !dir.ends_with("ci-rustc-sysroot") { | ||
let _ = fs::remove_dir_all(dir); | ||
} | ||
} | ||
} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@jyn514 I couldn't repreduce the stage 1 issue you mentioned, but still added this to ensure current sysroots will not be deleted.
ps: highlighted code is updated, builder.top_stage
in 0..=builder.top_stage
replaced with 2
Signed-off-by: ozkanonur <work@onurozkan.dev>
c6cff01
to
654a4e8
Compare
Thanks. I am still a little unsure this is the right approach, but it seems strictly better than the current situation. @bors r+ |
🌲 The tree is currently closed for pull requests below priority 1. This pull request will be tested once the tree is reopened. |
…iaskrgr Rollup of 10 pull requests Successful merges: - rust-lang#103117 (Use `IsTerminal` in place of `atty`) - rust-lang#103969 (Partial support for running UI tests with `download-rustc`) - rust-lang#103989 (Fix build of std for thumbv7a-pc-windows-msvc) - rust-lang#104076 (fix sysroot issue which appears for ci downloaded rustc) - rust-lang#104469 (Make "long type" printing type aware and trim types in E0275) - rust-lang#104497 (detect () to avoid redundant <> suggestion for type) - rust-lang#104577 (Don't focus on notable trait parent when hiding it) - rust-lang#104587 (Update cargo) - rust-lang#104593 (Improve spans for RPITIT object-safety errors) - rust-lang#104604 (Migrate top buttons style to CSS variables) Failed merges: r? `@ghost` `@rustbot` modify labels: rollup
Currently when compiler is downloaded rather than compiled, sysroot is being
ci-rustc-sysroot
because ofrust/src/bootstrap/compile.rs
Lines 1125 to 1131 in 7eef946
And rustdoc is overriding the downloaded one at the end of the process.
With the condition I add, we simply check if the current compiler stage is target build stage, if so use the proper sysroot instead of
ci-rustc-sysroot
.Resolves #103206