-
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
Delay ambiguous intra-doc link resolution after Cache
has been populated
#131691
Delay ambiguous intra-doc link resolution after Cache
has been populated
#131691
Conversation
Cache
has been populated
Some changes occurred in src/librustdoc/clean/types.rs cc @camelid |
/// These links are ambiguous. We need for the cache to have its paths filled. Unfortunately, | ||
/// if we run the `LinkCollector` pass after `Cache::populate`, a lot of items that we need | ||
/// to go through will be removed, making a lot of intra-doc links to not be inferred. | ||
/// | ||
/// So instead, we store the ambiguous links and we wait for cache paths to be filled before | ||
/// inferring them (if possible). |
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.
This is how I would write it.
According to
rustc_resolve
, these links are ambiguous.However, we cannot link to an item that has been stripped from the documentation. If all but one of the "possibilities" are stripped, then there is no real ambiguity. To determine if an ambiguity is real, we delay resolving them until after
Cache::populate
, then filter every item that doesn't have a cachedpath
.We could get correct results by simply delaying everything. This would have fewer happy codepaths, but we want to distinguish different kinds of error conditions, and this is easy to do by resolving links as soon as possible.
I actually might still try refactoring this to use a single happy codepath, if I can do it in a way that doesn't regress the diagnostics and actually reduces the total amount of code.
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.
Can be done as a follow-up but if you have ideas for improvements, it's very welcome!
} | ||
|
||
pub(crate) struct AmbiguousLinks { | ||
disambiguator: Option<Disambiguator>, |
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.
How could this possibly be Some
? Doesn't using a disambiguator imply that the link can't be ambiguous?
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.
One word: reexports. :)
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.
I'm still confused. Is there a test case for this?
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.
So with:
//! [struct@X]
mod a {
pub struct X;
}
mod b {
pub struct X;
}
pub use a::*;
pub use b::*;
We get:
warning: public documentation for `bar` links to private item `X`
--> bar.rs:1:6
|
1 | //! [struct@X]
| ^^^^^^^^ this item is private
|
= note: this link will resolve properly if you pass `--document-private-items`
= note: `#[warn(rustdoc::private_intra_doc_links)]` on by default
I think it's a bug but for now, let's simply remove this disambiguator
as it doesn't seem to work.
self.cx.tcx, | ||
BROKEN_INTRA_DOC_LINKS, | ||
format!( | ||
"all items matching `{path_str}` are either private or doc(hidden)" |
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.
"all items matching `{path_str}` are either private or doc(hidden)" | |
"all items matching `{path_str}` are private or doc(hidden)" |
@@ -2061,6 +2251,8 @@ fn ambiguity_error( | |||
// There is no way for users to disambiguate at this point, so better return the first | |||
// candidate and not show a warning. | |||
return false; | |||
} else if !emit_error { | |||
return true; |
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.
Can this first chunk be separated out into a has_ambiguity_error
function that ambiguity_error
simply calls?
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.
Good idea!
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.
Ah actually not really: we need the kinds
variable that is computed at the beginning...
Updated! |
This comment has been minimized.
This comment has been minimized.
e0fb0dc
to
10f2395
Compare
Updated! |
@bors ready |
@bors r+ rollup |
…er-out-2, r=notriddle Delay ambiguous intra-doc link resolution after `Cache` has been populated Fixes rust-lang#130233. I was getting nowhere with rust-lang#130278. I took a wrong turn at some point and ended making way too many changes so instead I started again back from 0 and this time it worked out as expected. r? `@notriddle`
…er-out-2, r=notriddle Delay ambiguous intra-doc link resolution after `Cache` has been populated Fixes rust-lang#130233. I was getting nowhere with rust-lang#130278. I took a wrong turn at some point and ended making way too many changes so instead I started again back from 0 and this time it worked out as expected. r? ``@notriddle``
Rollup of 10 pull requests Successful merges: - rust-lang#129181 (Pass end position of span through inline ASM cookie) - rust-lang#130989 (Don't check unsize goal in MIR validation when opaques remain) - rust-lang#131657 (Rustfmt `for<'a> async` correctly) - rust-lang#131691 (Delay ambiguous intra-doc link resolution after `Cache` has been populated) - rust-lang#131730 (Refactor some `core::fmt` macros) - rust-lang#131751 (Rename `can_coerce` to `may_coerce`, and then structurally resolve correctly in the probe) - rust-lang#131753 (Unify `secondary_span` and `swap_secondary_and_primary` args in `note_type_err`) - rust-lang#131776 (Emscripten: Xfail backtrace ui tests) - rust-lang#131777 (Fix trivially_copy_pass_by_ref in stable_mir) - rust-lang#131778 (Fix needless_lifetimes in stable_mir) r? `@ghost` `@rustbot` modify labels: rollup
…iaskrgr Rollup of 9 pull requests Successful merges: - rust-lang#130989 (Don't check unsize goal in MIR validation when opaques remain) - rust-lang#131657 (Rustfmt `for<'a> async` correctly) - rust-lang#131691 (Delay ambiguous intra-doc link resolution after `Cache` has been populated) - rust-lang#131730 (Refactor some `core::fmt` macros) - rust-lang#131751 (Rename `can_coerce` to `may_coerce`, and then structurally resolve correctly in the probe) - rust-lang#131753 (Unify `secondary_span` and `swap_secondary_and_primary` args in `note_type_err`) - rust-lang#131776 (Emscripten: Xfail backtrace ui tests) - rust-lang#131777 (Fix trivially_copy_pass_by_ref in stable_mir) - rust-lang#131778 (Fix needless_lifetimes in stable_mir) r? `@ghost` `@rustbot` modify labels: rollup
Rollup merge of rust-lang#131691 - GuillaumeGomez:intra-doc-link-filter-out-2, r=notriddle Delay ambiguous intra-doc link resolution after `Cache` has been populated Fixes rust-lang#130233. I was getting nowhere with rust-lang#130278. I took a wrong turn at some point and ended making way too many changes so instead I started again back from 0 and this time it worked out as expected. r? ```@notriddle```
Fixes #130233.
I was getting nowhere with #130278. I took a wrong turn at some point and ended making way too many changes so instead I started again back from 0 and this time it worked out as expected.
r? @notriddle