bjoernager/rust
bjoernager/rust
1
Fork 0
Code Issues Pull requests Activity

Fix a comment.

Browse source 
Imagine you have replace ranges (2..20,X) and (5..15,Y), and these tokens:
```
a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x
```
If we replace (5..15,Y) first, then (2..20,X) we get this sequence
```
a,b,c,d,e,Y,_,_,_,_,_,_,_,_,_,p,q,r,s,t,u,v,w,x
a,b,X,_,_,_,_,_,_,_,_,_,_,_,_,_,_,_,_,_,u,v,w,x
```
which is what we want.

If we do it in the other order, we get this:
```
a,b,X,_,_,_,_,_,_,_,_,_,_,_,_,p,q,r,s,t,u,v,w,x
a,b,X,_,_,Y,_,_,_,_,_,_,_,_,_,_,_,_,_,_,u,v,w,x
```
which is wrong. So it's true that we need the `.rev()` but the comment
is wrong about why.
This commit is contained in:
Nicholas Nethercote 2024-07-26 14:02:15 +10:00
parent a560810a69
commit 6e87858f26
1 changed files with 3 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

6
compiler/rustc_parse/src/parser/attr_wrapper.rs
View file

@ -137,9 +137,9 @@ impl ToAttrTokenStream for LazyAttrTokenStreamImpl {
// `#[cfg(FALSE)] struct Foo { #[cfg(FALSE)] field: bool }`
//
// By starting processing from the replace range with the greatest
// start position, we ensure that any replace range which encloses
// another replace range will capture the *replaced* tokens for the inner
// range, not the original tokens.
// start position, we ensure that any (outer) replace range which
// encloses another (inner) replace range will fully overwrite the
// inner range's replacement.
for (range, target) in replace_ranges.into_iter().rev() {
assert!(!range.is_empty(), "Cannot replace an empty range: {range:?}");