Don't use guess_head_span in predicates_of for foreign span

Previously, the result of `predicates_of` for a foreign trait
would depend on the *current* state of the corresponding source
file in the foreign crate. This could lead to ICEs during incremental
compilation, since the on-disk contents of the upstream source file
could potentially change without the upstream crate being recompiled.

Additionally, this ensure that that the metadata we produce for a crate
only depends on its *compiled* upstream dependencies (e.g an rlib or
rmeta file), *not* the current on-disk state of the upstream crate
source files.
This commit is contained in:
Aaron Hill 2021-08-27 21:13:08 -05:00
parent ac50a53359
commit c9157efad6
No known key found for this signature in database
GPG key ID: B4087E510E98B164
9 changed files with 97 additions and 15 deletions

View file

@ -567,6 +567,17 @@ impl SourceMap {
}
}
/// Returns whether or not this span points into a file
/// in the current crate. This may be `false` for spans
/// produced by a macro expansion, or for spans associated
/// with the definition of an item in a foreign crate
pub fn is_local_span(&self, sp: Span) -> bool {
let local_begin = self.lookup_byte_offset(sp.lo());
let local_end = self.lookup_byte_offset(sp.hi());
// This might be a weird span that covers multiple files
local_begin.sf.src.is_some() && local_end.sf.src.is_some()
}
/// Returns the source snippet as `String` corresponding to the given `Span`.
pub fn span_to_snippet(&self, sp: Span) -> Result<String, SpanSnippetError> {
self.span_to_source(sp, |src, start_index, end_index| {