Auto merge of #143882 - cjgillot:relative-span-file, r=<try>

Also hash spans inside the same file as relative.

r? `@ghost` for perf

Author: bors

compiler/rustc_span/src/lib.rs

@@ -2660,12 +2660,12 @@ where
         }
 
         if let Some(parent) = span.parent {
-            let def_span = ctx.def_span(parent).data_untracked();
-            if def_span.contains(span) {
+            let parent_span = ctx.def_span(parent).data_untracked();
+            if parent_span.contains(span) {
                 // This span is enclosed in a definition: only hash the relative position.
                 Hash::hash(&TAG_RELATIVE_SPAN, hasher);
-                (span.lo - def_span.lo).to_u32().hash_stable(ctx, hasher);
-                (span.hi - def_span.lo).to_u32().hash_stable(ctx, hasher);
+                Hash::hash(&(span.lo - parent_span.lo), hasher);
+                Hash::hash(&(span.hi - parent_span.lo), hasher);
                 return;
             }
         }
@@ -2682,6 +2682,25 @@ where
         Hash::hash(&TAG_VALID_SPAN, hasher);
         Hash::hash(&file, hasher);
 
+        if let Some(parent) = span.parent {
+            let parent_span = ctx.def_span(parent).data_untracked();
+            let Some((parent_file, ..)) = ctx.span_data_to_lines_and_cols(&parent_span) else {
+                Hash::hash(&TAG_INVALID_SPAN, hasher);
+                return;
+            };
+
+            if parent_file == file {
+                // This span is relative to another span in the same file,
+                // only hash the relative position.
+                Hash::hash(&TAG_RELATIVE_SPAN, hasher);
+                // Use signed difference as `span` may start before `parent_span`,
+                // for instance attributes start before their item's span.
+                Hash::hash(&(span.lo.to_u32() as isize - parent_span.lo.to_u32() as isize), hasher);
+                Hash::hash(&(span.hi.to_u32() as isize - parent_span.lo.to_u32() as isize), hasher);
+                return;
+            }
+        }
+
         // Hash both the length and the end location (line/column) of a span. If we
         // hash only the length, for example, then two otherwise equal spans with
         // different end locations will have the same hash. This can cause a problem