Auto merge of #76256 - tgnottingham:issue-74890, r=nikomatsakis

incr-comp: hash and serialize span end line/column

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 during incremental compilation wherein a previous result for a
query that depends on the end location of a span will be incorrectly
reused when the end location of the span it depends on has changed. A
similar analysis applies if some query depends specifically on the
length of the span, but we only hash the end location. So hash both.

Fix #46744, fix #59954, fix #63161, fix #73640, fix #73967, fix #74890, fix #75900

---

See #74890 for a more in-depth analysis.

I haven't thought about what other problems this root cause could be responsible for. Please let me know if anything springs to mind. I believe the issue has existed since the inception of incremental compilation.

1 files changed, 26 insertions, 5 deletions

diff --git a/compiler/rustc_span/src/lib.rs b/compiler/rustc_span/src/lib.rs
index f52b64f4883..0926561f4c5 100644
--- a/compiler/rustc_span/src/lib.rs
+++ b/compiler/rustc_span/src/lib.rs
@@ -1894,16 +1894,37 @@ where
             return;
         }
 
+        let (_, line_hi, col_hi) = match ctx.byte_pos_to_line_and_col(span.hi) {
+            Some(pos) => pos,
+            None => {
+                Hash::hash(&TAG_INVALID_SPAN, hasher);
+                span.ctxt.hash_stable(ctx, hasher);
+                return;
+            }
+        };
+
         Hash::hash(&TAG_VALID_SPAN, hasher);
         // We truncate the stable ID hash and line and column numbers. The chances
         // of causing a collision this way should be minimal.
         Hash::hash(&(file_lo.name_hash as u64), hasher);
 
-        let col = (col_lo.0 as u64) & 0xFF;
-        let line = ((line_lo as u64) & 0xFF_FF_FF) << 8;
-        let len = ((span.hi - span.lo).0 as u64) << 32;
-        let line_col_len = col | line | len;
-        Hash::hash(&line_col_len, hasher);
+        // 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
+        // during incremental compilation wherein a previous result for a query that
+        // depends on the end location of a span will be incorrectly reused when the
+        // end location of the span it depends on has changed (see issue #74890). A
+        // similar analysis applies if some query depends specifically on the length
+        // of the span, but we only hash the end location. So hash both.
+
+        let col_lo_trunc = (col_lo.0 as u64) & 0xFF;
+        let line_lo_trunc = ((line_lo as u64) & 0xFF_FF_FF) << 8;
+        let col_hi_trunc = (col_hi.0 as u64) & 0xFF << 32;
+        let line_hi_trunc = ((line_hi as u64) & 0xFF_FF_FF) << 40;
+        let col_line = col_lo_trunc | line_lo_trunc | col_hi_trunc | line_hi_trunc;
+        let len = (span.hi - span.lo).0;
+        Hash::hash(&col_line, hasher);
+        Hash::hash(&len, hasher);
         span.ctxt.hash_stable(ctx, hasher);
     }
 }