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.

4 files changed, 48 insertions, 0 deletions

diff --git a/src/test/run-make/incr-prev-body-beyond-eof/Makefile b/src/test/run-make/incr-prev-body-beyond-eof/Makefile
new file mode 100644
index 00000000000..49a7ee5f900
--- /dev/null
+++ b/src/test/run-make/incr-prev-body-beyond-eof/Makefile
@@ -0,0 +1,19 @@
+include ../../run-make-fulldeps/tools.mk
+
+# FIXME https://github.com/rust-lang/rust/issues/78911
+# ignore-32bit wrong/no cross compiler and sometimes we pass wrong gcc args (-m64)
+
+# Tests that we don't ICE during incremental compilation after modifying a
+# function span such that its previous end line exceeds the number of lines
+# in the new file, but its start line/column and length remain the same.
+
+SRC=$(TMPDIR)/src
+INCR=$(TMPDIR)/incr
+
+all:
+	mkdir $(SRC)
+	mkdir $(INCR)
+	cp a.rs $(SRC)/main.rs
+	$(RUSTC) -C incremental=$(INCR) $(SRC)/main.rs
+	cp b.rs $(SRC)/main.rs
+	$(RUSTC) -C incremental=$(INCR) $(SRC)/main.rs
diff --git a/src/test/run-make/incr-prev-body-beyond-eof/a.rs b/src/test/run-make/incr-prev-body-beyond-eof/a.rs
new file mode 100644
index 00000000000..ca70fb56334
--- /dev/null
+++ b/src/test/run-make/incr-prev-body-beyond-eof/a.rs
@@ -0,0 +1,16 @@
+fn main() {
+    // foo must be used.
+    foo();
+}
+
+// For this test to operate correctly, foo's body must start on exactly the same
+// line and column and have the exact same length in bytes in a.rs and b.rs. In
+// a.rs, the body must end on a line number which does not exist in b.rs.
+// Basically, avoid modifying this file, including adding or removing whitespace!
+fn foo() {
+    assert_eq!(1, 1);
+
+
+
+
+}
diff --git a/src/test/run-make/incr-prev-body-beyond-eof/b.rs b/src/test/run-make/incr-prev-body-beyond-eof/b.rs
new file mode 100644
index 00000000000..a272e44a632
--- /dev/null
+++ b/src/test/run-make/incr-prev-body-beyond-eof/b.rs
@@ -0,0 +1,12 @@
+fn main() {
+    // foo must be used.
+    foo();
+}
+
+// For this test to operate correctly, foo's body must start on exactly the same
+// line and column and have the exact same length in bytes in a.rs and b.rs. In
+// a.rs, the body must end on a line number which does not exist in b.rs.
+// Basically, avoid modifying this file, including adding or removing whitespace!
+fn foo() {
+    assert_eq!(1, 1);////
+}
diff --git a/src/test/run-make/issue-36710/Makefile b/src/test/run-make/issue-36710/Makefile
index b7bf366c918..b0e8451ff5d 100644
--- a/src/test/run-make/issue-36710/Makefile
+++ b/src/test/run-make/issue-36710/Makefile
@@ -1,5 +1,6 @@
 include ../../run-make-fulldeps/tools.mk
 
+# FIXME https://github.com/rust-lang/rust/issues/78911
 # ignore-32bit wrong/no cross compiler and sometimes we pass wrong gcc args (-m64)
 
 all: foo