26e1ce7394
apparently I missed some tests in the last commit. Rather than having dozens of tests use the long version, use the short version in `run-make` and the long version in `run-make-fulldeps` (which is now only three tests)
38 lines
1.0 KiB
Makefile
38 lines
1.0 KiB
Makefile
include ../tools.mk
|
|
|
|
OUT=$(TMPDIR)/emit
|
|
|
|
# --emit KIND=PATH should not affect crate hash vs --emit KIND
|
|
all: $(OUT)/a/libfoo.rlib $(OUT)/b/libfoo.rlib $(OUT)/c/libfoo.rlib \
|
|
$(TMPDIR)/libfoo.rlib
|
|
$(RUSTC) -Zls $(TMPDIR)/libfoo.rlib > $(TMPDIR)/base.txt
|
|
$(RUSTC) -Zls $(OUT)/a/libfoo.rlib > $(TMPDIR)/a.txt
|
|
$(RUSTC) -Zls $(OUT)/b/libfoo.rlib > $(TMPDIR)/b.txt
|
|
$(RUSTC) -Zls $(OUT)/c/libfoo.rlib > $(TMPDIR)/c.txt
|
|
|
|
diff $(TMPDIR)/base.txt $(TMPDIR)/a.txt
|
|
diff $(TMPDIR)/base.txt $(TMPDIR)/b.txt
|
|
|
|
# Different KIND parameters do affect hash.
|
|
# diff exits 1 on difference, 2 on trouble
|
|
diff $(TMPDIR)/base.txt $(TMPDIR)/c.txt ; test "$$?" -eq 1
|
|
|
|
# Default output name
|
|
$(TMPDIR)/libfoo.rlib: foo.rs
|
|
$(RUSTC) --emit link foo.rs
|
|
|
|
# Output named with -o
|
|
$(OUT)/a/libfoo.rlib: foo.rs
|
|
mkdir -p $(OUT)/a
|
|
$(RUSTC) --emit link -o $@ foo.rs
|
|
|
|
# Output named with KIND=PATH
|
|
$(OUT)/b/libfoo.rlib: foo.rs
|
|
mkdir -p $(OUT)/b
|
|
$(RUSTC) --emit link=$@ foo.rs
|
|
|
|
# Output multiple kinds
|
|
$(OUT)/c/libfoo.rlib: foo.rs
|
|
mkdir -p $(OUT)/c
|
|
$(RUSTC) --emit link=$@,metadata foo.rs
|