Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Dependency loops cause targets to be missing from ninja graph with --gen_all_targets #206

Open
danw opened this issue Aug 10, 2020 · 1 comment

Comments

@danw
Copy link
Collaborator

danw commented Aug 10, 2020

(From internal google bug 162918277)

This test case can show the problem:

testcase/circular_dep_self.mk

test:
	echo PASS

loop: loop test
	true
$ go test --ckati --ninja --all -test.run=TestKati/circular_dep_self.mk
--- FAIL: TestKati (0.00s)
    --- FAIL: TestKati/phony_loop.mk (0.02s)
        run_test.go:484: Different output from kati (red) to the expected value from make (green):
            src/ninja.cc:671: default_target_PASS

Ninja is erroring out because there are no default targets defined:

# Generated by kati ...

pool local_pool
 depth = 72

build _kati_always_build_: phony

Removing the circular dependency causes things to work:

# Generated by kati ...

pool local_pool
 depth = 72

build _kati_always_build_: phony

rule rule0
 description = build $out
 command = /bin/bash -c "true"
build loop: rule0 test
rule rule1
 description = build $out
 command = /bin/bash -c "echo PASS"
build test: rule1

default test
@danw
Copy link
Collaborator Author

danw commented Aug 10, 2020

This is due to the calculation of non_root_targets here. We could handle the self-referential loop rather easily by comparing p.first/t, but deeper loops would be more difficult to detect and handle with the current code.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant