gitlab error | Dagger | Page 1

Nope, this subpath does exist: we loop over that ; but I just checked and it's the same for Erik's error case.

This seems to be the issue: parseRefString returning / receinving a wrong ref

ModuleSource.resolveFromCaller: ModuleSource!
                    multi.go:85: 126 : │ parseRefString: gitlab.com/dagger-modules/test@323d56c9ece3492d13f58b8b603d31a7c511cd41

Now, another thing (for reference), the expected error on the sad test is about: escaping the module source. If it is wrongly parsed now, it is probably the same root cause

 dagger install gitlab.com/dagger-modules/test/more/dagger-test-modules-public/../../@323d56c9ece3492d13f58b8b603d31a7c511cd41 DONE [43.1s]
                    testctx.go:204: 
                                Error Trace:    /app/core/integration/module_cli_test.go:893
                                                                        /app/testctx/testctx.go:168
                                Error:          An error is expected but got nil.
                                Test:           TestCLI/TestDaggerInstall/GitLab_public/git/sad
                    testctx.go:187: Cleaned up SSH agent: /tmp/ssh-agent678586485/ssh-agent.sock

This PR seems very suspicious, as we now loop over the filteredModuleRefs that use parseRefString: https://github.com/dagger/dagger/commit/ab4ae221c5dd80a16d342b757522c849042206ab#diff-35a72c75f37c6dcddd3e1df88cccf5a8b22ac1f6304301c66c3a1c7394ff0eafR464-R481 and as Alex mentioned, relies on maps that could lead to some nondeterminism

#gitlab error