This behaviour is already documented in got-worktree(5) but wasn't actually implemented. ok stsp@

The fix for CVE-2022-39253 in Git made our regression tests fail since creating submodules from local clones is no longer allowed by default. Add an override to our invocations of "git submodule add", the same which was added to Git's regression test suite as part of their fix. This CVE doesn't otherwise affect us. We do not implement submodules. Our use of them in regression tests exists only to ensure that Got does not freak out when it sees one. Precisely because automated support for nested repositories and/or working copies (as in Git Submodules, or Subversion Externals, etc.) have been an endless source of security problems in other systems. sudden test failures pointed out by naddy@

The idea is that got patch can simply look for a line such as: commit - abcde1234567... to find the merge base commit ID to show in diff3 conflict markers. got log -p now displays commit IDs in its diff header, instead ofl tree or blob IDs. ok op@

Otherwise the test directory it's not left in place; ok tracey@

This brings the rest of the regression test scripts in line with patch.sh.

Since the interpreter for the regress shell scripts are using /bin/sh this will usually imply some level of POSIX compliance (that isn't bash-specific, for instance). Some systems use dash as their sh shell and as such is stricter POSIX compliance. To help -portable, make the shell test checks use a single '=' for equality, rather than '=='.

Annoying bug which we inherited from OpenRCS which inherited it from OpenBSD's diff3 program. ok tracey millert

Commands which perform merges will now install bad symlinks as symlinks in the work tree, instead of creating them as regular files. This means bad symlinks committed with 'got commit -S' (or Git) will be preserved. The decision to introduce a bad symlink is done at commit-time and merges should not forcefully reverse this decision. The cherrypick and backout commands require a manual commit step, and a merge result with bad symlinks will require use of 'got commit -S'. Additional testing by thomas adam

reported by + ok naddy

This was a long-standing and very annoying bug. The two xfail tests in the cherrypick test suite are passing now.

The test case I am using here is a bit large but I could not yet find a way to make it smaller. If someone knows a smaller test case, please let me know. naddy caught a small bug in the new test code I had written, thanks!

one of these tests is still failing; there is another edge case left to fix

ok stsp

Previous default output remains when test cases are run individually. ok tracey

If the work tree's path prefix does not exist in the first of the two trees, then 'got cherrypick' failed with "no such entry found in tree". But this is a legitimate situation, as shown in the new test added here. The first tree could be the empty tree, for example, which should result in 'got cherrypick' adding all files from the second tree instead of complaining about a non-existent path-prefix directory in the first tree.