[PATCH v2] scalar: avoid segfault in reconfigure --all

From: Derrick Stolee <stolee@gmail.com>

During the latest v2.45.0 update, 'scalar reconfigure --all' started to segfault on my machine. Breaking it down via the debugger, it was faulting on a NULL reference to the_hash_algo, which is a macro pointing to the_repository->hash_algo.

This NULL reference appears to be due to the way the loop is abusing the the_repository pointer, pointing it to a local repository struct after discovering that the current directory is a valid Git repository. This repo-swapping bit was in the original implementation from 4582676075 (scalar: teach 'reconfigure' to optionally handle all registered enlistments, 2021-12-03), but only recently started segfaulting while trying to parse the HEAD reference. This also only happens on the _second_ repository in the list, so does not reproduce if there is only one registered repo.

My first inclination was to try to refactor cmd_reconfigure() to execute 'git for-each-repo' instead of this loop. In addition to the difficulty of executing 'scalar reconfigure' within 'git for-each-repo', it would be difficult to perform the clean-up logic for non-existent repos if we relied on that child process.

Instead, I chose to move the temporary repo to be within the loop and reinstate the_repository to its old value after we are done performing logic on the current array item.

Add tests to t9210-scalar.sh to test 'scalar reconfigure --all' with multiple registered repos. There are two different ways that the old use of the_repository could trigger bugs. These issues are being solved independently to be more careful about the_repository being uninitialized, but the change in this patch around the use of the_repository is still a good safety precaution.

Co-authored-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Derrick Stolee <stolee@gmail.com>
---
    scalar: avoid segfault in reconfigure --all
    
    Update 'scalar reconfigure --all' to be more careful with the_repository
    pointer, avoiding sudden halts in some scenarios.
    
    ------------------------------------------------------------------------
    
    I noticed this while validating the v2.45.0 release (specifically the
    microsoft/git fork, but this applies to the core project).
    
    Thanks, Patrick, for digging in and finding the critical reasons why
    this issue can happen. I've included Patrick's test cases and given him
    co-authorship. I forged his sign-off, so could you please ACK that
    sign-off, Patrick?
    
    -Stolee

Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-1724%2Fderrickstolee%2Fscalar-reconfigure-repo-handle-v2
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-1724/derrickstolee/scalar-reconfigure-repo-handle-v2
Pull-Request: https://github.com/gitgitgadget/git/pull/1724

Range-diff vs v1:

 1:  6f0f5fa8f00 ! 1:  5be703b6c70 scalar: avoid segfault in reconfigure --all
     @@ Commit message
          reinstate the_repository to its old value after we are done performing
          logic on the current array item.
      
     -    Add a test to t9210-scalar.sh to test 'scalar reconfigure --all' with
     -    multiple registered repos, as a precaution. Unfortunately, I was unable
     -    to reproduce the segfault using this test, so there is some coverage
     -    left to be desired. What exactly causes my setup to hit this bug but not
     -    this test structure? Unclear.
     +    Add tests to t9210-scalar.sh to test 'scalar reconfigure --all' with
     +    multiple registered repos. There are two different ways that the old
     +    use of the_repository could trigger bugs. These issues are being solved
     +    independently to be more careful about the_repository being
     +    uninitialized, but the change in this patch around the use of
     +    the_repository is still a good safety precaution.
      
     +    Co-authored-by: Patrick Steinhardt <ps@pks.im>
     +    Signed-off-by: Patrick Steinhardt <ps@pks.im>
          Signed-off-by: Derrick Stolee <stolee@gmail.com>
      
       ## scalar.c ##
     @@ t/t9210-scalar.sh: test_expect_success 'scalar reconfigure' '
       	test true = "$(git -C one/src config core.preloadIndex)"
       '
       
     -+test_expect_success 'scalar reconfigure --all' '
     ++test_expect_success 'scalar reconfigure --all with includeIf.onbranch' '
      +	repos="two three four" &&
      +	for num in $repos
      +	do
      +		git init $num/src &&
      +		scalar register $num/src &&
     ++		git -C $num/src config includeif."onbranch:foo".path something &&
      +		git -C $num/src config core.preloadIndex false || return 1
      +	done &&
      +
     @@ t/t9210-scalar.sh: test_expect_success 'scalar reconfigure' '
      +		test true = "$(git -C $num/src config core.preloadIndex)" || return 1
      +	done
      +'
     ++
     ++ test_expect_success 'scalar reconfigure --all with detached HEADs' '
     ++	repos="two three four" &&
     ++	for num in $repos
     ++	do
     ++		rm -rf $num/src &&
     ++		git init $num/src &&
     ++		scalar register $num/src &&
     ++		git -C $num/src config core.preloadIndex false &&
     ++		test_commit -C $num/src initial &&
     ++		git -C $num/src switch --detach HEAD || return 1
     ++	done &&
     ++
     ++	scalar reconfigure --all &&
     ++
     ++	for num in $repos
     ++	do
     ++		test true = "$(git -C $num/src config core.preloadIndex)" || return 1
     ++	done
     ++'
      +
       test_expect_success '`reconfigure -a` removes stale config entries' '
       	git init stale/src &&

 scalar.c          | 10 +++++++---
 t/t9210-scalar.sh | 38 ++++++++++++++++++++++++++++++++++++++
 2 files changed, 45 insertions(+), 3 deletions(-)