Description

Using:

• git version 2.4.6
• gitdb==0.6.4
• GitPython==1.0.1

The use case forgit merge-base --is-ancestor from git manpage:

Check if the first is an ancestor of the second , and exit with status 0 if true, or with status 1 if not. Errors are signaled by a non-zero status that is not 1.

Related to#169 (comment)

The problem is that the output ofgit merge-base --is-ancestor comes from the return code and unlike without the--is-ancestor flag, this command doesn't return commit SHAs but is a simple yes-no check. The current implementation doesn't take into account the return code and the command cannot be usefully run.

Current Behaviour

revA is ancestor to revB

$ git merge-base --is-ancestor$revA$revB;echo$?0

>>>repo.merge_base(revA,revB,is_ancestor=True)[ ]

revA is not ancestor to revB

$ git merge-base --is-ancestor$revA$revB;echo$?1

>>>repo.merge_base(revA,revB,is_ancestor=True)[ ]

revA is not ancestor to revB and revB doesn't exist

$ git merge-base --is-ancestor$revA$revB;echo$?fatal: Not a valid commit name$revB128

>>>repo.merge_base(revA,revB,is_ancestor=True)GitCommandError:'git merge-base --is-ancestor $revA $revB'returnedwithexitcode128stderr:'fatal: Not a valid commit name $revB'

Desired Behaviour

When thekwargis_ancestor is used, don't return a list, but return a bool that is true if the return code is 0. Return False if return code is 1. Raise Error as before for all other cases

$ git merge-base --is-ancestor$revA$revB;echo$?0

>>>repo.merge_base(revA,revB,is_ancestor=True)True

revA is not ancestor to revB

$ git merge-base --is-ancestor$revA$revB;echo$?1

>>>repo.merge_base(revA,revB,is_ancestor=True)False

revA is not ancestor to revB and revB doesn't exist

$ git merge-base --is-ancestor$revA$revB;echo$?fatal: Not a valid commit name$revB128

>>>repo.merge_base(revA,revB,is_ancestor=True)GitCommandError:'git merge-base --is-ancestor $revA $revB'returnedwithexitcode128stderr:'fatal: Not a valid commit name $revB'