The logic for deciding what error to emit under various update-ref failure scenarios is nontrivial. Add a bunch of tests of these error messages. Then improve a few error messages that were not ideal. One or two other error messages could still be improved, but that would take more work so I will skip it for now. This patch series applies on top of mh/split-under-lock. Improving the testing of that patch series was in fact the main motivation for this one. It didn't turn up any serious errors, so I don't think that this patch series needs to hold up the other one. On the other hand, each of the patch series changes some error messages, so if possible it would be preferable to ship them in the same Git release. Happily, this branch doesn't conflict with either of the other large patch series that are queued up in this area (mh/ref-iterators [2] and ref-store [3]). These patches are also available from my GitHub account [4] as branch update-ref-errors. Michael [1] http://thread.gmane.org/gmane.comp.version-control.git/293800 [2] http://thread.gmane.org/gmane.comp.version-control.git/296322 [3] http://thread.gmane.org/gmane.comp.version-control.git/296409 [4] https://github.com/mhagger/git Michael Haggerty (5): t1404: rename file to t1404-update-ref-errors.sh t1404: document function test_update_rejected t1404: add more tests of update-ref error handling lock_ref_for_update(): make error handling more uniform lock_ref_for_update(): avoid a symref resolution refs/files-backend.c | 77 ++++--- t/t1404-update-ref-df-conflicts.sh | 181 ----------------- t/t1404-update-ref-errors.sh | 402 +++++++++++++++++++++++++++++++++++++ 3 files changed, 447 insertions(+), 213 deletions(-) delete mode 100755 t/t1404-update-ref-df-conflicts.sh create mode 100755 t/t1404-update-ref-errors.sh -- 2.8.1 -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
