Skip to content

remote: handle negative refspecs with show

Jacob Keller git 3 versions – New

Get series with b4
b4 am [email protected]

Versions newest first · linked by author and title

  1. v3 Latest6 patches 0 repliesNew
  2. v25 patches 8 repliesSuperseded
  3. v11 patch 10 repliesSuperseded

Changes in v3 from the cover letter

* re-ordered so that the fix is first
* cleaned up helper functions to honor -C instead of hard coding 'test'
* added expected failures for the short comings of this fix

Patches in v3 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/6remote: handle negative refspecs in git remote showbuiltin/remote.c remote.c remote.h +1 more+63 -30 replies
2/6t5505: remove sub shell use in favor of git -Ct/t5505-remote.sh+415 -6460 replies
3/6tests: handle --global directly in test_config/test_unconfigt/lib-parallel-checkout.sh t/lib-proto-disable.sh t/lib-submodule-update.sh +18 more+144 -1080 replies
4/6tests: only automatically unset matching values from test_configt/test-lib-functions.sh+33 -30 replies
5/6t5505: use test_config where appropriatet/t5505-remote.sh t/test-lib-functions.sh+30 -280 replies
6/6remote: add test for negative refspec with prunet/t5505-remote.sh+13 -00 replies

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v3 yet.

Read the v3 thread

Cover letter [PATCH v3 0/6] remote: handle negative refspecs with show

The git remote show command does not handle negative refspecs. In fact, the function will happily show that the refs will be fetched, even though they won't be.

The first patch in this series fixes the show issue for new refs that should be marked as skipped. The later patches do some cleanup on the t5505-remote.sh script, improving the test file, along with some improved test infrastructure.

Finally, I also noticed that git remote prune doesn't interact as expected with negative refspecs either, so there is a patch to add an expected failure for this as well.

I suspect we need to dig into get_stale_heads and make that handle negative refspecs properly, but I wasn't able to figure that out in the time I have today.

Changes since v2:
* re-ordered so that the fix is first
* cleaned up helper functions to honor -C instead of hard coding 'test'
* added expected failures for the short comings of this fix

Jacob Keller (6):
remote: handle negative refspecs in git remote show
t5505: remove sub shell use in favor of git -C
tests: handle --global directly in test_config/test_unconfig
tests: only automatically unset matching values from test_config
t5505: use test_config where appropriate
remote: add test for negative refspec with prune

 builtin/remote.c                        |   12 +-
 remote.c                                |    2 +-
 remote.h                                |    6 +
 t/lib-parallel-checkout.sh              |    4 +-
 t/lib-proto-disable.sh                  |   12 +-
 t/lib-submodule-update.sh               |    2 +-
 t/t0001-init.sh                         |   14 +-
 t/t0021-conversion.sh                   |   58 +-
 t/t0210-trace2-normal.sh                |   10 +-
 t/t0211-trace2-perf.sh                  |    4 +-
 t/t0212-trace2-event.sh                 |    2 +-
 t/t1309-early-config.sh                 |    2 +-
 t/t2081-parallel-checkout-collisions.sh |    2 +-
 t/t2082-parallel-checkout-attributes.sh |    4 +-
 t/t3431-rebase-fork-point.sh            |    2 +-
 t/t5505-remote.sh                       | 1132 ++++++++++-------------
 t/t5544-pack-objects-hook.sh            |    6 +-
 t/t5550-http-fetch-dumb.sh              |    8 +-
 t/t5573-pull-verify-signatures.sh       |    6 +-
 t/t5606-clone-options.sh                |   10 +-
 t/t5617-clone-submodules-remote.sh      |    4 +-
 t/t5702-protocol-v2.sh                  |    2 +-
 t/t7814-grep-recurse-submodules.sh      |   16 +-
 t/test-lib-functions.sh                 |  104 ++-
 24 files changed, 667 insertions(+), 757 deletions(-)
Signature · 1 line
2.36.1