[v2,0/6] Fix GIT_TEST_SPLIT_INDEX

Message

SZEDER Gábor Aug. 26, 2021, 8:59 p.m. UTC

To recap from v1's cover letter:

Running tests with GIT_TEST_SPLIT_INDEX=1 is supposed to turn on the
split index feature and trigger index splitting (mostly) randomly.
Alas, this has been broken for ~2.5 years, and it hasn't triggered any
index splitting since then.

The last patch in this series makes GIT_TEST_SPLIT_INDEX=1 work again,
although it slightly changes its behavior; see its log message for all
the details.

Patches 3 and 4 fix old tests that started to fail because
of this slight behavior change.  Bsides the final fix, patch 4 is the
most interesting in this series.

Patch 5 fixes recent sparse index tests that were introduced while
GIT_TEST_SPLIT_INDEX was broken.

The first two patches are just minor while-at-it test cleanups.


Changes since v1, both in patch 5:
  - Removed a remark about making split and sparse index compatible.
  - Stolle suggested to rebase this patch series on top of topic
    'ds/sparse-index-ignored-files' to avoid merge conflicts in a
    sparse index test.  Well, I didn't do that, but instead I rebased
    this series on top of the merge of master and 86d6ca5886 (t7519:
    rewrite sparse index test, 2021-08-10), and the latter commit:
      - is the one responsible for said merge conflicts, and
      - is at the bottom of 'ds/sparse-index-ignored-files', and could
	be considered more as a followup fix for topic
	'ds/status-with-sparse-index' (already in v2.33.0) than an
	integral part of 'ds/sparse-index-ignored-files'.
    The rest of 'ds/sparse-index-ignored-files' doesn't have any
    conflicts with this patch series that auto-merge couldn't resolve.

SZEDER Gábor (6):
  t1600-index: remove unnecessary redirection
  t1600-index: don't run git commands upstream of a pipe
  t1600-index: disable GIT_TEST_SPLIT_INDEX
  read-cache: look for shared index files next to the index, too
  tests: disable GIT_TEST_SPLIT_INDEX for sparse index tests
  read-cache: fix GIT_TEST_SPLIT_INDEX

 read-cache.c                       | 37 +++++++++++-----
 t/t1091-sparse-checkout-builtin.sh | 25 ++++++-----
 t/t1600-index.sh                   | 13 ++++--
 t/t1700-split-index.sh             | 34 +++++++++++++++
 t/t7519-status-fsmonitor.sh        | 68 ++++++++++++++++--------------
 5 files changed, 121 insertions(+), 56 deletions(-)

Range-diff against v1:
1:  7e1d38d215 = 1:  ab555030cb t1600-index: remove unnecessary redirection
2:  d86c91d26e = 2:  38d350b02d t1600-index: don't run git commands upstream of a pipe
3:  792b6739e5 = 3:  57b5386aee t1600-index: disable GIT_TEST_SPLIT_INDEX
4:  b12467bab7 = 4:  58c444bcd0 read-cache: look for shared index files next to the index, too
5:  51e13df905 ! 5:  24b8b7738c tests: disable GIT_TEST_SPLIT_INDEX for sparse index tests
    @@ Commit message
         by later sanity checks finding that the index is not in fact a sparse
         index.  This indicates that 'git sparse-checkout init --sparse-index'
         lacks some error checking and its tests lack coverage related to split
    -    index, but fixing those issues (let alone making sparse index
    -    comparible with split index) is beyond the scope of this patch series.
    +    index, but fixing those issues is beyond the scope of this patch
    +    series.
     
         [1] https://public-inbox.org/git/[e-mail]/
     
    @@ t/t7519-status-fsmonitor.sh: check_sparse_index_behavior () {
      }
      
      test_expect_success 'status succeeds with sparse index' '

  I took the liberty to cut the part of the range-diff showing the
  changes of changes to this test, because 86d6ca5886 totally rewrites
  this test, while this patch changes the test's indentation,
  resulting in an unusable range-diff.
  Instead, to show the only relevant change in this patch here is the
  first hunk touching this test with --ignore-all-space from v1:

    @@ -402,6 +402,10 @@ check_sparse_index_behavior () {
     }
    
     test_expect_success 'status succeeds with sparse index' '
    +       (
    +               sane_unset GIT_TEST_SPLIT_INDEX &&
    +               git update-index --no-split-index &&
    +
                    git reset --hard &&
    
                    test_config core.fsmonitor "$TEST_DIRECTORY/t7519/fsmonitor-all" &&

  And from v2:

    @@ -399,6 +399,9 @@ check_sparse_index_behavior () {
     }
    
     test_expect_success 'status succeeds with sparse index' '
    +       (
    +               sane_unset GIT_TEST_SPLIT_INDEX &&
    +
                    git clone . full &&
                    git clone . sparse &&
                    git -C sparse sparse-checkout init --cone --sparse-index &&

6:  a5c790e622 = 6:  a43a12b0ce read-cache: fix GIT_TEST_SPLIT_INDEX

Comments

Derrick Stolee Aug. 31, 2021, 2:47 p.m. UTC | #1

On 8/26/2021 4:59 PM, SZEDER Gábor wrote:
> To recap from v1's cover letter:
> 
> Running tests with GIT_TEST_SPLIT_INDEX=1 is supposed to turn on the
> split index feature and trigger index splitting (mostly) randomly.
> Alas, this has been broken for ~2.5 years, and it hasn't triggered any
> index splitting since then.
> 
> The last patch in this series makes GIT_TEST_SPLIT_INDEX=1 work again,
> although it slightly changes its behavior; see its log message for all
> the details.

I checked the range-diff and reread the patches. This version looks
good to me. Thanks for pointing out the new failures that were
happening in my patches. I think I fixed them in my latest versions
of the sparse-index work, so the latest 'seen' should pass with these
changes to GIT_TEST_SPLIT_INDEX.

Thanks,
-Stolee

Junio C Hamano Aug. 31, 2021, 5:38 p.m. UTC | #2

Derrick Stolee <[e-mail]> writes:

> On 8/26/2021 4:59 PM, SZEDER Gábor wrote:
>> To recap from v1's cover letter:
>> 
>> Running tests with GIT_TEST_SPLIT_INDEX=1 is supposed to turn on the
>> split index feature and trigger index splitting (mostly) randomly.
>> Alas, this has been broken for ~2.5 years, and it hasn't triggered any
>> index splitting since then.
>> 
>> The last patch in this series makes GIT_TEST_SPLIT_INDEX=1 work again,
>> although it slightly changes its behavior; see its log message for all
>> the details.
>
> I checked the range-diff and reread the patches. This version looks
> good to me. Thanks for pointing out the new failures that were
> happening in my patches. I think I fixed them in my latest versions
> of the sparse-index work, so the latest 'seen' should pass with these
> changes to GIT_TEST_SPLIT_INDEX.
>
> Thanks,
> -Stolee


This, at least the t7519-status-fsmonitor.sh part, seems to depend
on the 'ds/sparse-index-ignored-files' topic, so I'll prepare a
merge of the topic on top of 'master' then queue these with your
Acked-by.

Thanks, both.