Re: [PATCH v3 2/2] tests: add tests for grep --max-count

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



"Carlos López via GitGitGadget"  <gitgitgadget@xxxxxxxxx> writes:

> From: =?UTF-8?q?Carlos=20L=C3=B3pez?= <00xc@xxxxxxxxxxxxxx>
>
> Add tests for grep's -m / --max-count to check if the option correctly
> outputs limited results, and that it interacts properly with other flags
> that could likely be used in conjunction.
>
> Signed-off-by: Carlos López 00xc@xxxxxxxxxxxxxx
> ---
>  t/t7810-grep.sh | 83 +++++++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 83 insertions(+)

This is better done as part of the previous patch.  The new tests
protect the new code from future breakage.

> diff --git a/t/t7810-grep.sh b/t/t7810-grep.sh
> index 69356011713..7b1b8a3cd93 100755
> --- a/t/t7810-grep.sh
> +++ b/t/t7810-grep.sh
> @@ -77,6 +77,7 @@ test_expect_success setup '
>  	# Say hello.
>  	function hello() {
>  	  echo "Hello world."
> +	  echo "Hello again."
>  	} # hello
>  
>  	# Still a no-op.
> @@ -595,6 +596,88 @@ test_expect_success 'grep --files-without-match --quiet' '
>  	test_must_be_empty actual
>  '
>  
> +cat >expected <<EOF &&
> +EOF
> +
> +test_expect_success 'grep --max-count 0 (must exit with non-zero)' '
> +	test_must_fail git grep --max-count 0 foo >actual &&
> +	test_cmp expected actual
> +'

For this particular one, "test_must_be_empty actual" would suffice,
without comparing with the expected output.

> +cat >expected <<EOF &&
> +file:foo mmap bar
> +EOF
> +
> +test_expect_success 'grep --max-count 1' '
> +	git grep --max-count 1 foo >actual &&
> +	test_cmp expected actual
> +'

Writing expected output outside test_expect_success that uses it is
a quite old style but that is because this test script is pretty
much ancient, so mimicking it is OK.  We'd need to come back later
when the tree is quiescent to clean them up, though (#leftoverbits).

> ...
> +	test_cmp expected actual
> +'

The new tests seem to give us a reasonable test coverage.  We could
discard one of the "-m1" vs "-m3" in the early ones, as they do not
give much extra test coverage over the other, to reduce repetition.

We do not test a case where we pick up-to N matches each from
multiple files, though.  Perhaps

    git grep -m1 -e o -- hello.\*

may stop after hitting "No-op." in hello.ps1 and "stdio" in hello.c,
which may make a good test, perhaps?

Thanks.




[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]

  Powered by Linux