git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [GSoC][PATCH v1] t9811: avoid pipe in git commands in test script

From
Junio C Hamano <gitster@pobox.com>
Date
Apr 2, 2019, 02:39 UTC
Message-ID
<xmqqk1gdt9ke.fsf@gitster-ct.c.googlers.com>
In-Reply-To
<20190402014115.22478-1-khalludi123@gmail.com>
Khalid Ali <khalludi123@gmail.com> writes:
> The exit code of the upstream in a pipe is ignored thus we
> should avoid using it. By writing out the output of the git command to a
> file, we can test the exit codes of both the commands.

End the log message here by moving the next two paragraphs below the three-dash line.

Show 23 quoted lines
>
> Aside from the commit message, I plan to apply for GSoC. Planning to
> solve the rebase/cherry-pick issue or adding functionality for the
> rebase interactive command. 
>
> Any review on this commit would be appreciated!
>
> Signed-off-by: Khalid Ali <khalludi123@gmail.com>
> ---
>  t/t9811-git-p4-label-import.sh | 11 ++++++-----
>  1 file changed, 6 insertions(+), 5 deletions(-)
>
> diff --git a/t/t9811-git-p4-label-import.sh b/t/t9811-git-p4-label-import.sh
> index 602b0a5d5c..2f4e80ed55 100755
> --- a/t/t9811-git-p4-label-import.sh
> +++ b/t/t9811-git-p4-label-import.sh
> @@ -63,7 +63,8 @@ test_expect_success 'basic p4 labels' '
>  		git checkout TAG_WITH\$_SHELL_CHAR &&
>  		test -f f1 && test -f f2 && test -f file_with_\$metachar &&
>  
> -		git show TAG_LONG_LABEL | grep -q "A Label second line"
> +		git show TAG_LONG_LABEL >actual &&
> +		grep -q "A Label second line" actual

There is no mention of file 'actual' before this patch, so we can reasonably be sure that we are not breaking expectations of existing test that a file with that name has contents different from what the above command produces. On the other hand, if later tests have things like "git add ." or "git status -s" and expects that there is no such file called 'actual', this change may have broken the expectation.

I *think* the above is done after going inside $git directory, which will be removed with "test_when_finished cleanup_git", so it is fairly clear that leaving an extra file 'actual' behind at this point is safe.

This is not a problem with this conversion, but we tend to avoid squelching standard output to help debugging tests. An independent clean-up patch may want to replace "grep -q" with just "grep".

Show 10 quoted lines
> @@ -92,9 +93,9 @@ test_expect_success 'two labels on the same changelist' '
>  		cd "$git" &&
>  		git p4 sync --import-labels &&
>  
> -		git tag | grep TAG_F1 &&
> -		git tag | grep -q TAG_F1_1 &&
> -		git tag | grep -q TAG_F1_2 &&
> +		git tag >actual && grep TAG_F1 actual &&
> +		git tag >actual && grep -q TAG_F1_1 actual &&
> +		git tag >actual && grep -q TAG_F1_2 actual &&
Ditto.
Show 10 quoted lines
>  
>  		cd main &&
>  
> @@ -205,7 +206,7 @@ test_expect_success 'use git config to enable import/export of tags' '
>  		git p4 rebase --verbose &&
>  		git p4 submit --verbose &&
>  		git tag &&
> -		git tag | grep TAG_F1_1
> +		git tag >actual && grep TAG_F1_1 actual
>  	) &&

I notice that this test piece does not call cleanup_git when it is done. I think that is a mistake (not introduced by this patch) that ma want to get fixed independently.

>  	(
>  		cd "$cli" &&
Previous: Khalid AliNext: Khalid Ali
Message 2 of 3 in “t9811: avoid pipe in git commands in test script”
  1. Khalid AliApr 2, 2019
  2. Junio C HamanoApr 2, 2019
  3. Khalid AliApr 2, 2019

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.