Git development
 help / color / mirror / Atom feed
From: Kaartic Sivaraam <kaartic.sivaraam@gmail.com>
To: Karthik Nayak <karthik.188@gmail.com>, git@vger.kernel.org
Cc: gitster@pobox.com
Subject: Re: [PATCH v2 02/20] t9300: make the series B tests self-contained
Date: Thu, 19 Nov 2026 17:52:33 +0530	[thread overview]
Message-ID: <db459a4f-9454-4855-b1c5-c8ffc9b277ce@gmail.com> (raw)
In-Reply-To: <20261112-kn-fast-import-tests-1-v2-2-b8c3a0ab27e8@gmail.com>

On 11/12/26 23:04, Karthik Nayak wrote:
> The 't/t9300-fast-import.sh' test file contains numerous tests and is
> currently just shy of 4000 lines. This is because the file contains
> tests over multiple series, each series testing particular functionality
> or sections of 'git-fast-import(1)'. In upcoming commits, we want to
> split out these tests into individual files for better discoverability.
> 
> First, we need to remove the dependency tests have on preceding tests.
> Modify the tests in 'series B' to each run in a new repository, which is
> removed when the test completes, so that they neither depend on nor
> leave behind any state.
> 
> The 'B: fail on invalid blob sha1' and 'B: accept branch name
> "TEMP_TAG"' tests import on top of 'main', which used to exist from
> earlier tests. Create it with 'test_commit', so that the first import
> can only fail because of the invalid blob ID.
> 
> The 'fsck' tests require the whole repository to be free of unreachable
> objects. They only passed because of the garbage collection in the
> cleanup of other tests. A new repository makes that unnecessary.
> 
> Signed-off-by: Karthik Nayak <karthik.188@gmail.com>
> ---
>   t/t9300-fast-import.sh | 472 ++++++++++++++++++++++++++++---------------------
>   1 file changed, 275 insertions(+), 197 deletions(-)
> 
> diff --git a/t/t9300-fast-import.sh b/t/t9300-fast-import.sh
> index 4cffe43a71..e2cf0b7a22 100755
> --- a/t/t9300-fast-import.sh
> +++ b/t/t9300-fast-import.sh
> @@ -358,274 +358,352 @@ test_expect_success 'A: export marks with large values' '
>   ###
>   
>   test_expect_success 'B: fail on invalid blob sha1' '
> -	test_tick &&
> -	cat >input <<-INPUT_END &&
> -	commit refs/heads/branch
> -	mark :1
> -	committer $GIT_COMMITTER_NAME <$GIT_COMMITTER_EMAIL> $GIT_COMMITTER_DATE
> -	data <<COMMIT
> -	corrupt
> -	COMMIT
> +	test_when_finished "rm -rf repo" &&
> +	git init repo &&
> +	(
> +		cd repo &&
> +		test_commit base &&

I suppose we need this test_commit after making wrapping this in a 
subshell since we can't lean on the commit created by the previous test. 
This is just the commit message mentions it. Makes sense.

>   
> -	from refs/heads/main
> -	M 755 $(echo $ZERO_OID | sed -e "s/0$/1/") zero1
> +		test_tick &&
> +		cat >input <<-INPUT_END &&
> +		commit refs/heads/branch
> +		mark :1
> +		committer $GIT_COMMITTER_NAME <$GIT_COMMITTER_EMAIL> $GIT_COMMITTER_DATE
> +		data <<COMMIT
> +		corrupt
> +		COMMIT
>   
> -	INPUT_END
> +		from refs/heads/main
> +		M 755 $(echo $ZERO_OID | sed -e "s/0$/1/") zero1
>   
> -	test_when_finished "rm -f .git/objects/pack_* .git/objects/index_*" &&

I suppose we're also losing this because the 'repo' itself would cease 
to exist after this change.

> -	test_must_fail git fast-import <input
> +		INPUT_END
> +
> +		test_must_fail git fast-import <input
> +	)
>   '
> 
 >   test_expect_success 'B: accept branch name "TEMP_TAG"' '> -	cat 
 >input <<-INPUT_END &&
> -	commit TEMP_TAG
> -	committer $GIT_COMMITTER_NAME <$GIT_COMMITTER_EMAIL> $GIT_COMMITTER_DATE
> -	data <<COMMIT
> -	tag base
> -	COMMIT
> +	test_when_finished "rm -rf repo" &&
> +	git init repo &&
> +	(
> +		cd repo &&
> +		test_commit new &&

I think this additional test_commit is deserved, too.

>   test_expect_success 'B: accept invalid timezone with raw-permissive' '
> -	cat >input <<-INPUT_END &&
> -	commit refs/heads/invalid-timezone
> -	committer $GIT_COMMITTER_NAME <$GIT_COMMITTER_EMAIL> 1234567890 +051800
> -	data <<COMMIT
> -	empty commit
> -	COMMIT
> -	INPUT_END
> +	test_when_finished "rm -rf repo" &&
> +	git init repo &&
> +	(
> +		cd repo &&
>   
> -	git init invalid-timezone &&
> -	git -C invalid-timezone fast-import --date-format=raw-permissive <input &&
> -	git -C invalid-timezone cat-file -p invalid-timezone >out &&
> -	test_grep "1234567890 [+]051800" out
> +		cat >input <<-INPUT_END &&
> +		commit refs/heads/invalid-timezone
> +		committer $GIT_COMMITTER_NAME <$GIT_COMMITTER_EMAIL> 1234567890 +051800
> +		data <<COMMIT
> +		empty commit
> +		COMMIT
> +		INPUT_END
> +
> +		git fast-import --date-format=raw-permissive <input &&
> +		git cat-file -p invalid-timezone >out &&

We're losing the '-C invalid-timezone' here since we are already in a 
repo created specifically for this test. Looks good.

> +		test_grep "1234567890 [+]051800" out
> +	)
>   '
>  [ ... snip ... ]
>   
>   test_expect_success 'B: fail on invalid file path of .gitmodules' '
> -	cat >input <<-INPUT_END &&
> -	blob
> -	mark :1
> -	data <<EOF
> -	File contents
> -	EOF
> +	test_when_finished "rm -rf repo" &&
> +	git init repo &&
> +	(
> +		cd repo &&
>   
> -	commit refs/heads/badpath
> -	committer Name <email> $GIT_COMMITTER_DATE
> -	data <<COMMIT
> -	Commit Message
> -	COMMIT
> -	M 120000 :1 .gitmodules
> -	INPUT_END
> +		cat >input <<-INPUT_END &&
> +		blob
> +		mark :1
> +		data <<EOF
> +		File contents
> +		EOF
>   
> -	test_when_finished "git update-ref -d refs/heads/badpath" &&
> -	test_must_fail git fast-import <input
> +		commit refs/heads/badpath
> +		committer Name <email> $GIT_COMMITTER_DATE
> +		data <<COMMIT
> +		Commit Message
> +		COMMIT
> +		M 120000 :1 .gitmodules
> +		INPUT_END
> +
> +		test_must_fail git fast-import <input
> +	)
>   '
>   
>   ###
> 


An --ignore-all-space diff helped me grok that most of this change is 
just the existing test getting their own repository and the test code 
indented to put them within a subshell.

I just wonder how much additional time or work this has resulted in 16 
additional repos being created and torn down.

-- 
Sivaraam


  reply	other threads:[~2026-11-19 12:22 UTC|newest]

Thread overview: 49+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-10 11:49 [PATCH 00/20] t9300: group and move tests to other files Karthik Nayak
2026-11-10 11:49 ` [PATCH 01/20] t9300: move 'fastimport.unpackLimit' to the required test Karthik Nayak
2026-11-10 11:49 ` [PATCH 02/20] t9300: make the series B tests self-contained Karthik Nayak
2026-11-10 11:49 ` [PATCH 03/20] t9300: make the series T " Karthik Nayak
2026-11-10 11:49 ` [PATCH 04/20] t9300: make the series S " Karthik Nayak
2026-11-10 11:49 ` [PATCH 05/20] t9300: make the series V " Karthik Nayak
2026-11-10 11:49 ` [PATCH 06/20] t9300: make the series Q " Karthik Nayak
2026-11-10 11:49 ` [PATCH 07/20] t9300: make the series P " Karthik Nayak
2026-11-10 11:49 ` [PATCH 08/20] t9300: make the series O " Karthik Nayak
2026-11-10 11:49 ` [PATCH 09/20] t9300: make the big blob " Karthik Nayak
2026-11-10 11:49 ` [PATCH 10/20] t9300: make the ref update " Karthik Nayak
2026-11-10 11:49 ` [PATCH 11/20] t9300: make the date format " Karthik Nayak
2026-11-10 11:49 ` [PATCH 12/20] t9300: make the pack edges test self-contained Karthik Nayak
2026-11-10 11:49 ` [PATCH 13/20] t9300: make the marks tests self-contained Karthik Nayak
2026-11-10 11:49 ` [PATCH 14/20] t9300: make the series Y " Karthik Nayak
2026-11-10 11:50 ` [PATCH 15/20] t9300: make the cat-blob " Karthik Nayak
2026-11-10 11:50 ` [PATCH 16/20] t9300: make the feature and option " Karthik Nayak
2026-11-10 11:50 ` [PATCH 17/20] t9300: make the stream " Karthik Nayak
2026-11-10 11:50 ` [PATCH 18/20] t9300: make the get-mark and encoding " Karthik Nayak
2026-11-10 11:50 ` [PATCH 19/20] t9300: move self-contained tests into dedicated files Karthik Nayak
2026-11-10 11:50 ` [PATCH 20/20] t93xx: drop the series letter from the moved test names Karthik Nayak
2026-11-10 16:45 ` [PATCH 00/20] t9300: group and move tests to other files Junio C Hamano
2026-11-10 20:06   ` Junio C Hamano
2026-11-11 11:22     ` Karthik Nayak
2026-11-12 17:34 ` [PATCH v2 " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 01/20] t9300: move 'fastimport.unpackLimit' to the required test Karthik Nayak
2026-11-19 10:06     ` Kaartic Sivaraam
2026-11-12 17:34   ` [PATCH v2 02/20] t9300: make the series B tests self-contained Karthik Nayak
2026-11-19 12:22     ` Kaartic Sivaraam [this message]
2026-11-12 17:34   ` [PATCH v2 03/20] t9300: make the series T " Karthik Nayak
2026-11-19 12:45     ` Kaartic Sivaraam
2026-11-12 17:34   ` [PATCH v2 04/20] t9300: make the series S " Karthik Nayak
2026-11-19 18:27     ` Kaartic Sivaraam
2026-11-12 17:34   ` [PATCH v2 05/20] t9300: make the series V " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 06/20] t9300: make the series Q " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 07/20] t9300: make the series P " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 08/20] t9300: make the series O " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 09/20] t9300: make the big blob " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 10/20] t9300: make the ref update " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 11/20] t9300: make the date format " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 12/20] t9300: make the pack edges test self-contained Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 13/20] t9300: make the marks tests self-contained Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 14/20] t9300: make the series Y " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 15/20] t9300: make the cat-blob " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 16/20] t9300: make the feature and option " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 17/20] t9300: make the stream " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 18/20] t9300: make the get-mark and encoding " Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 19/20] t9300: move self-contained tests into dedicated files Karthik Nayak
2026-11-12 17:34   ` [PATCH v2 20/20] t93xx: drop the series letter from the moved test names Karthik Nayak

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=db459a4f-9454-4855-b1c5-c8ffc9b277ce@gmail.com \
    --to=kaartic.sivaraam@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=karthik.188@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox