Re: [PATCH] log: add log.excludeDecoration config option

[Taylor Blau <me@ttaylorr.com>]

Hi Stolee,

On Mon, Apr 13, 2020 at 03:28:39PM +0000, Derrick Stolee via GitGitGadget wrote:
> From: Derrick Stolee <dstolee@microsoft.com>
>
> In 'git log', the --decorate-refs-exclude option appends a pattern
> to a string_list. This list is used to prevent showing some refs
> in the decoration output, or even by --simplify-by-decoration.
>
> Users may want to use their refs space to store utility refs that
> should not appear in the decoration output. For example, Scalar [1]
> runs a background fetch but places the "new" refs inside the
> refs/scalar/hidden/<remote>/* refspace instead of refs/<remote>/*
> to avoid updating remote refs when the user is not looking. However,
> these "hidden" refs appear during regular 'git log' queries.
>
> A similar idea to use "hidden" refs is under consideration for core
> Git [2].
>
> Add the 'log.excludeDecoration' config option so users can exclude
> some refs from decorations by default instead of needing to use
> --decorate-refs-exclude manually. The config value is multi-valued
> much like the command-line option.
>
> There are several tests in t4202-log.sh that test the
> --decorate-refs-(include|exclude) options, so these are extended.
> Since the expected output is already stored as a file, simply add
> new calls that replace a "--decorate-refs-exclude" option with an
> in-line config setting.
>
> [1] https://github.com/microsoft/scalar
> [2] https://lore.kernel.org/git/77b1da5d3063a2404cd750adfe3bb8be9b6c497d.1585946894.git.gitgitgadget@gmail.com/
>
> Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
> ---
>     log: add log.excludeDecoration config option
>
>     This was something hinted at in the "fetch" step of the background
>     maintenance RFC. Should be a relatively minor addition to our config
>     options.
>
>     We definitely want this feature for microsoft/git (we would set
>     log.excludeDecoration=refs/scalar/* in all Scalar repos), but we will
>     wait for feedback from the community.
>
>     Thanks, -Stolee
>
> Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-610%2Fderrickstolee%2Flog-exclude-decoration-v1
> Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-610/derrickstolee/log-exclude-decoration-v1
> Pull-Request: https://github.com/gitgitgadget/git/pull/610
>
>  Documentation/config/log.txt |  5 +++++
>  builtin/log.c                | 14 ++++++++++++++
>  t/t4202-log.sh               | 22 ++++++++++++++++++++++
>  3 files changed, 41 insertions(+)
>
> diff --git a/Documentation/config/log.txt b/Documentation/config/log.txt
> index e9e1e397f3f..1a158324f79 100644
> --- a/Documentation/config/log.txt
> +++ b/Documentation/config/log.txt
> @@ -18,6 +18,11 @@ log.decorate::
>  	names are shown. This is the same as the `--decorate` option
>  	of the `git log`.
>
> +log.excludeDecoration::
> +	Exclude the specified patterns from the log decorations. This multi-
> +	valued config option is the same as the `--decorate-refs-exclude`
> +	option of `git log`.
> +

Thanks for this documentation update. Do you think that it would be
worth updating the entry for '--decorate-refs-exclude', too? I figure
that it would be good, since scripters who look at 'git log's man page
before 'git config's would have a trail to follow in case they want a
persistent '--decorate-refs-exclude' option.

>  log.follow::
>  	If `true`, `git log` will act as if the `--follow` option was used when
>  	a single <path> is given.  This has the same limitations as `--follow`,
> diff --git a/builtin/log.c b/builtin/log.c
> index 83a4a6188e2..d7d1d5b7143 100644
> --- a/builtin/log.c
> +++ b/builtin/log.c
> @@ -236,7 +236,21 @@ static void cmd_log_init_finish(int argc, const char **argv, const char *prefix,
>  	}
>
>  	if (decoration_style) {
> +		const struct string_list *config_exclude =
> +			repo_config_get_value_multi(the_repository,
> +						    "log.excludeDecoration");
> +
> +		if (config_exclude) {
> +			struct string_list_item *item;
> +			for (item = config_exclude->items;
> +			     item && item < config_exclude->items + config_exclude->nr;
> +			     item++)

Any reason not to use the 'for_each_string_list_item' macro in
'string-list.h' for this?

> +				string_list_append(&decorate_refs_exclude,
> +						item->string);
> +		}
> +

On my first read, I thought that there might be some value in using a
hash set here, since I (incorrectly) thought that only literal reference
names were added here, in which case we'd benefit from a constant time
lookup.

But, since we store patterns here, it's not helpful for us to have a
constant time lookup, since we don't have anything to look up in the
first place! I guess you could expand out all of these patterns ahead of
time, but I don't think that this is universally a good thing to do. We
may have a pattern like 'refs/heads/*', but we're only doing a 'git log'
over a part of history that doesn't have many/any refs pointing at it,
in which case expanding ahead of time was a bad thing to do.

Of course, 'string_list_append' is going to potentially introduce
duplicate patterns, but I don't think that should be a huge problem,
since 'ref_filter_match' stops as soon as it finds an exclude/include. I
guess you could use 'string_list_insert', but it makes insertions more
expensive without a clear benefit.

>  		rev->show_decorations = 1;
> +
>  		load_ref_decorations(&decoration_filter, decoration_style);
>  	}
>
> diff --git a/t/t4202-log.sh b/t/t4202-log.sh
> index 0f766ba65f5..b5de449e510 100755
> --- a/t/t4202-log.sh
> +++ b/t/t4202-log.sh
> @@ -787,6 +787,9 @@ test_expect_success 'decorate-refs-exclude with glob' '
>  	EOF
>  	git log -n6 --decorate=short --pretty="tformat:%f%d" \
>  		--decorate-refs-exclude="heads/octopus*" >actual &&
> +	test_cmp expect.decorate actual &&
> +	git -c log.excludeDecoration="heads/octopus*" log \
> +		-n6 --decorate=short --pretty="tformat:%f%d" >actual &&
>  	test_cmp expect.decorate actual
>  '
>
> @@ -801,6 +804,9 @@ test_expect_success 'decorate-refs-exclude without globs' '
>  	EOF
>  	git log -n6 --decorate=short --pretty="tformat:%f%d" \
>  		--decorate-refs-exclude="tags/reach" >actual &&
> +	test_cmp expect.decorate actual &&
> +	git -c log.excludeDecoration="tags/reach" log \
> +		-n6 --decorate=short --pretty="tformat:%f%d" >actual &&
>  	test_cmp expect.decorate actual
>  '
>
> @@ -816,6 +822,14 @@ test_expect_success 'multiple decorate-refs-exclude' '
>  	git log -n6 --decorate=short --pretty="tformat:%f%d" \
>  		--decorate-refs-exclude="heads/octopus*" \
>  		--decorate-refs-exclude="tags/reach" >actual &&
> +	test_cmp expect.decorate actual &&
> +	git -c log.excludeDecoration="heads/octopus*" \
> +		-c log.excludeDecoration="tags/reach" log \
> +		-n6 --decorate=short --pretty="tformat:%f%d" >actual &&
> +	test_cmp expect.decorate actual &&
> +	git -c log.excludeDecoration="heads/octopus*" log \
> +		--decorate-refs-exclude="tags/reach" \
> +		-n6 --decorate=short --pretty="tformat:%f%d" >actual &&
>  	test_cmp expect.decorate actual
>  '
>
> @@ -831,6 +845,10 @@ test_expect_success 'decorate-refs and decorate-refs-exclude' '
>  	git log -n6 --decorate=short --pretty="tformat:%f%d" \
>  		--decorate-refs="heads/*" \
>  		--decorate-refs-exclude="heads/oc*" >actual &&
> +	test_cmp expect.decorate actual &&
> +	git -c log.excludeDecoration="heads/oc*" log \
> +		--decorate-refs="heads/*" \
> +		-n6 --decorate=short --pretty="tformat:%f%d" >actual &&
>  	test_cmp expect.decorate actual
>  '
>
> @@ -846,6 +864,10 @@ test_expect_success 'decorate-refs-exclude and simplify-by-decoration' '
>  	git log -n6 --decorate=short --pretty="tformat:%f%d" \
>  		--decorate-refs-exclude="*octopus*" \
>  		--simplify-by-decoration >actual &&
> +	test_cmp expect.decorate actual &&
> +	git -c log.excludeDecoration="*octopus*" log \
> +		-n6 --decorate=short --pretty="tformat:%f%d" \
> +		--simplify-by-decoration >actual &&
>  	test_cmp expect.decorate actual
>  '
>

Thanks for the tests, they look good. I like your strategy of repeating
the test with and without the new config settings.

>
> base-commit: 274b9cc25322d9ee79aa8e6d4e86f0ffe5ced925
> --
> gitgitgadget

Thanks,
Taylor