Re: [PATCH v2] rev-list: support human-readable output for `--disk-usage`

From: Jeff King <peff@peff.net>
Date: Aug 9, 2022, 13:22 UTC
Message-ID: <YvJfpNSKMIPqVQmD@coredump.intra.peff.net>
In-Reply-To: <pull.1313.v2.git.1659947722132.gitgitgadget@gmail.com>

On Mon, Aug 08, 2022 at 08:35:21AM +0000, Li Linchao via GitGitGadget wrote:

Show 7 quoted lines

> The '--disk-usage' option for git-rev-list was introduced in 16950f8384
> (rev-list: add --disk-usage option for calculating disk usage, 2021-02-09).
> This is very useful for people inspect their git repo's objects usage
> infomation, but the resulting number is quit hard for a human to read.
> 
> Teach git rev-list to output a human readable result when using
> '--disk-usage'.

OK. When adding --disk-usage, I never really dreamed people would use it for human output, since "du .git" is usually a suitable approximation. :) But I don't have any real objection. I'm curious what your use case is like, if you don't mind sharing. We used it at GitHub for computing per-fork sizes for analysis, etc (so the result was always fed into another script).

>  Documentation/rev-list-options.txt |  5 +++-
>  builtin/rev-list.c                 | 42 ++++++++++++++++++++++++------
>  t/t6115-rev-list-du.sh             | 22 ++++++++++++++++
>  3 files changed, 60 insertions(+), 9 deletions(-)

The patch itself looks pretty sensible (and thanks Ævar for the first round of review; the suggestions there all looked good). A few small comments:

Show 13 quoted lines

> @@ -481,8 +485,13 @@ static int try_bitmap_disk_usage(struct rev_info *revs,
>  	if (!bitmap_git)
>  		return -1;
>  
> -	printf("%"PRIuMAX"\n",
> -	       (uintmax_t)get_disk_usage_from_bitmap(bitmap_git, revs));
> +	size_from_bitmap = get_disk_usage_from_bitmap(bitmap_git, revs);
> +	if (human_readable)
> +		strbuf_humanise_bytes(&disk_buf, size_from_bitmap);
> +	else
> +		strbuf_addf(&disk_buf, "%"PRIuMAX"", (uintmax_t)size_from_bitmap);
> +	puts(disk_buf.buf);
> +	strbuf_release(&disk_buf);

It's not a lot of duplicated lines, but since it is implementing policy logic, I think it would be nice to move the formatting decision into a function. Something like:

  static void show_disk_usage(off_t size)
  {
	struct strbuf sb = STRBUF_INIT;
	if (human_readable)
		strbuf_humanise_bytes(&sb, size);
	else
		strbuf_addf(&sb, "%"PRIuMAX, (uintmax_t)size_from_bitmap);
	puts(sb.buf);
	strbuf_release(&sb);
  }

and then you can call it from here, and from the non-bitmap path below.

(Also, while typing it out, I noticed that you don't need the extra "" after PRIuMAX; that just concatenates an empty string).

Show 19 quoted lines

> -		if (!strcmp(arg, "--disk-usage")) {
> -			show_disk_usage = 1;
> -			info.flags |= REV_LIST_QUIET;
> -			continue;
> +		if (skip_prefix(arg, "--disk-usage", &arg)) {
> +			if (*arg == '=') {
> +				if (!strcmp(++arg, "human")) {
> +					human_readable = 1;
> +					show_disk_usage = 1;
> +					info.flags |= REV_LIST_QUIET;
> +					continue;
> +				} else
> +					die(_("invalid value for '%s': '%s', try --disk-usage=human"), "--disk-usage", arg);
> +			} else {
> +				show_disk_usage = 1;
> +				info.flags |= REV_LIST_QUIET;
> +				continue;
> +			}
>  		}

We can put the common parts of each side of the conditional into the outer block to avoid repeating ourselves. Also, your code matches --show-disk-usage-without-an-equals, since it nows uses skip_prefix(). You could fix that by checking for '\0' in *arg. So together, something like:

  if (skip_prefix(arg, "--disk-usage", &arg)) {
	if (*arg == '=') {
		if (!strcmp(++arg, "human"))
			human_readable = 1;
		else
			die(...);
	} else if (*arg) {
		/*
		 * Arguably should goto a label to continue chain of ifs?
		 * Doesn't matter unless we try to add --disk-usage-foo
		 * afterwards
		 */
		usage(rev_list_usage);
	}
	show_disk_usage = 1;
	info.flags |= REV_LIST_QUIET;
	continue;
  }

-Peff