git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: Junio C Hamano <gitster@pobox.com>
To: Jeff King <peff@peff.net>
Cc: git@vger.kernel.org, Duy Nguyen <pclouds@gmail.com>,
	"Shawn O. Pearce" <spearce@spearce.org>
Subject: Re: [PATCH 4/6] introduce a commit metapack
Date: Tue, 29 Jan 2013 09:38:10 -0800	[thread overview]
Message-ID: <7vy5fbq48t.fsf@alter.siamese.dyndns.org> (raw)
In-Reply-To: 20130129091610.GD9999@sigill.intra.peff.net

Jeff King <peff@peff.net> writes:

> +int commit_metapack(unsigned char *sha1,
> +		    uint32_t *timestamp,
> +		    unsigned char **tree,
> +		    unsigned char **parent1,
> +		    unsigned char **parent2)
> +{
> +	struct commit_metapack *p;
> +
> +	prepare_commit_metapacks();
> +	for (p = commit_metapacks; p; p = p->next) {
> +		unsigned char *data;
> +		int pos = sha1_entry_pos(p->index, 20, 0, 0, p->nr, p->nr, sha1);

This is a tangent, but isn't it about time to rip out the check for
GIT_USE_LOOKUP in find_pack_entry_one(), I wonder.

> +	prepare_commit_metapacks();
> +	for (p = commit_metapacks; p; p = p->next) {
> +		unsigned char *data;
> +		int pos = sha1_entry_pos(p->index, 20, 0, 0, p->nr, p->nr, sha1);
> +		if (pos < 0)
> +			continue;
> +
> +		/* timestamp(4) + tree(20) + parents(40) */
> +		data = p->data + 64 * pos;
> +		*timestamp = *(uint32_t *)data;
> +		*timestamp = ntohl(*timestamp);
> +		data += 4;
> +		*tree = data;
> +		data += 20;
> +		*parent1 = data;
> +		data += 20;
> +		*parent2 = data;
> +
> +		return 0;
> +	}
> +
> +	return -1;
> +}

I am torn on this one.

These cached properties of a single commit will not change no matter
which pack it appears in, and it feels logically wrong, especially
when you record these object names in the full SHA-1 form, to tie a
"commit metapack" to a pack.  Logically there needs only one commit
metapack that describes all the commits known to the repository when
the metapack was created.

In order to reduce the disk footprint and I/O cost, the future
direction for this mechanism may want to point into an existing
store of SHA-1 hashes with a shorter file offset, and the .idx file
could be such a store, and in order to move in that direction, you
cannot avoid tying a metapack to a pack.

> +static void get_commits(struct metapack_writer *mw,
> +			const unsigned char *sha1,
> +			void *data)
> +{
> +	struct commit_list ***tail = data;
> +	enum object_type type = sha1_object_info(sha1, NULL);
> +	struct commit *c;
> +
> +	if (type != OBJ_COMMIT)
> +		return;
> +
> +	c = lookup_commit(sha1);
> +	if (!c || parse_commit(c))
> +		die("unable to read commit %s", sha1_to_hex(sha1));
> +
> +	/*
> +	 * Our fixed-size parent list cannot represent root commits, nor
> +	 * octopus merges. Just skip those commits, as we can fallback
> +	 * in those rare cases to reading the actual commit object.
> +	 */
> +	if (!c->parents ||
> +	    (c->parents && c->parents->next && c->parents->next->next))
> +		return;
> +
> +	*tail = &commit_list_insert(c, *tail)->next;
> +}

It feels somewhat wasteful to:

 - use commit_list for this, rather than an array of commit
   objects.  If you have a rough estimate of the number of commits
   in the pack, you could just preallocate a single array and use
   ALLOC_GROW() on it, no?

 - iterate over the .idx file and run sha1_object_info() and
   parse_commit() on many objects in the SHA-1 order.  Iterating in
   the way builtin/pack-objects.c::get_object_details() does avoids
   jumping around in existing packfiles, which may be more
   efficient, no?

> +void commit_metapack_write(const char *idx)
> +{
> +	struct metapack_writer mw;
> +	struct commit_list *commits = NULL, *p;
> +	struct commit_list **tail = &commits;
> +	uint32_t nr = 0;
> +
> +	metapack_writer_init(&mw, idx, "commits", 1);
> +
> +	/* Figure out how many eligible commits we've got in this pack. */
> +	metapack_writer_foreach(&mw, get_commits, &tail);
> +	for (p = commits; p; p = p->next)
> +		nr++;

  parent reply	other threads:[~2013-01-29 17:38 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-01-29  9:14 [PATCH/RFC 0/6] commit caching Jeff King
2013-01-29  9:15 ` [PATCH 1/6] csum-file: make sha1write const-correct Jeff King
2013-01-29  9:15 ` [PATCH 2/6] strbuf: add string-chomping functions Jeff King
2013-01-29 10:15   ` Michael Haggerty
2013-01-29 11:10     ` Jeff King
2013-01-30  5:00       ` Michael Haggerty
2013-01-29  9:15 ` [PATCH 3/6] introduce pack metadata cache files Jeff King
2013-01-29 17:35   ` Junio C Hamano
2013-01-30  6:47     ` Jeff King
2013-01-30  1:30   ` Duy Nguyen
2013-01-30  6:50     ` Jeff King
2013-01-29  9:16 ` [PATCH 4/6] introduce a commit metapack Jeff King
2013-01-29 10:24   ` Michael Haggerty
2013-01-29 11:13     ` Jeff King
2013-01-29 17:38   ` Junio C Hamano [this message]
2013-01-29 18:08     ` Junio C Hamano
2013-01-30  7:12       ` Jeff King
2013-01-30  7:17         ` Junio C Hamano
2013-02-01  9:21           ` Jeff King
2013-01-30 15:56         ` Junio C Hamano
2013-01-31 17:03           ` Shawn Pearce
2013-02-01  9:42             ` Jeff King
2013-02-02 17:49               ` Junio C Hamano
2013-01-30  7:07     ` Jeff King
2013-01-30  3:36   ` Duy Nguyen
2013-01-30  7:12     ` Jeff King
2013-01-30 13:56   ` Duy Nguyen
2013-01-30 14:16     ` Duy Nguyen
2013-01-31 11:06       ` Duy Nguyen
2013-02-01 10:15         ` Jeff King
2013-02-02  9:49           ` Duy Nguyen
2013-02-01 10:40         ` Jeff King
2013-03-17 13:21         ` Duy Nguyen
2013-03-18 12:20           ` Jeff King
2013-02-01 10:00     ` Jeff King
2013-01-29  9:16 ` [PATCH 5/6] add git-metapack command Jeff King
2013-01-29  9:16 ` [PATCH 6/6] commit: look up commit info in metapack Jeff King
2013-01-30  3:31 ` [PATCH/RFC 0/6] commit caching Duy Nguyen
2013-01-30  7:18   ` Jeff King
2013-01-30  8:32     ` Duy Nguyen
2013-01-31 17:14 ` Shawn Pearce
2013-02-01  9:11   ` Jeff King
2013-02-02 10:04     ` Shawn Pearce

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

  List information: http://vger.kernel.org/majordomo-info.html

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

  git send-email \
    --in-reply-to=7vy5fbq48t.fsf@alter.siamese.dyndns.org \
    --to=gitster@pobox.com \
    --cc=git@vger.kernel.org \
    --cc=pclouds@gmail.com \
    --cc=peff@peff.net \
    --cc=spearce@spearce.org \
    /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.
Code repositories for project(s) associated with this public inbox

	https://80x24.org/mirrors/git.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).