public inbox for git@vger.kernel.org 
 help / color / mirror / Atom feed
From: Junio C Hamano <gitster@pobox•com>
To: Nicolas Pitre <nico@fluxnic•net>
Cc: git@vger•kernel.org
Subject: Re: [PATCH 12/23] pack v4: creation code
Date: Tue, 27 Aug 2013 08:48:24 -0700	[thread overview]
Message-ID: <xmqqppszdtiv.fsf@gitster.dls.corp.google.com> (raw)
In-Reply-To: <1377577567-27655-13-git-send-email-nico@fluxnic.net> (Nicolas Pitre's message of "Tue, 27 Aug 2013 00:25:56 -0400")

Nicolas Pitre <nico@fluxnic•net> writes:

> Let's actually open the destination pack file and write the header and
> the tables.
>
> The header isn't much different from pack v3, except for the pack version
> number of course.
>
> The first table is the sorted SHA1 table normally found in the pack index
> file.  With pack v4 we write this table in the main pack file instead as
> it is index referenced by subsequent objects in the pack.  Doing so has
> many advantages:
>
> - The SHA1 references used to be duplicated on disk: once in the pack
>   index file, and then at least once or more within commit and tree
>   objects referencing them.  The only SHA1 which is not being listed more
>   than once this way is the one for a branch tip commit object and those
>   are normally very few.  Now all that SHA1 data is represented only once.
>

This tickles my curiosity. Why isn't this SHA-1 table sorted by
reference count the same way as the tree path and the people name
tables to keep the average length of varint references short?

> - The SHA1 references found in commit and tree objects can be obtained
>   on disk directly without having to deflate those objects first.
>
> The SHA1 table size is obtained by multiplying the number of objects by 20.
>
> And then the commit and path dictionary tables are written right after
> the SHA1 table.

> Signed-off-by: Nicolas Pitre <nico@fluxnic•net>
> ---
>  packv4-create.c | 60 ++++++++++++++++++++++++++++++++++++++++++++++++++++-----
>  1 file changed, 55 insertions(+), 5 deletions(-)
>
> diff --git a/packv4-create.c b/packv4-create.c
> index 2956fda..5211f9c 100644
> --- a/packv4-create.c
> +++ b/packv4-create.c
> @@ -605,6 +605,48 @@ static unsigned long write_dict_table(struct sha1file *f, struct dict_table *t)
>  	return hdrlen + datalen;
>  }
>  
> +static struct sha1file * packv4_open(char *path)
> +{
> +	int fd;
> +
> +	fd = open(path, O_CREAT|O_EXCL|O_WRONLY, 0600);
> +	if (fd < 0)
> +		die_errno("unable to create '%s'", path);
> +	return sha1fd(fd, path);
> +}
> +
> +static unsigned int packv4_write_header(struct sha1file *f, unsigned nr_objects)
> +{
> +	struct pack_header hdr;
> +
> +	hdr.hdr_signature = htonl(PACK_SIGNATURE);
> +	hdr.hdr_version = htonl(4);
> +	hdr.hdr_entries = htonl(nr_objects);
> +	sha1write(f, &hdr, sizeof(hdr));
> +
> +	return sizeof(hdr);
> +}
> +
> +static unsigned long packv4_write_tables(struct sha1file *f, unsigned nr_objects,
> +					 struct pack_idx_entry *objs)
> +{
> +	unsigned i;
> +	unsigned long written = 0;
> +
> +	/* The sorted list of object SHA1's is always first */
> +	for (i = 0; i < nr_objects; i++)
> +		sha1write(f, objs[i].sha1, 20);
> +	written = 20 * nr_objects;
> +
> +	/* Then the commit dictionary table */
> +	written += write_dict_table(f, commit_name_table);
> +
> +	/* Followed by the path component dictionary table */
> +	written += write_dict_table(f, tree_path_table);
> +
> +	return written;
> +}
> +
>  static struct packed_git *open_pack(const char *path)
>  {
>  	char arg[PATH_MAX];
> @@ -658,9 +700,10 @@ static struct packed_git *open_pack(const char *path)
>  	return p;
>  }
>  
> -static void process_one_pack(char *src_pack)
> +static void process_one_pack(char *src_pack, char *dst_pack)
>  {
>  	struct packed_git *p;
> +	struct sha1file *f;
>  	struct pack_idx_entry *objs, **p_objs;
>  	unsigned nr_objects;
>  
> @@ -673,15 +716,22 @@ static void process_one_pack(char *src_pack)
>  	p_objs = sort_objs_by_offset(objs, nr_objects);
>  
>  	create_pack_dictionaries(p, p_objs);
> +
> +	f = packv4_open(dst_pack);
> +	if (!f)
> +		die("unable to open destination pack");
> +	packv4_write_header(f, nr_objects);
> +	packv4_write_tables(f, nr_objects, objs);
>  }
>  
>  int main(int argc, char *argv[])
>  {
> -	if (argc != 2) {
> -		fprintf(stderr, "Usage: %s <packfile>\n", argv[0]);
> +	if (argc != 3) {
> +		fprintf(stderr, "Usage: %s <src_packfile> <dst_packfile>\n", argv[0]);
>  		exit(1);
>  	}
> -	process_one_pack(argv[1]);
> -	dict_dump();
> +	process_one_pack(argv[1], argv[2]);
> +	if (0)
> +		dict_dump();
>  	return 0;
>  }

  reply	other threads:[~2013-08-27 15:48 UTC|newest]

Thread overview: 83+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-08-27  4:25 [PATCH 00/23] Preliminary pack v4 support Nicolas Pitre
2013-08-27  4:25 ` [PATCH 01/23] pack v4: initial pack dictionary structure and code Nicolas Pitre
2013-08-27 15:08   ` Junio C Hamano
2013-08-27 16:13     ` Nicolas Pitre
2013-08-27  4:25 ` [PATCH 02/23] export packed_object_info() Nicolas Pitre
2013-08-27  4:25 ` [PATCH 03/23] pack v4: scan tree objects Nicolas Pitre
2013-08-27  4:25 ` [PATCH 04/23] pack v4: add tree entry mode support to dictionary entries Nicolas Pitre
2013-08-27  4:25 ` [PATCH 05/23] pack v4: add commit object parsing Nicolas Pitre
2013-08-27 15:26   ` Junio C Hamano
2013-08-27 16:47     ` Nicolas Pitre
2013-08-27 17:42       ` Junio C Hamano
2013-08-27  4:25 ` [PATCH 06/23] pack v4: split the object list and dictionary creation Nicolas Pitre
2013-08-27  4:25 ` [PATCH 07/23] pack v4: move to struct pack_idx_entry and get rid of our own struct idx_entry Nicolas Pitre
2013-08-27  4:25 ` [PATCH 08/23] pack v4: basic references encoding Nicolas Pitre
2013-08-27 15:29   ` Junio C Hamano
2013-08-27 15:53     ` Nicolas Pitre
2013-08-27  4:25 ` [PATCH 09/23] pack v4: commit object encoding Nicolas Pitre
2013-08-27 15:39   ` Junio C Hamano
2013-08-27 16:50     ` Nicolas Pitre
2013-08-27 19:59     ` Nicolas Pitre
2013-08-27 20:15       ` Junio C Hamano
2013-08-27 21:43         ` Nicolas Pitre
2013-09-02 20:48   ` Duy Nguyen
2013-09-03  6:30     ` Nicolas Pitre
2013-09-03  7:41       ` Duy Nguyen
2013-09-05  3:50         ` Nicolas Pitre
2013-08-27  4:25 ` [PATCH 10/23] pack v4: tree " Nicolas Pitre
2013-08-27 15:44   ` Junio C Hamano
2013-08-27 16:52     ` Nicolas Pitre
2013-08-27  4:25 ` [PATCH 11/23] pack v4: dictionary table output Nicolas Pitre
2013-08-27  4:25 ` [PATCH 12/23] pack v4: creation code Nicolas Pitre
2013-08-27 15:48   ` Junio C Hamano [this message]
2013-08-27 16:59     ` Nicolas Pitre
2013-08-27  4:25 ` [PATCH 13/23] pack v4: object headers Nicolas Pitre
2013-08-27  4:25 ` [PATCH 14/23] pack v4: object data copy Nicolas Pitre
2013-08-27 15:53   ` Junio C Hamano
2013-08-27 18:24     ` Nicolas Pitre
2013-08-27  4:25 ` [PATCH 15/23] pack v4: object writing Nicolas Pitre
2013-08-27  4:26 ` [PATCH 16/23] pack v4: tree object delta encoding Nicolas Pitre
2013-08-27  4:26 ` [PATCH 17/23] pack v4: load delta candidate for encoding tree objects Nicolas Pitre
2013-08-27  4:26 ` [PATCH 18/23] pack v4: honor pack.compression config option Nicolas Pitre
2013-08-27  4:26 ` [PATCH 19/23] pack v4: relax commit parsing a bit Nicolas Pitre
2013-08-27  4:26 ` [PATCH 20/23] pack index v3 Nicolas Pitre
2013-08-27  4:26 ` [PATCH 21/23] pack v4: normalize pack name to properly generate the pack index file name Nicolas Pitre
2013-08-27  4:26 ` [PATCH 22/23] pack v4: add progress display Nicolas Pitre
2013-08-27  4:26 ` [PATCH 23/23] initial pack index v3 support on the read side Nicolas Pitre
2013-08-31 11:45   ` Duy Nguyen
2013-09-03  6:09     ` Nicolas Pitre
2013-09-03  7:34       ` Duy Nguyen
2013-08-27 11:17 ` [PATCH] Document pack v4 format Nguyễn Thái Ngọc Duy
2013-08-27 18:25   ` Junio C Hamano
2013-08-27 18:53   ` Nicolas Pitre
2013-08-31  2:49   ` [PATCH v2] " Nguyễn Thái Ngọc Duy
2013-09-03  6:00     ` Nicolas Pitre
2013-09-03  6:46       ` Nicolas Pitre
2013-09-03 11:49         ` Duy Nguyen
2013-09-03 14:54           ` Duy Nguyen
2013-09-05  4:12             ` Nicolas Pitre
2013-09-05  4:19               ` Duy Nguyen
2013-09-05  4:40                 ` Nicolas Pitre
2013-09-05  5:04                   ` Duy Nguyen
2013-09-05  5:39                     ` Nicolas Pitre
2013-09-05 16:52                       ` Duy Nguyen
2013-09-05 17:14                         ` Nicolas Pitre
2013-09-05 20:26                           ` Junio C Hamano
2013-09-05 21:04                             ` Nicolas Pitre
2013-09-06  4:18                         ` Duy Nguyen
2013-09-06 13:19                           ` Nicolas Pitre
2013-09-06  2:14     ` [PATCH v3] " Nguyễn Thái Ngọc Duy
2013-09-06  3:23       ` Nicolas Pitre
2013-09-06  9:48         ` Duy Nguyen
2013-09-06 13:25           ` Nicolas Pitre
2013-09-06 13:44             ` Duy Nguyen
2013-09-06 16:44               ` Nicolas Pitre
2013-09-07  4:57                 ` Nicolas Pitre
2013-09-07  4:52       ` Nicolas Pitre
2013-09-07  8:05         ` Duy Nguyen
2013-08-27 15:03 ` [PATCH 00/23] Preliminary pack v4 support Junio C Hamano
2013-08-27 15:59   ` Nicolas Pitre
2013-08-27 16:44     ` Junio C Hamano
2013-08-28  2:30       ` Duy Nguyen
2013-08-28  2:58         ` Nicolas Pitre
2013-08-28  3:06           ` Duy Nguyen

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=xmqqppszdtiv.fsf@gitster.dls.corp.google.com \
    --to=gitster@pobox$(echo .)com \
    --cc=git@vger$(echo .)kernel.org \
    --cc=nico@fluxnic$(echo .)net \
    /path/to/YOUR_REPLY

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