All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jonathan Nieder <jrnieder@gmail.com>
To: Stefan Beller <sbeller@google.com>
Cc: git@vger.kernel.org, gitster@pobox.com,
	Ronnie Sahlberg <sahlberg@google.com>
Subject: Re: [PATCH v6] refs.c: use a string_list for repack_without_refs
Date: Thu, 20 Nov 2014 12:36:48 -0800	[thread overview]
Message-ID: <20141120203648.GI6527@google.com> (raw)
In-Reply-To: <1416514066-17049-1-git-send-email-sbeller@google.com>

Stefan Beller wrote:

> Signed-off-by: Ronnie Sahlberg <sahlberg@google.com>
> Signed-off-by: Stefan Beller <sbeller@google.com>
> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
> ---

Yep, looks good now.  Thanks for bearing with me.

[...]
> +++ b/refs.h
> @@ -163,8 +163,16 @@ extern void rollback_packed_refs(void);
[...]
> +/*
> + * Remove the refs listed in the unsorted string list 'without' from
> + * the packed-refs file. On error, packed-refs will be unchanged, the
> + * return value is nonzero, and a message about the error is written
> + * to the 'err' strbuf.
> + *
> + * The refs in 'without' may be unsorted.
> + * 'err' must not be NULL.

I think we've gone back and forth enough on this text and it's not
worth the transactional cost to tweak it further, so I'm not
suggesting a change --- just explaining how I read it for future
reference.

"may be unsorted" is confusing to me.  It sounds like the reader of
this comment (someone calling repack_without_refs) has to be prepared
for that possibility.  But we are saying the opposite --- not "be
prepared", but "don't worry about sorting 'without', since
repack_without_refs can handle it".

It's also redundant, since the paragraph above already says that
'without' is an unsorted string list.

The way I see it, there are four types that for various reasons (lack
of language-level support for subclassing, etc) are conflated into a
single struct in the string-list API:

 * sorted string list that owns its items (i.e., created with DUP)
 * sorted string list that does not own its items (i.e., created with NODUP)
 * unsorted string list that owns its items
 * unsorted string list that does not own its items

Different functions are valid to call on each type, as documented in
the comments in string-list.h.

repack_without_refs accepts all 4 types of string-list.  That's what
it means when the documentation says its argument is unsorted.

Thanks,
Jonathan

  reply	other threads:[~2014-11-20 20:36 UTC|newest]

Thread overview: 61+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-11-18 22:43 [PATCH] refs.c: use a stringlist for repack_without_refs Stefan Beller
2014-11-18 23:06 ` Junio C Hamano
2014-11-18 23:39   ` Junio C Hamano
2014-11-18 23:45 ` Jonathan Nieder
2014-11-19  0:28   ` Stefan Beller
2014-11-19  1:08   ` Stefan Beller
2014-11-19 18:00     ` Junio C Hamano
2014-11-19 18:50       ` Stefan Beller
2014-11-19 20:44         ` Jonathan Nieder
2014-11-19 21:54           ` Stefan Beller
2014-11-19 21:59             ` [PATCH v4] " Stefan Beller
2014-11-20  2:15               ` Jonathan Nieder
2014-11-20 16:47                 ` Junio C Hamano
2014-11-20 18:04                 ` [PATCH v5 1/1] " Stefan Beller
2014-11-20 18:10                   ` [PATCH] refs.c: repack_without_refs may be called without error string buffer Stefan Beller
2014-11-20 18:15                     ` Ronnie Sahlberg
2014-11-20 18:35                     ` Jonathan Nieder
2014-11-20 18:36                       ` Ronnie Sahlberg
2014-11-20 18:56                         ` Stefan Beller
2014-11-20 18:29                   ` [PATCH v5 1/1] refs.c: use a stringlist for repack_without_refs Jonathan Nieder
2014-11-20 18:37                   ` Jonathan Nieder
2014-11-20 19:01                   ` Junio C Hamano
2014-11-20 19:05                     ` Stefan Beller
2014-11-20 20:07                       ` [PATCH v6] refs.c: use a string_list " Stefan Beller
2014-11-20 20:36                         ` Jonathan Nieder [this message]
2014-11-21 14:09         ` [PATCH 0/6] repack_without_refs(): convert to string_list Michael Haggerty
2014-11-21 14:09           ` [PATCH 1/6] prune_remote(): exit early if there are no stale references Michael Haggerty
2014-11-22 21:07             ` Jonathan Nieder
2014-11-21 14:09           ` [PATCH 2/6] prune_remote(): initialize both delete_refs lists in a single loop Michael Haggerty
2014-11-21 14:09           ` [PATCH 3/6] prune_remote(): sort delete_refs_list references en masse Michael Haggerty
2014-11-21 16:44             ` Junio C Hamano
2014-11-25  7:21               ` Michael Haggerty
2014-11-25  8:04                 ` Michael Haggerty
2014-11-22 21:08             ` Jonathan Nieder
2014-11-21 14:09           ` [PATCH 4/6] repack_without_refs(): make the refnames argument a string_list Michael Haggerty
2014-11-22 21:17             ` Jonathan Nieder
2014-11-25  7:42               ` Michael Haggerty
2014-11-21 14:09           ` [PATCH 5/6] prune_remote(): rename local variable Michael Haggerty
2014-11-22 21:18             ` Jonathan Nieder
2014-11-21 14:09           ` [PATCH 6/6] prune_remote(): iterate using for_each_string_list_item() Michael Haggerty
2014-11-22 21:19             ` Jonathan Nieder
2014-11-21 14:25           ` [PATCH 0/6] repack_without_refs(): convert to string_list Michael Haggerty
2014-11-21 18:00             ` Junio C Hamano
2014-11-21 19:57               ` Stefan Beller
2014-11-25  0:28               ` Our cumbersome mailing list workflow (was: Re: [PATCH 0/6] repack_without_refs(): convert to string_list) Michael Haggerty
2014-11-27 17:46                 ` Our cumbersome mailing list workflow Torsten Bögershausen
2014-11-27 18:24                   ` Matthieu Moy
2014-11-28 12:09                     ` Philip Oakley
2014-11-27 22:53                   ` Eric Wong
2014-11-28 15:34                     ` Michael Haggerty
2014-11-28 16:24                       ` brian m. carlson
2014-12-01  2:46                       ` Junio C Hamano
2014-12-03  2:20                         ` Stefan Beller
2014-12-03  3:53                           ` Jonathan Nieder
2014-12-03 17:18                             ` Junio C Hamano
2014-12-03 17:28                           ` Torsten Bögershausen
2014-11-28 14:31                   ` Michael Haggerty
2014-11-28 15:42                     ` Marc Branchaud
2014-11-28 21:39                       ` Damien Robert
2014-12-03 23:57                 ` Philip Oakley
2014-12-04  2:03                   ` Stefan Beller

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=20141120203648.GI6527@google.com \
    --to=jrnieder@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=sahlberg@google.com \
    --cc=sbeller@google.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.