git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Junio C Hamano <gitster@pobox.com>
To: Kevin Bracey <kevin@bracey.fi>
Cc: git@vger.kernel.org, David Aguilar <davvid@gmail.com>,
	Ciaran Jessup <ciaranj@gmail.com>,
	Scott Chacon <schacon@gmail.com>,
	Alex Riesen <raa.lkml@gmail.com>
Subject: Re: [PATCH v2 2/3] mergetools/p4merge: create a base if none available
Date: Sat, 09 Mar 2013 20:55:55 -0800	[thread overview]
Message-ID: <7v7glfetus.fsf@alter.siamese.dyndns.org> (raw)
In-Reply-To: <1362856860-15205-3-git-send-email-kevin@bracey.fi> (Kevin Bracey's message of "Sat, 9 Mar 2013 21:20:59 +0200")

Kevin Bracey <kevin@bracey.fi> writes:

> diff --git a/git-sh-setup.sh b/git-sh-setup.sh
> index 795edd2..aa9a732 100644
> --- a/git-sh-setup.sh
> +++ b/git-sh-setup.sh
> @@ -249,6 +249,19 @@ clear_local_git_env() {
>  	unset $(git rev-parse --local-env-vars)
>  }
>  
> +# Generate a virtual base file for a two-file merge. On entry the
> +# base file $1 should be a copy of $2. Uses git apply to remove
> +# lines from $1 that are not in $3, leaving only common lines.
> +create_virtual_base() {
> +	sz0=$(wc -c <"$1")
> +	@@DIFF@@ -u -La/"$1" -Lb/"$1" "$2" "$3" | git apply --no-add
> +	sz1=$(wc -c <"$1")
> +
> +	# If we do not have enough common material, it is not
> +	# worth trying two-file merge using common subsections.
> +	expr $sz0 \< $sz1 \* 2 >/dev/null || : >"$1"
> +}
> +

This rewrite is wrong.  It should be

> +	sz0=$(wc -c <"$1")
> +	@@DIFF@@ -u -La/"$1" -Lb/"$1" "$1" "$3" | git apply --no-add
> +	sz1=$(wc -c <"$1")

for it to make sense.  "diff $1 $3" is a change to go from $1 to $3;
with "-La/$1 -Lb/$1", we declare that the change is to be applied to
$1, and use --no-add to only use the removal from the diff when we
edit $1 using this mechanism.

The end effect is to in-place edit "$1" to remove what is not common
with "$3", and sz0/sz1 computation is done on "$1" for this reason.
Does it (i.e. "$1") shrink sufficiently when we remove the material
that is not common in it (i.e. "$1") and "$3"?

This part is a two-file operation between $1 and $3; there is
nothing you would want to pass $2 to influence what the above three
lines do.

It may happen that the caller has two copies of the same thing,
$orig and $src1, and uses one for $1 and the other for $2, so you
won't observe the damage from the incorrect rewriting of the above
logic, but it invites the next caller to incorrectly feed something
totally unrelated to $1 and $2.

Please fix it to a function that takes two temporary paths, not
three.

  reply	other threads:[~2013-03-10  4:56 UTC|newest]

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-03-06 20:32 [PATCH 0/2] Improve P4Merge mergetool invocation Kevin Bracey
2013-03-06 20:32 ` [PATCH 1/2] p4merge: swap LOCAL and REMOTE for mergetool Kevin Bracey
2013-03-07  0:36   ` Junio C Hamano
2013-03-07  6:16     ` Kevin Bracey
2013-03-07  7:23       ` Junio C Hamano
2013-03-07 17:14         ` Kevin Bracey
2013-03-07 19:10           ` Junio C Hamano
2013-03-07 19:50             ` David Aguilar
2013-03-07 20:31               ` Junio C Hamano
2013-03-06 20:32 ` [PATCH 2/2] p4merge: create a virtual base if none available Kevin Bracey
2013-03-07  2:23   ` David Aguilar
2013-03-07  6:28     ` Kevin Bracey
2013-03-07  7:25     ` Junio C Hamano
2013-03-07  3:33   ` David Aguilar
2013-03-09 19:20 ` [PATCH v2 0/3] Improve P4Merge mergetool invocation Kevin Bracey
2013-03-09 19:20   ` [PATCH v2 1/3] mergetools/p4merge: swap LOCAL and REMOTE Kevin Bracey
2013-03-09 19:20   ` [PATCH v2 2/3] mergetools/p4merge: create a base if none available Kevin Bracey
2013-03-10  4:55     ` Junio C Hamano [this message]
2013-03-09 19:21   ` [PATCH v2 3/3] git-merge-one-file: revise merge error reporting Kevin Bracey
2013-03-13  1:12 ` [PATCH v3 1/3] mergetools/p4merge: swap LOCAL and REMOTE Kevin Bracey
2013-03-13  1:12   ` [PATCH v3 2/3] mergetools/p4merge: create a base if none available Kevin Bracey
2013-03-13  1:12   ` [PATCH v3 3/3] git-merge-one-file: revise merge error reporting Kevin Bracey
2013-03-13  9:03     ` David Aguilar
2013-03-24 12:26       ` [PATCH v2 0/3] git-merge-one-file " Kevin Bracey
2013-03-24 12:26         ` [PATCH v2 1/3] git-merge-one-file: style cleanup Kevin Bracey
2013-03-24 12:26         ` [PATCH v2 2/3] git-merge-one-file: send "ERROR:" messages to stderr Kevin Bracey
2013-03-24 12:26         ` [PATCH v2 3/3] git-merge-one-file: revise merge error reporting Kevin Bracey
2013-03-25 17:04           ` Junio C Hamano
2013-03-25 17:17             ` Junio C Hamano
2013-03-25 17:20               ` Junio C Hamano
2013-03-25 19:24               ` Eric Sunshine
2013-03-13 17:57     ` [PATCH v3 " Junio C Hamano
2013-03-14  6:27       ` Kevin Bracey
2013-03-14 14:56         ` Junio C Hamano
2013-03-14 17:31           ` Kevin Bracey
2013-03-14 17:39             ` Kevin Bracey
2013-03-13  2:05   ` [PATCH v3 1/3] mergetools/p4merge: swap LOCAL and REMOTE David Aguilar
2013-03-24 11:54   ` [PATCH v4 1/2] " Kevin Bracey
2013-03-24 11:54     ` [PATCH v4 2/2] mergetools/p4merge: create a base if none available Kevin Bracey
2013-03-25 17:47       ` Junio C Hamano

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=7v7glfetus.fsf@alter.siamese.dyndns.org \
    --to=gitster@pobox.com \
    --cc=ciaranj@gmail.com \
    --cc=davvid@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=kevin@bracey.fi \
    --cc=raa.lkml@gmail.com \
    --cc=schacon@gmail.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 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 NNTP newsgroup(s).