From: Junio C Hamano <gitster@pobox.com>
To: Tay Ray Chuan <rctay89@gmail.com>
Cc: "Git Mailing List" <git@vger.kernel.org>,
"Jeff King" <peff@peff.net>, Thomas Rast <trast@student.ethz.ch>
Subject: Re: [PATCH v2 2/4] help.c::exclude_cmds: realloc() before copy, plug a leak
Date: Wed, 25 Jul 2012 10:39:37 -0700 [thread overview]
Message-ID: <7v394fd9k6.fsf@alter.siamese.dyndns.org> (raw)
In-Reply-To: <1343232982-10540-3-git-send-email-rctay89@gmail.com> (Tay Ray Chuan's message of "Thu, 26 Jul 2012 00:16:20 +0800")
Tay Ray Chuan <rctay89@gmail.com> writes:
> Copying with structural assignment may not take into account that the
> LHS struct has sufficient memory, especially since the cmdname->name
> member is nonfixed in size. Be unambiguous about it by realloc()'ing it
> to be of sufficient size.
If the original code were
*(cmd->names[cj++]) = *(cmd->names[ci++]);
there may be a structural assignment involved, but
cmds->names[dst] = cmd->names[src]
just copies the pointer that points at a struct cmdname that records
the src command name to another slot of cmds->names[] array, whose
elements are pointers, no? What's there to realloc?
> @@ -58,20 +69,25 @@ void exclude_cmds(struct cmdnames *cmds, struct cmdnames *excludes)
> {
> int ci, cj, ei;
> int cmp;
> + int last_cj;
>
> ci = cj = ei = 0;
> while (ci < cmds->cnt && ei < excludes->cnt) {
> cmp = strcmp(cmds->names[ci]->name, excludes->names[ei]->name);
> if (cmp < 0)
> - cmds->names[cj++] = cmds->names[ci++];
> + copy_cmdname(&cmds->names[cj++], cmds->names[ci++]);
> else if (cmp == 0)
> ci++, ei++;
> else if (cmp > 0)
> ei++;
> }
> + last_cj = cj;
>
> while (ci < cmds->cnt)
> - cmds->names[cj++] = cmds->names[ci++];
> + copy_cmdname(&cmds->names[cj++], cmds->names[ci++]);
> +
> + while (last_cj < cmds->cnt)
> + free(cmds->names[last_cj++]);
>
> cmds->cnt = cj;
> }
We shifted cmds->names[] array to skip entries that appear in
excludes. If original cmds->names[] had "0", "1", "2", "3", ...
and excludes had "0" and "1", cmds->names[] would contain "2", "3",
"2", "3"; the first two are copied over "0" and "1" that are
excluded, and the latter two are leftover beyond last_cj. The
corresponding names share the same structure (cmds->names[] is an
array of pointers). Doesn't freeing cmds->names[2] free the
structure that is used by both cmds->names[0] and cmds->names[2]?
Confused.
The function drops cmds->names[ci] when it appears in excludes, so
you may want to free it when it happens, though.
help.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/help.c b/help.c
index 6991492..cae389b 100644
--- a/help.c
+++ b/help.c
@@ -64,9 +64,10 @@ void exclude_cmds(struct cmdnames *cmds, struct cmdnames *excludes)
cmp = strcmp(cmds->names[ci]->name, excludes->names[ei]->name);
if (cmp < 0)
cmds->names[cj++] = cmds->names[ci++];
- else if (cmp == 0)
- ci++, ei++;
- else if (cmp > 0)
+ else if (cmp == 0) {
+ ei++;
+ free(cmd->names[ci++]);
+ } else if (cmp > 0)
ei++;
}
next prev parent reply other threads:[~2012-07-25 17:39 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-05-06 6:55 [PATCH 0/4] allow recovery from command name typos Tay Ray Chuan
2012-05-06 6:55 ` [PATCH 1/4] help.c::uniq: plug a leak Tay Ray Chuan
2012-05-06 6:55 ` [PATCH 2/4] help.c::exclude_cmds: " Tay Ray Chuan
2012-05-06 6:55 ` [PATCH 3/4] help.c: plug a leak when help.autocorrect is set Tay Ray Chuan
2012-05-06 6:55 ` [PATCH 4/4] allow recovery from command name typos Tay Ray Chuan
2012-05-06 8:21 ` Jeff King
2012-05-06 16:07 ` Tay Ray Chuan
2012-05-07 9:43 ` Thomas Rast
2012-05-07 15:49 ` Tay Ray Chuan
2012-05-07 17:41 ` Junio C Hamano
2012-05-09 15:06 ` Tay Ray Chuan
2012-05-09 17:03 ` Junio C Hamano
[not found] ` <CAOBOgRaDEgAqXWmdC6hrudkL5OwzeMffbj2RtKMxf2TsYWzotA@mail.gmail.com>
2012-05-06 16:04 ` Tay Ray Chuan
2012-05-06 8:12 ` [PATCH 1/4] help.c::uniq: plug a leak Jeff King
2012-05-06 15:54 ` Tay Ray Chuan
2012-05-07 7:30 ` Jeff King
2012-07-25 16:16 ` [PATCH v2 0/4] allow recovery from command name typos Tay Ray Chuan
2012-07-25 16:16 ` [PATCH v2 1/4] help.c::uniq: plug a leak Tay Ray Chuan
2012-07-25 16:16 ` [PATCH v2 2/4] help.c::exclude_cmds: realloc() before copy, " Tay Ray Chuan
2012-07-25 16:16 ` [PATCH v2 3/4] help.c: plug leaks with(out) help.autocorrect Tay Ray Chuan
2012-07-25 16:16 ` [PATCH v2 4/4] allow recovery from command name typos Tay Ray Chuan
2012-07-25 17:57 ` Junio C Hamano
2012-07-26 17:08 ` Tay Ray Chuan
2012-07-26 17:26 ` Jeff King
2012-07-26 17:59 ` Junio C Hamano
2012-07-26 18:37 ` Jeff King
2012-07-26 17:53 ` Junio C Hamano
2012-07-25 17:47 ` [PATCH v2 3/4] help.c: plug leaks with(out) help.autocorrect Junio C Hamano
2012-07-25 17:39 ` Junio C Hamano [this message]
2012-08-05 18:45 ` [PATCH v3 0/2] allow recovery from command name typos Tay Ray Chuan
2012-08-05 18:45 ` [PATCH v3 1/2] add interface for /dev/tty interaction Tay Ray Chuan
2012-08-05 18:45 ` [PATCH v3 2/2] allow recovery from command name typos Tay Ray Chuan
2012-08-06 0:50 ` Junio C Hamano
2012-08-05 20:11 ` [PATCH v3 1/2] add interface for /dev/tty interaction Junio C Hamano
2012-08-06 19:45 ` Jeff King
2012-08-06 19:56 ` Jeff King
2012-08-06 20:01 ` 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=7v394fd9k6.fsf@alter.siamese.dyndns.org \
--to=gitster@pobox.com \
--cc=git@vger.kernel.org \
--cc=peff@peff.net \
--cc=rctay89@gmail.com \
--cc=trast@student.ethz.ch \
/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).