git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] Fix git_mkstemp to return an error when path is too long.
@ 2007-07-26  1:32 Carlos Rica
  2007-07-26  3:36 ` Johannes Schindelin
  2007-07-26  4:26 ` Junio C Hamano
  0 siblings, 2 replies; 4+ messages in thread
From: Carlos Rica @ 2007-07-26  1:32 UTC (permalink / raw)
  To: git, Junio C Hamano, Johannes Schindelin

Now the function returns -2 to the caller if the given buffer
is too short to save the entire path for the temporary file.

Signed-off-by: Carlos Rica <jasampler@gmail.com>
---
 diff.c |    2 ++
 path.c |    9 ++++++---
 2 files changed, 8 insertions(+), 3 deletions(-)

diff --git a/diff.c b/diff.c
index cd6b0c4..8735467 100644
--- a/diff.c
+++ b/diff.c
@@ -1694,6 +1694,8 @@ static void prep_temp_blob(struct diff_tempfile *temp,
 	int fd;

 	fd = git_mkstemp(temp->tmp_path, PATH_MAX, ".diff_XXXXXX");
+	if (fd == -2)
+		die("path too long for temp-file");
 	if (fd < 0)
 		die("unable to create temp-file");
 	if (write_in_full(fd, blob, size) != size)
diff --git a/path.c b/path.c
index c4ce962..f33d15d 100644
--- a/path.c
+++ b/path.c
@@ -68,7 +68,8 @@ char *git_path(const char *fmt, ...)
 }


-/* git_mkstemp() - create tmp file honoring TMPDIR variable */
+/* git_mkstemp() - create tmp file honoring TMPDIR variable.
+ * return -2 if path is too long to have it concatenated. */
 int git_mkstemp(char *path, size_t len, const char *template)
 {
 	char *env, *pch = path;
@@ -79,12 +80,14 @@ int git_mkstemp(char *path, size_t len, const char *template)
 		pch += 5;
 	} else {
 		size_t n = snprintf(pch, len, "%s/", env);
-
+		if (n >= len)
+			return -2;
 		len -= n;
 		pch += n;
 	}

-	strlcpy(pch, template, len);
+	if (strlcpy(pch, template, len) >= len)
+		return -2;

 	return mkstemp(path);
 }
-- 
1.5.0

^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH] Fix git_mkstemp to return an error when path is too long.
  2007-07-26  1:32 [PATCH] Fix git_mkstemp to return an error when path is too long Carlos Rica
@ 2007-07-26  3:36 ` Johannes Schindelin
  2007-07-26  4:26 ` Junio C Hamano
  1 sibling, 0 replies; 4+ messages in thread
From: Johannes Schindelin @ 2007-07-26  3:36 UTC (permalink / raw)
  To: Carlos Rica; +Cc: git, Junio C Hamano

Hi,

On Thu, 26 Jul 2007, Carlos Rica wrote:

> @@ -79,12 +80,14 @@ int git_mkstemp(char *path, size_t len, const char *template)
>  		pch += 5;
>  	} else {
>  		size_t n = snprintf(pch, len, "%s/", env);
> -
> +		if (n >= len)
> +			return -2;

That is certainly a bug fixed (even if few people have an insanely long 
TMPDIR...)

>  		len -= n;
>  		pch += n;
>  	}
> 
> -	strlcpy(pch, template, len);
> +	if (strlcpy(pch, template, len) >= len)
> +		return -2;

Maybe just "return error("filename too long: %.*s", 60, pch);"?  So that 
all callers to git_mkstemp() get the message for free?


Ciao,
Dscho

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] Fix git_mkstemp to return an error when path is too long.
  2007-07-26  1:32 [PATCH] Fix git_mkstemp to return an error when path is too long Carlos Rica
  2007-07-26  3:36 ` Johannes Schindelin
@ 2007-07-26  4:26 ` Junio C Hamano
  2007-07-26 19:42   ` Carlos Rica
  1 sibling, 1 reply; 4+ messages in thread
From: Junio C Hamano @ 2007-07-26  4:26 UTC (permalink / raw)
  To: Carlos Rica; +Cc: git, Johannes Schindelin

Carlos Rica <jasampler@gmail.com> writes:

> Now the function returns -2 to the caller if the given buffer
> is too short to save the entire path for the temporary file.
>
> Signed-off-by: Carlos Rica <jasampler@gmail.com>

Trying not to overstep the static buffer is of course a good
thing, and I think it is probably Ok to error out on a TMPDIR
environment variable that is insanely long, instead of using an
extra malloc/free, as underlying mkstemp() would error out if it
is given an insanely long template buffer anyway.

However, I think it is not so useful to return -2, even if that
is done so that you can differentiate the case where the TMPDIR
and/or the template were too long and the case mkstemp() errored
out.

Stop and think for a minute: what does the underlying mkstemp()
do, if the given template is too long?

That's right.  You would get ENAMETOOLONG.  So why don't we do
this instead?

---
 diff.c |    2 +-
 path.c |   25 ++++++++++---------------
 2 files changed, 11 insertions(+), 16 deletions(-)

diff --git a/diff.c b/diff.c
index cd6b0c4..a5fc56b 100644
--- a/diff.c
+++ b/diff.c
@@ -1695,7 +1695,7 @@ static void prep_temp_blob(struct diff_tempfile *temp,
 
 	fd = git_mkstemp(temp->tmp_path, PATH_MAX, ".diff_XXXXXX");
 	if (fd < 0)
-		die("unable to create temp-file");
+		die("unable to create temp-file: %s", strerror(errno));
 	if (write_in_full(fd, blob, size) != size)
 		die("unable to write temp-file");
 	close(fd);
diff --git a/path.c b/path.c
index c4ce962..dc7ded9 100644
--- a/path.c
+++ b/path.c
@@ -71,25 +71,20 @@ char *git_path(const char *fmt, ...)
 /* git_mkstemp() - create tmp file honoring TMPDIR variable */
 int git_mkstemp(char *path, size_t len, const char *template)
 {
-	char *env, *pch = path;
-
-	if ((env = getenv("TMPDIR")) == NULL) {
-		strcpy(pch, "/tmp/");
-		len -= 5;
-		pch += 5;
-	} else {
-		size_t n = snprintf(pch, len, "%s/", env);
-
-		len -= n;
-		pch += n;
+	const char *tmp;
+	size_t n;
+
+	tmp = getenv("TMPDIR");
+	if (!tmp)
+		tmp = "/tmp";
+	n = snprintf(path, len, "%s/%s", tmp, template);
+	if (len <= n) {
+		errno = ENAMETOOLONG;
+		return -1;
 	}
-
-	strlcpy(pch, template, len);
-
 	return mkstemp(path);
 }
 
-
 int validate_headref(const char *path)
 {
 	struct stat st;

^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH] Fix git_mkstemp to return an error when path is too long.
  2007-07-26  4:26 ` Junio C Hamano
@ 2007-07-26 19:42   ` Carlos Rica
  0 siblings, 0 replies; 4+ messages in thread
From: Carlos Rica @ 2007-07-26 19:42 UTC (permalink / raw)
  To: Junio C Hamano; +Cc: git, Johannes Schindelin

I like that new solution. ENAMETOOLONG is fantastic!

Usually, the function is called using the long PATH_MAX constant,
however, the function doesn't force you to supply an insanely long
buffer like that, and therefore is nice to say when the limit is exceeded.

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2007-07-26 19:42 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-07-26  1:32 [PATCH] Fix git_mkstemp to return an error when path is too long Carlos Rica
2007-07-26  3:36 ` Johannes Schindelin
2007-07-26  4:26 ` Junio C Hamano
2007-07-26 19:42   ` Carlos Rica

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).