[PATCH v2 0/2] Check for lock failures early

Callers outside of refs.c use either lock_ref_sha1() or 
lock_any_ref_for_update() to lock a ref during an update.

Two of these places we do not immediately check the lock for failure
making reading the code harder.

One place we do some unrelated string manipulation fucntions before we
check for failure and the other place we rely on that write_ref_sha1()
will check the lock for failure and return an error.


These two patches updates these two places so that we immediately check the
lock for failure and act on it.
It does not change any functionality or logic but makes the code easier to
read by being more consistent.

Version 2:
* Simplify the return on error case in sequencer.c.


Ronnie Sahlberg (2):
  sequencer.c: check for lock failure and bail early in fast_forward_to
  commit.c: check for lock error and return early

 builtin/commit.c | 8 ++++----
 sequencer.c      | 4 ++++
 2 files changed, 8 insertions(+), 4 deletions(-)

-- 
1.9.1.504.g5a62d94

Move the check for the lock failure to happen immediately after
lock_any_ref_for_update().
Previously the lock and the check-if-lock-failed was separated by a handful
of string manipulation statements.

Moving the check to occur immediately after the failed lock makes the
code slightly easier to read and makes it follow the pattern of
 try-to-take-a-lock()
 if (check-if-lock-failed){
    error
 }
---
 builtin/commit.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/builtin/commit.c b/builtin/commit.c
index d9550c5..c6320f1 100644
--- a/builtin/commit.c
+++ b/builtin/commit.c

@@ -1672,6 +1672,10 @@ int cmd_commit(int argc, const char **argv, const char *prefix)
 					   ? NULL
 					   : current_head->object.sha1,
 					   0, NULL);
+	if (!ref_lock) {
+		rollback_index_files();
+		die(_("cannot lock HEAD ref"));
+	}
 
 	nl = strchr(sb.buf, '\n');
 	if (nl)

@@ -1681,10 +1685,6 @@ int cmd_commit(int argc, const char **argv, const char *prefix)
 	strbuf_insert(&sb, 0, reflog_msg, strlen(reflog_msg));
 	strbuf_insert(&sb, strlen(reflog_msg), ": ", 2);
 
-	if (!ref_lock) {
-		rollback_index_files();
-		die(_("cannot lock HEAD ref"));
-	}
 	if (write_ref_sha1(ref_lock, sha1, sb.buf) < 0) {
 		rollback_index_files();
 		die(_("cannot update HEAD ref"));

-- 
1.9.1.504.g5a62d94

Change fast_forward_to() to check if locking the ref failed, print a nice
error message and bail out early.
The old code did not check if ref_lock was NULL and relied on the fact
that the write_ref_sha1() would safely detect this condition and set the
return variable ret to indicate an error.
While that is safe, it makes the code harder to read for two reasons:
* Inconsistency.  Almost all other places we do check the lock for NULL
  explicitely, so the naive reader is confused "why don't we check here".
* And relying on write_ref_sha1() to detect and return an error for when
  a previous lock_any_ref_for_update() feels obfuscated.

This change should not change any functionality or logic
aside from adding an extra error message when this condition is triggered.
(write_ref_sha1() returns an error silently for this condition)

Signed-off-by: Ronnie Sahlberg <redacted>
---
 sequencer.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/sequencer.c b/sequencer.c
index bde5f04..0a80c58 100644
--- a/sequencer.c
+++ b/sequencer.c

@@ -281,8 +281,12 @@ static int fast_forward_to(const unsigned char *to, const unsigned char *from,
 		exit(1); /* the callee should have complained already */
 	ref_lock = lock_any_ref_for_update("HEAD", unborn ? null_sha1 : from,
 					   0, NULL);
+	if (!ref_lock)
+		return error(_("Failed to lock HEAD during fast_forward_to"));
+
 	strbuf_addf(&sb, "%s: fast-forward", action_name(opts));
 	ret = write_ref_sha1(ref_lock, to, sb.buf);
+
 	strbuf_release(&sb);
 	return ret;
 }

-- 
1.9.1.504.g5a62d94

On 04/16/2014 08:56 PM, Ronnie Sahlberg wrote:

Change fast_forward_to() to check if locking the ref failed, print a nice
error message and bail out early.
The old code did not check if ref_lock was NULL and relied on the fact
that the write_ref_sha1() would safely detect this condition and set the

s/the write_ref_sha1()/write_ref_sha1()/

return variable ret to indicate an error.
While that is safe, it makes the code harder to read for two reasons:
* Inconsistency.  Almost all other places we do check the lock for NULL
  explicitely, so the naive reader is confused "why don't we check here".

s/explicitely/explicitly/
s/here"/here?"/

* And relying on write_ref_sha1() to detect and return an error for when
  a previous lock_any_ref_for_update() feels obfuscated.

s/feels/failed feels/ maybe?

This change should not change any functionality or logic
aside from adding an extra error message when this condition is triggered.
(write_ref_sha1() returns an error silently for this condition)

You need a period inside the parentheses.

This error message can be emitted to the user in the normal course of
things (i.e., it is not a bug).  So the message should make sense to the
user.  Is "fast_forward_to" a user-facing term that the user will
understand?  I suspect that you took it from the name of the function,
which is *not* meaningful to a user.

But unfortunately I'm not familiar enough with the sequencer to be able
to suggest a better error message.

+
 	strbuf_addf(&sb, "%s: fast-forward", action_name(opts));
 	ret = write_ref_sha1(ref_lock, to, sb.buf);
+
 	strbuf_release(&sb);
 	return ret;
 }

Michael

-- 
Michael Haggerty
mhagger@alum.mit.edu
http://softwareswirl.blogspot.com/

On 04/16/2014 08:56 PM, Ronnie Sahlberg wrote:

Move the check for the lock failure to happen immediately after
lock_any_ref_for_update().
Previously the lock and the check-if-lock-failed was separated by a handful
of string manipulation statements.

Please flow sentences together into paragraphs for easier reading,
rather than having an extremely ragged right-hand margin.

The rest looks good.

Michael

-- 
Michael Haggerty
mhagger@alum.mit.edu
http://softwareswirl.blogspot.com/