Re: [PATCH 6/8] imap-send: change msg_data from storing (char *, len) to storing strbuf

From: Michael Haggerty <mhagger@alum.mit.edu>
Date: Nov 30, 2012, 13:36 UTC
Message-ID: <50B8B66F.3090300@alum.mit.edu>
In-Reply-To: <7vboegp04x.fsf@alter.siamese.dyndns.org>

On 11/29/2012 10:30 PM, Junio C Hamano wrote:

Show 6 quoted lines

> Michael Haggerty <mhagger@alum.mit.edu> writes:
> 
>> struct msg_data stored (char *, len) of the data to be included in a
> 
> That (<type>, <varname>) is a bit funny notation, even though it is
> understandable.

I understand that it is funny, but it seems like the clearest way to express what is meant in a way that fits in the summary line. Feel free to change it if you like.

Show 11 quoted lines

>> message, kept the character data NUL-terminated, etc., much like a
>> strbuf would do.  So change it to use a struct strbuf.  This makes the
>> code clearer and reduces copying a little bit.
>>
>> A side effect of this change is that the memory for each message is
>> freed after it is used rather than leaked, though that detail is
>> unimportant given that imap-send is a top-level command.
>>
>> --
> 
> ?

If by "?" you are wondering where the memory leak was, it was:

* The while loop in main() called split_msg()

  * split_msg() cleared the msg_data structure using
    memset(msg, 0, sizeof *msg)

  * split_msg() copied the first message out of all_msgs using
    xmemdupz() and stored the result to msg->data

* The msg_data was passed to imap_store_msg().  Its contents were
  copied to cb.data (which will be freed in the imap functions) but
  the original was left unfreed.

* The next time through the loop, split_msg() zeroed the msg_data
  structure again, thus discarding the pointer to the xmemdupz()ed
  memory.

The leak caused more memory than necessary to be allocated (worst case: nearly the total size of all_msgs). But (a) all_msgs is already stored in memory, so the wastage is at most a factor of 2; and (b) this all happens in main() shortly before program exit erases all sins.

I didn't bother documenting this in the commit message because the patch changes the code anyway, but feel free to add the above explanation to the commit message if you think it is useful.

Show 9 quoted lines

>> For some reason, there is a bunch of infrastructure in this file for
>> dealing with IMAP flags, although there is nothing in the code that
>> actually allows any flags to be set.  If there is no plan to add
>> support for flags in the future, a bunch of code could be ripped out
>> and "struct msg_data" could be completely replaced with strbuf.
> 
> Yeah, after all these years we have kept the unused flags field
> there and nobody needed anything out of it.  I am OK with a removal
> if it is done at the very end of the series.

I don't think the removal of flags needs to be part of the same series.
 I suggest a separate patch series dedicated to deleting *all* the extra
imap infrastructure at once.  That being said, I'm not committing to do
so.  (We could add it to an "straightforward projects for aspiring git
developers" list, if we had such a thing.)

Michael

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