On 11/29/2012 10:33 PM, Junio C Hamano wrote:
> Michael Haggerty <mhagger@alum.mit.edu> writes:
> 
>> Now that we can xml-quote an arbitrary string in O(N), there is no
>> reason to process the message line by line.  This change saves lots of
>> memory allocations and copying.
>>
>> The old code would have created invalid output for a malformed input
>> message (one that does not contain a blank line separating the header
>> from the body).  The new code die()s in this situation.
> 
> Given that imap-send is about sending a patch the distinction would
> not matter in practice, but isn't the difference between the two
> that the new version would not allow sending a header-only message
> without a body, while the old one allowed it?

I was thinking that the end-of-header line is a required part of an
RFC2282 email message, but I was wrong.  If you squash the attached
patch onto this commit, it will handle emails without bodies correctly.

Nevertheless, the old code was even *more* broken because it added a
"</pre>" regardless of whether the separator line had been seen, and
therefore a message without an end-of-header line would come out like

    Header1: foo
    Header2: bar
    </pre>

with no content_type line, no pre_open, and </pre> appended to the
header without a blank line in between.  This is the "invalid output"
that I was referring to.

Michael

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