Re: [gmime-devel] GMime 3.0.0 released

[Jeffrey Stedfast <jestedfa microsoft com>]

Hi Peter,

Thanks for this bug report! Looks like you’re right about the cause of the problem.

What the code *should* be doing is removing the “Content-Transfer-Encoding” header (if it exists) instead of 
trying to set the value to NULL.

I’ve just fixed the code in git master – I can try to make a new release this weekend if needed (or I can 
wait if you think you might have a few more of these type cases).

Jeff 

On 6/22/17, 12:26 PM, "gmime-devel-list on behalf of Peter Bloomfield" <gmime-devel-list-bounces gnome org on 
behalf of peterbloomfield bellsouth net> wrote:

    Hi Jeff:
    
    On 04/23/2017 12:17:22 PM Sun, Jeffrey Stedfast wrote:
    > After a long hiatus from GMime development, I am pleased to announce the release of GMime 3.0.0!
    …
    
    Many thanks for your work on GMime!
    
    I'm testing Balsa with GMime 3.0.1, and getting occasional gmime-CRITICAL warnings. These are very 
helpful in identifying mistakes in my port, but they're now down to the point where they seem to result from 
the GMime code itself. Here's a recent stack trace:
    
    #0  0x00007fffeef41a31 in _g_log_abort () at /lib64/libglib-2.0.so.0
    #1  0x00007fffeef42e3d in g_logv () at /lib64/libglib-2.0.so.0
    #2  0x00007fffeef42f9f in g_log () at /lib64/libglib-2.0.so.0
    #3  0x00007ffff71467a2 in g_mime_header_list_set (headers=0x17ccc80, name=name@entry=0x7ffff716c99b 
"Content-Transfer-Encoding", value=value@entry=0x0, charset=charset@entry=0x0)
         at gmime-header.c:1322
    #4  0x00007ffff715486a in g_mime_part_set_content_encoding (mime_part=0x9f9ae0, 
encoding=GMIME_CONTENT_ENCODING_DEFAULT) at gmime-part.c:791
    #5  0x00007ffff715490d in mime_part_encode (object=0x9f9ae0, constraint=GMIME_ENCODING_CONSTRAINT_7BIT) 
at gmime-part.c:470
    #6  0x00007ffff714a928 in multipart_encode (object=0x93c2c0, constraint=GMIME_ENCODING_CONSTRAINT_7BIT) 
at gmime-multipart.c:240
    #7  0x00000000004ae52c in lbm_mbox_armored_object (stream=0xb638a0) at mailbox_mbox.c:1916
    
    The call from lbm_mbox_armored_object is to g_mime_object_encode, but gdb didn't include a frame for it.
    
    The reason for the warning is that g_mime_content_encoding_to_string (GMIME_CONTENT_ENCODING_DEFAULT) 
returns a NULL value, but g_mime_part_set_content_encoding passes it to g_mime_header_list_set without 
checking it.
    
    Possibly mime_part_encode could skip calling g_mime_part_set_content_encoding when encoding == 
GMIME_CONTENT_ENCODING_DEFAULT, as that's the initial encoding of a GMimePart. Or perhaps 
g_mime_part_set_content_encoding could simply return if mime_part->encoding == encoding.
    
    Thanks again for GMime!
    
    Best,
    
    Peter