Re: [Pan-devel] message/partial support (Update)

[Charles Kerr]

This is great news Jeff, as soon as this gets plugged in we'll
not only be able to support message/partial, but we'll also be
able to get rid of the one-big-temp-file required in decoding
multiparts now.

Did you want to plug glib2 into pan's cvs head, or do you want
me to do it?

cheers,
Charles


On Sun, May 12, 2002 at 03:34:50AM -0400, Jeffrey Stedfast wrote:
> n/m, figured it out - there were a few issues actually :-)
> 
> 1. I was forgetting to check bounds in GMimeStreamCat::read (and also
> ::write)
> 
> 2. I was forgetting that not all source streams have to have a
> bound_start of 0 (duh!) ;-)
> 
> Jeff
> 
> On Sun, 2002-05-12 at 00:29, Jeffrey Stedfast wrote:
> > I've basically "finished" this, although I may have to change the
> > gmime-message-partial.c API a bit. Instead of:
> > 
> > GMimeMessage *
> > g_mime_message_partial_reconstruct_message (GMimeMessagePartial **,
> > size_t num);
> > 
> > I may have to pass in GMimeMessages, since rfc1521 says that I must
> > merge the raw headers in the message/partial part with the headers of
> > the original message (discluding Content-* headers and maybe a few
> > others, I forget).
> > 
> > 
> > now for the Bad News (tm):
> > 
> > GMimeStreamCat seems to be buggy and I'm having trouble finding the
> > problem. I think it has to do with substreaming a Cat stream. When I
> > write out the reconstructed message, it seems to write the wrong section
> > of the stream for at least one of the parts?
> > If I write the Cat stream to a mem stream before parsing it into a
> > GMimeMessage, then writing out the resultant message is fine.
> > 
> > 
> > I'll work on this some more tomorrow...
> > 
> > Jeff
> > 
> > On Tue, 2002-05-07 at 14:23, Jeffrey Stedfast wrote:
> > > I've started hacking this last night and am nearly finished but I'm not
> > > sure what the best interface would be for splitting a message into
> > > multiple messages containing message/partial parts.
> > > 
> > > The way I see it, I could either return an array of GMimeMessagePartial
> > > parts and leave it up to our caller to construct the GMimeMessages and
> > > embed the GMimeMessagePartial parts themselves, or I could return an
> > > array of GMimeMessages.
> > > 
> > > I think I'm leaning toward the second one, but I'm not sure...
> > > 
> > > In case you're interested in how it will be implemented, the interface
> > > will look something like this:
> > > 
> > > GMimeMessage **g_mime_message_partial_split_message (GMimeMessage
> > > *message, size_t max_size, int *nparts);
> > > 
> > > 'max_size' will be a way to tell GMime that you don't want any mime part
> > > to contain any more than max_size bytes of the original message.
> > > 'nparts' will be set to the number of parts in the array (which you will
> > > have to g_free when you're done with it - the array, not nparts).
> > > 
> > > 
> > > On a related note, has Pan 0.12.x been ported to use GMime2 yet? If not,
> > > I might look into working on that in my spare time this week/weekend.
> > > 
> > > Jeff
> > > 
> > > 
> > > _______________________________________________
> > > Pan-devel mailing list
> > > address@hidden
> > > http://mail.freesoftware.fsf.org/mailman/listinfo/pan-devel
> > 
> > 
> > _______________________________________________
> > Pan-devel mailing list
> > address@hidden
> > http://mail.freesoftware.fsf.org/mailman/listinfo/pan-devel
> 
> 
> _______________________________________________
> Pan-devel mailing list
> address@hidden
> http://mail.freesoftware.fsf.org/mailman/listinfo/pan-devel