Re: Experimental IMAP branch

[Eric Gillespie]

David Levine <levinedl@acm.org> writes:

> In the case of gmail servers, will inc be able to use the gmail APIs?

No, but it does work with Gmail's IMAP.

> The gmailapis branch has (Python) code to do that in docs/contrib/gmailapis/.

You could write a gmailapi Python script implementing the folder
manager interface.

I'm only interested in widely implemented open standards like
IMAP, but there's nothing getting in the way of anyone who wants
to build something like that for one provider's APi flavor of the month :)

Paul Fox <pgf@foxharp.boston.ma.us> writes:

> Can you say a few words about how you handle imap id to mh message
> number mappings?  I'm thinking of refiling between folders, and also
> refiling (moving) within a folder.

I added sections to docs/imap/DESIGN discussing both of these.

> Will there be a 1-to-1 correspondence between imap and mh folders?

Yes.  I think the new inc section explains this.

Ken Hornstein <kenh@pobox.com> writes:

> I've often thought that the job of inc is to takes messages from a
> maildrop/file and put them in nmh mailboxes.  Since normally messages

That's the job inc is doing for IMAP.  I added a section
explaining this.  The documentation still has a long way to go :)

Michael Richardson <mcr@sandelman.ca> writes:

> is it in imap-prototype branch then?

Yes.  Sorry I omitted that detail.

> ./configure

And I guess I also omitted that you need to configure with
--enable-experimental-imap-folder-manager as this is all off
by default.

> make
> saw some .rs stuff go bye... error...

You shouldn't have seen any unless you ran 'make cargo'.
Eventually I'll hook it up to the all and install targets, but
neither is done yet.

> sbr/folder_manager_client.c:1291:1: error: conflicting types for 
> ‘mh_folder_manager_connect’
>  mh_folder_manager_connect(
>  ^~~~~~~~~~~~~~~~~~~~~~~~~
> In file included from sbr/folder_manager_client.c:21:
> ./h/managed_folders.h:52:1: note: previous declaration of 
> ‘mh_folder_manager_connect’ was here
>  mh_folder_manager_connect(mh_managed_folder *);
>  ^~~~~~~~~~~~~~~~~~~~~~~~~

Since you missed the flag to configure, this suggests my
non-experimental build has bitrotted.  I intend for a clean build
when the flag is off.  Not sure when I last tested that.  Sorry.

Thanks everyone!