<<< Date Index >>>     <<< Thread Index >>>

Re: Rolling in sidebar, other mutt-ng type bits?



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thursday, October 23 at 02:23 PM, quoth Robin Lee Powell:
> I gather there's now another sidebar patch.

Oh? Are you sure you aren't getting confused by the transfer of 
maintainer from Thomer M. Gil to Terry P. Chan?

> Is there some reason it, and anything else useful from mutt-ng, 
> hasn't been rolled into mutt proper?

Much of what was mutt-ng *has* been rolled into mutt proper. The 
sidebar patch is the only significant thing I'm aware of that hasn't. 
The question about that particular sidebar patch is one that comes up 
every now and then, and has been answered several times. Rocco Rutte 
(one of the primary mutt bug squishers, and one of the three founders 
of mutt-ng) reviewed the patch in detail. He said the following 
(http://marc.info/?l=mutt-dev&m=112133798519807&w=2):

     For example, the sidebar patch available for mutt looks to work at
     first sight but there're many things just heavily broken or things
     you really don't want to stay in the code (like using snprintf()
     and strlen() to "calculate" the amount of digits of a number.).

     Really to integrate a patch by means of merging its functionality
     with the existing to get a better code base is just much more
     difficult than it sounds especially when you keep in mind that it
     also takes time to get an idea of what the source is supposed to
     do [and] how.

Essentially: the sidebar patch is large (1789 lines), complex, and 
appears to be poorly written. Those are all characteristics of code 
you want to stay as far away from as possible. My understanding is 
that that patch touches pieces of mutt's source that should have 
nothing to do with displaying a "sidebar", such as the mbox parsing 
code. Because of this, it produces unfortunate side-effects. You have 
only to look into the mutt archives a few weeks back to find people 
complaining that the patch causes mutt to hang in some circumstances.

And yet, the developer of the sidebar patch does not appear to provide 
support to those who use it (at least, not on the mutt users mailing 
list), and does not seem interested in cleaning it up, explaining it, 
or doing anything else that would be necessary or useful to getting it 
integrated with mainline mutt.

> The sidebar in particular sure seems like a really, really nice 
> feature.

Are you willing to reimplement it? Cleanly? Or explain the innards of 
the current sidebar patch to the primary mutt developers?

> Does current mutt have header caching?

Yes, and has for almost four years.

Mutt also has message caching (and has for almost two years), which 
mutt-ng does not have.

Mutt enjoys ongoing development. Mutt-ng provided a good catalyst for  
more development in mutt, but essentially atrophied as mutt 
development was renewed---mutt-ng hasn't been modified since April 
2006. Mutt-ng has, for all intents and purposes, been subsumed by the 
original mutt. The sidebar patch, as it stands, was rejected.

~Kyle
- -- 
The whole art of government consists in the art of being honest.
                   -- Thomas Jefferson: Rights of British America, 1774
-----BEGIN PGP SIGNATURE-----
Comment: Thank you for using encryption!

iEYEARECAAYFAkkA//AACgkQBkIOoMqOI16LvACfTMDYD9ZLJEuvKVmUf66Y+br+
+L4AnjzC4r/1eSqI4CMZXpERZAF4ZiTR
=WfCU
-----END PGP SIGNATURE-----