Hi, * Enrico Weigelt [06-05-18 00:56:21 +0200] wrote:
Sounds really good. Would be even better, if several things could be moved to separate libraries with clear dependencies (ie. linked with -no-undefied). This would clean up a lot.
Maybe someday we could move out things like the mailbox handling into an generic library, as I already suggested quite some time ago. (I personally don't have the time resources for this job currently, since other projects like Xserver occopy me a lot)
Cool, one more sharing these ideas.Brendan agrees and he came up with another very important issue: there should be 1.6 release some time.
Personally I think there're almost enough features added since 1.4 worth making up 1.6 (missing, IMHO and obviously: caching for POP and the conversion of <screen/> to <table/> for the docbook manual's function listings I made, UPGRADING and some documentation bits like the alias groups which are only "documented" in mail by TLR). Brendan plans to do more work to stableize the IMAP pipelining and I'm sure there're more very important things to do prior to it.
Moving the code around is not that difficult nor is it to start to implement the mailbox drivers as really independent modules (since that part of mutt is quite clearly structured and abstract already). Brendan once made a patch and so did I so this shouldn't be too difficult after all.
But: moving such code could cause trouble and needs lots of testing, esp. for the build process. And since it introduces no functional changes, I think it should be delayed past 1.6 to a 1.7 series since it's working right now.
I know that many people requested a mutt roadmap or sort of and noone has come up with one. Maybe we could use the wiki to write down what things should be done past 1.6 and prior to it so others like me can help by contributing patches to work towards 1.6. An idea would be to go with wiki.mutt.org or, if that doesn't work out, go with an actively maintained TODO document in CVS HEAD.
bye, Rocco -- :wq!