-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Thu, Sep 30, 2004 at 09:46:22AM +0200, Paul J Stevens wrote: > > I think you misunderstand. Simplification of the code is primary. Doing so > requires getting rid of large parts of the code. I'm currently removing all > list code (he leif) and replace it with glist where necessary. I'm also > removing all cache code, and I'm replacing all mimeparsing with gmime. > However, since I'm learning about gmime and glib as I go along, it's much > easier to develop all this new code in a stand alone situation where I > don't have to fireup the daemons sprinkled with trace and sleep calls, > connect a gdb to the sleeping process, set breakpoints, etc.... > Glib disgusts me. Although I beleive in outsourcing code to other programs, glib is horrible partly due to the fact that its another monolithic bloated library. What benefits are you getting by adding another dependency vs. using the existing mime parsing.
> The reason I'm removing the cache code is because it's an optimization hack > that's totally intertwined with the imap code, making that code unreadable > at best. Also, I'm convinced that cleaning up the code will allow us to > focus more clearly on the important bottlenecks. If a memory cache of some > sort is required it should be reimplemented *after* the cleanup, and fully > integrated in the new message retrieval api in a totally opaque manner. > Cool. > -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFBXcDOF6i3K/AxoQERAjusAKCZvhaL8dVZ/vPEKZcdbCgVeQ0vkgCfeCCU AAYTAtEHbNHvtWENIYmpf48= =LUyZ -----END PGP SIGNATURE-----
