From: bsa@kf8nh.wariat.org Subject: Re: Porting MH-6.8 to Linux.... Date: Sun, 28 Feb 1993 02:08:17 GMT
In article <CGD.93Feb25171515@eden.CS.Berkeley.EDU> cgd@eden.CS.Berkeley.EDU (Chris G. Demetriou) writes:
>In article <1993Feb22.225620.16415@serval.net.wsu.edu> hlu@eecs.wsu.edu (H.J. Lu) writes:
>=>|> "badly coded" implies without justification. There's sufficiently
>=>|> good reasoning as to why that chunk of MH is coded that way -- read
>=>|> the comments, and you'll find out...
>=>
>=>I don't have the time.
>
>if you don't have time to read the comments, then how do you plan
>to modify the code? the m_getfld code is pretty hairy...
He wasn't. I was talking about doing it, as was someone else in this
newsgroup. And I *did* read the comments. I just need to read the GNU
iostreams code now... hopefully I'll get time Monday, so I can install a real
mailer :-)
>>Why not leave that to stdio? Not all stdio's are the same.
>
>umm, because stdio is meant to be a generalized package,
>good for some things, but not meant to be the most
>optimized solution?
The counter-argument is that if m_getfld() needs speed it has no business
using stdio. Would it have taken that much effort to add to MH an optimized
but incomplete (relative to stdio), reasonably portable, buffered file package
for the benefit of m_getfld()? Or even an unoptimized one that let m_getfld()
get at its internals so it could continue to optimize itself?
++Brandon