libraries vs. file system performance

[The Beach Bum]

In article <15106 at mimsy.UUCP> chris at mimsy.UUCP (Chris Torek) writes:
>In article <10211 at rpp386.Dallas.TX.US> jfh at rpp386.Dallas.TX.US
>(The Beach Bum) writes:
>>[directory libraries might] display quadratic file system behavior
>>[ or worse ;-) ] ...
>
>(And *that* is the real problem, not the side issues several others
>have named.  Nonetheless, `directory libraries' would probably be handy
>during library development, and damn the scan time.  When it gets too
>bad, you give in and make the .a file.  Or the .a groups....)

Excellent suggestion.  Ranlib'ing a library every few minutes as you
debug modules is a serious pain - and of course the behavior is O(n**2),
for those of us WITHOUT [ sigh ;-( ] name caches, or zillions of block
buffers [ sigh ;-( ], directory flogging is a real loser.  namei's per
second is not a high number [ Version 7 file system being a total piece
of garbage compared to Berzerkeley FFS. ]  [ That's right - it's 2pm,
where is YOUR i-list? ]

And I do think I made some argument that below a certain size directory
libraries made sense.  So, Chris, when are you going to write the code?
Better still - when can we expect to see FFS on a '386 box?
-- 
John F. Haugh II                        +-Quote of the Week:-------------------
VoiceNet: (214) 250-3311   Data: -6272  |"Unix doesn't have bugs,
InterNet: jfh at rpp386.Dallas.TX.US       | Unix is a bug"
UucpNet : <backbone>!killer!rpp386!jfh  +--              -- author forgotten --