[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
- To: tech_(_at_)_openbsd_(_dot_)_org
- Subject: yp optimization
- From: Theo de Raadt <deraadt_(_at_)_theos_(_dot_)_com>
- Date: Tue, 23 Apr 1996 20:46:36 -0600
text data bss dec hex total size
217088 12288 28376 257752 3eed8 csh.orig 229376
217088 12288 28648 258024 3efe8 csh.after_xdr 229376
212992 12288 28904 254184 3e0e8 csh.after_yp_and_xdr 225280 (-4096)
How's that? By changing some YP stuff around I've saved a page in the
csh binary... an i386, binary, btw.
I expect greater or equal wins in most other binaries. Anything that
does a database lookup. (Quick, name a program that doesn't!)a
While I was fixing a bunch of other bugs (grr, horrible ones) in the
YP & XDR stuff, I decided to investigate splitting the YP & RPC
functions into a bunch of seperate files. Hence, programs no longer
needs to link in a bunch of unneccessary YP functions like yp_order(),
yp_all() ... *and* their associated XDR functions. I think in some
simpler cases programs also avoid pulling in a few RPC functions.
I'm still trying to figure out why the bss grew. This is a static
binary, so I'm not sure. Any ideas?
There might be a shared library compatibility issue, because 4 _'d
variables can no longer be static. But that's largely irrelevant
because there is another bug fix in there which kills shared library