method stripper progress

Thu Feb 19 18:48:00 GMT 2004

Okay, things are progressing very well; I'm down to roughly
 15% .data
 15% .eh_frame
 70% .text
... and all other sections are insignificantly small.
BTW, the huge .data segment in my last email was my fault, and totally
unrelated to gcj. Sorry about that ;)
I'm pretty happy with where things are right now, but I just wanted to
see if anybody has any ideas on shrinking the .text segment. Although
I don't plan to actually ship with this set of flags, I'm using these
to make sure I can find the "lower bound" on size:
 -Os
 -fnew-ra \
 -ffunction-sections \
 -fdata-sections \
 -fomit-frame-pointer \
 -fno-force-mem \
 -fno-force-addr \
 -fbranch-probabilities \
 -fno-schedule-insns \
 -fno-optimize-sibling-calls \
 -fno-if-conversion \
 -fno-thread-jumps \
 -foptimize-static-class-initialization \
 -fno-store-check \
 -fno-bounds-check \
 -fmerge-all-constants \
 -fno-inline-functions \
 -finline-limit=1 \
-- 
In 1845 the Potato Famine decimated Irish agriculture because of lack
of genetic diversity. Over 150 years later, we still haven't learned
this lesson, which is why the plague of Microsoft viruses will
continue for as long as the software monoculture does.