flow/edge changes broke this on x86?

[craig@jcb-sc.com]

Here's a simple Fortran test case (which I'll later add to the test suite)
that just started failing with the most recent snapshot (19960228):

      SUBROUTINE QUICK
      SAVE
C
      ASSIGN 101 TO JUMP
  101 Continue
C
      RETURN
      END

I stripped out the commentary from the file actually compiled in the
shell output below:

/develop/egcs/19990228/bin/g77 -v -W -Wall -O 960317-1.f
g77 version egcs-2.93.10 19990228 (gcc2 ss-980929 experimental) (from FSF-g77 version 0.5.24-19990219)
Driving: /develop/egcs/19990228/bin/g77 -v -W -Wall -O 960317-1.f -lg2c -lm
Reading specs from /develop/egcs/19990228/lib/gcc-lib/i686-pc-linux-gnu/egcs-2.93.10/specs
gcc version egcs-2.93.10 19990228 (gcc2 ss-980929 experimental)
 /develop/egcs/19990228/lib/gcc-lib/i686-pc-linux-gnu/egcs-2.93.10/f771 960317-1.f -quiet -dumpbase 960317-1.f -O -W -Wall -version -fversion -o /tmp/ccQiiaaa.s
GNU F77 version egcs-2.93.10 19990228 (gcc2 ss-980929 experimental) (i686-pc-linux-gnu) compiled by GNU C version egcs-2.93.10 19990228 (gcc2 ss-980929 experimental).
GNU Fortran Front End version 0.5.24-19990219
g77: Internal compiler error: program f771 got fatal signal 11
[craig@deer g77-test]$

        tq vm, (burley)