egcs-19981109 hpux10.20 build failure in libf2c

[Jeffrey A Law]

  > The following failure occurs when attempting to compile
  > libf2c/libU77/lbitbits.c while building with make or
  > make-bootstrap.  Reducing the optimization to -Os does not
  > affect this, but it compiles successfully at -O.
  > Dr. Timothy C. Prince
  > Consulting Engineer
  > Solar Turbines, a Caterpillar Company
  > alternate e-mail: tprince@computer.org
  > make[2]: Entering directory `/usr2/users/tim/src/gnu/egcs-19981109/hp/hppa1
  > .1-hp-hpux10.20/libf2c/libF77'
  > /usr2/users/tim/src/gnu/egcs-19981109/hp/gcc/xgcc -B/usr2/users/tim/src/gnu
  > /egcs-19981109/hp/gcc/ -B/usr2/users/tim/src/gnu/914/hppa1.1-
  > hp-
  > hpux10.20/bin/ -c -DSkip_f2c_Undefs -I. -I../../../../libf2c/libF77 -I.. -I
  > ../../../../libf2c/libF77/..  -DSTDC_HEADERS=1 -DRETSIGTYPE=v
  > oid
  >  -DIEEE_drem=1 -DSkip_f2c_Undefs=1  -O2 -pipe ../../../../libf2c/libF77/lbi
  > tbits.c
  > ../../../../libf2c/libF77/lbitbits.c: In function `lbit_cshift':
  > ../../../../libf2c/libF77/lbitbits.c:62: internal error--insn does not sati
  > sfy its constraints:
  > (jump_insn 237 92 97 (parallel[
  >             (set (pc)
                (label_ref 235))
  >             (set (reg:SI 88 SAR)
  >                 (reg/v:SI 25 %r25))
  >         ] ) 258 {decrement_and_branch_until_zero+5} (nil)
  >     (expr_list:REG_BR_PRED (const_int 14)
  >         (nil)))
  > ../../gcc/toplev.c:1390: Internal compiler error in function fatal_insn
  > {standard input}: Assembler messages:
  > {standard input}:70: Fatal error: Missing .exit
Odd.  hppa1.1-hp-hpux10.20 bootstrapped just fine for me.

Why is it that you're seeing so many failures and I see so few of them?

jeff