------- Comment #13 from dominiq at lps dot ens dot fr 2008-01-30 14:16 -------
I have applied the patch on top of rev. 131960 on i686-apple-darwin9. The only
change I have noticed is that the ICE for he code in comment #10 has been
replaced by the following (cryptic) error:
pr34828_1.f90:6.80:
teger(1), parameter :: ARR1(len(HEX1)/2) = (/( MSKa1(i), i=1,len(HEX1),2)/)
1
Error: Initialization expression didn't reduce (1)
...
(I can now see why the code is invalid!-).
For pr19925 codes, I still have the same ICEs as before the patch:
[karma] f90/bug% cat pr19925_1.f90
INTEGER, PARAMETER :: N=100000
INTEGER, PARAMETER :: I(N)=(/(MOD(K,2),K=1,N)/)
INTEGER, PARAMETER :: M(N)=I(N:1:-1)
END
[karma] f90/bug% gfc pr19925_1.f90
f951: internal compiler error: Bus error
...
[karma] f90/bug% cat pr19925_3.f90
program stuff
integer :: i_do
integer :: i(100001) = (/ (i_do, i_do=1,100001) /)
write (*,*) i(100001)
end program stuff
karma] f90/bug% gfc pr19925_3.f90
pr19925_3.f90: In function 'stuff':
pr19925_3.f90:1: internal compiler error: Possible frontend bug: array
constructor not expanded
...
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34828
