--- Comment #21 from paul dot thomas at jet dot uk 2006-09-23 16:18 ---
Created an attachment (id=12313)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=12313action=view)
Latest patch
This version fixes the problem with constructors of structures that have
allocatable components
--- Comment #16 from paul dot thomas at jet dot uk 2006-09-20 09:22 ---
(In reply to comment #15)
looks like there is agreement that the problem is fixed.
I am afraid to say that I do not agree entirely; the above takes 6 seconds to
compile with ifort on my Athlon 1700 a vapeur
--- Comment #2 from paul dot thomas at jet dot uk 2006-09-07 16:16 ---
Oskar,
This works for me with gcc-4.2.0 20060905; I know of nothing that would have
effected this change so I am at a bit of a loss. I hung a main program on it
and checked the output too.
Paul
--
http
--- Comment #3 from paul dot thomas at jet dot uk 2006-05-05 06:26 ---
The patch below fixes the problem by enclosing the expression for the array
reference outputs%signal_number in parentheses(You can verify that this does
the right thing by doing it explicitly in the fortran
--- Comment #2 from paul dot thomas at jet dot uk 2006-05-04 09:50 ---
(In reply to comment #1)
Confirmed, this is a front-end issue.
we have:
struct calc_signal_type D.904;
D.904 = (*(struct calc_signal_type[0:] *)
outputs-data)[outputs-dim[0].stride
--- Comment #1 from paul dot thomas at jet dot uk 2006-04-27 10:44 ---
This one is marvelous - it's one of Sherlock Holme's three pipers.
If the initialization is removed, all works correctly. If the module is
compiled separately and the object file linked with the main program
--- Comment #2 from paul dot thomas at jet dot uk 2006-04-27 13:24 ---
Holmes was wrong - it was a one piper.
The symbol for the module equivalence was being scrubbed for the lack of a
gfc_commit_symbols. Interchanging lines 1061 and 1064 in trans-common.c does
the job. Thus
--- Comment #2 from paul dot thomas at jet dot uk 2006-04-26 14:09 ---
The testcase reduces to:
module a
integer, parameter :: dp = kind (1d0)
real(dp) :: reM, M
equivalence (M, reM)
end module a
module b
use a, only : dp
end module b
use a
use b
print *, M
reM
error
Product: gcc
Version: unknown
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: fortran
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: paul dot thomas at jet dot uk
http
--- Comment #5 from paul dot thomas at jet dot uk 2006-04-24 09:10 ---
(In reply to comment #4)
This looks to be fixable at trans-intrinsic.c(gfc_conv_intrinsic_len), here a
special switch branch for constructors is needed.
I will make it so.
Paul
This is indeed the case; the ICE
10 matches
Mail list logo
