This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [Ada] flow.c sanity check breaks RTL build
> Jan Hubicka <jh@suse.cz> writes:
>
> > Hi,
> > calling update_life_info with nonzeroed aux fields creates all sord of weird
> > effects, so it is worthwhile to sanity check this. Fixes are following.
>
> Compiling urealp.adb triggers this sanity check. Here's the relevant
> part of a backtrace (with current CVS):
>
> #1 0x08272380 in calculate_global_regs_live (blocks_in=0x0, blocks_out=0x0,
> flags=0) at /home/fw/src-cvs/gnu/ORIG/gcc-MAIN/gcc/flow.c:1081
> #2 0x0827131b in update_life_info (blocks=0x0, extent=UPDATE_LIFE_GLOBAL,
> prop_flags=1) at /home/fw/src-cvs/gnu/ORIG/gcc-MAIN/gcc/flow.c:660
> #3 0x08270b5c in life_analysis (f=0x402fdd60, file=0x0, flags=1)
> at /home/fw/src-cvs/gnu/ORIG/gcc-MAIN/gcc/flow.c:478
> #4 0x084352e9 in reg_to_stack (first=0x1, file=0x0)
> at /home/fw/src-cvs/gnu/ORIG/gcc-MAIN/gcc/reg-stack.c:448
> #5 0x0838a958 in rest_of_compilation (decl=0x402a80e0)
> at /home/fw/src-cvs/gnu/ORIG/gcc-MAIN/gcc/toplev.c:3290
> #6 0x08064202 in end_subprog_body ()
> at /home/fw/src-cvs/gnu/ORIG/gcc-MAIN/gcc/ada/utils.c:1872
>
> What shall I do to track this down further?
I am just checking what makes my tree different from CVS, as C++ is unhappy
for regression tester, but hapy for me, so I guess my tree got out of sync.
CVS diff takes a while over modem line...
Honza