On 2/2/06, Patrick R. Michaud <[EMAIL PROTECTED]> wrote: > On Thu, Feb 02, 2006 at 11:33:25AM +0100, Leopold Toetsch wrote: > > On Feb 2, 2006, at 2:33, Allison Randal (via RT) wrote: > > > > >I've spent too much time on this error, so I'm routing around it, but > > >I'd love to figure out what's causing it. In my local version of > > >Punie I get this error when I run 'make test': > > > > While this all smells like a GC bug, I don't see it with r11401 on > > darwin nor x86/linux. I've also run io_print_6 through valgrind on the > > latter, which also doesn't show any indication re string truncation. > > > > Anyway, if it's a GC bug, the following should succeed: > > > > punie$ TEST_PROG_ARGS=-G make test > > This may or may not be related, but I'm getting a segfault for > building PGE itself (x86_64/linux), when trying to run mklib.pir > to generate the built-in rules. > i'm getting this, too, on win32. as are others, i think, on many platforms.
> This is one of the first times I've tried any of this under > 64-bit linux, so it might be totally unrelated. At any rate, > cd compilers/pge && ../../parrot -t mklib.pir >PGE/Library.pir near the end of it's output, reports: [snip] 3229 callmethodcc P0, "genfixedstr" - P0=Object(PGE::Exp::Literal)=PMC(0x417 020), 1973 get_params PMC_C[266] (7), P2, P0, S3, S2, S5, S4, S0 - , P2=PMCNULL, P0=PMCNULL, , , , , 1982 set_args PMC_C[24] (1), P2 - , P2=Object(PGE::Exp::Literal)=PMC(0x4 17020) 1985 get_results PMC_C[236] (5), I3, I0, I4, I1, S-1 - , I3=0, I0=2, I4=1, I1=0, S-1="(null)" 1992 callmethodcc P2, "quant" - P2=Object(PGE::Exp::Literal)=PMC(0x417 020), 1898 get_params PMC_C[24] (1), P0 - , P0=PMCNULL [snip] what the heck is register "S-1" in the get_results op?? when parrot got infinite registers, i don't think that meant they could be negative, too :) maybe that'll help track things down. > I've *finally* gotten many of my other tasks out of the way so > I'm back on the case. :-) > yay! welcome back. ~jerry
