Re: Version 3.37... (prg2lout crash)

Mark Summerfield <[email protected]> writes:

> I've also tried running in valgrind as per Ludovic's suggestion. The
> crash occurs exactly as before with valgrind reporting 0 errors:
>
>
> $ lout -r beginning run 5:
> lout file "gra_tick" (from "gra" line 38, from "all" line 46):
>    234,1: 23.7c object too high for 23.7c space; @Scale inserted
> lout file "gra_summ" (from "gra" line 44, from "all" line 46):
>     10,1: 25.7c object too high for 23.6c space; @Scale inserted
> lout file "prg_embe" (from "prg" line 136, from "all" line 48):
>    38,17: fatal error: failure (non-zero status) of filter: 
> prg2lout -r -lC                       -ilouti1 -olout1 -elout.err                     -t8 -T8s
> ==6261==
> ==6261== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 13 from 1)
> ==6261== malloc/free: in use at exit: 403,578,783 bytes in 99,368 blocks.
> ==6261== malloc/free: 353,380 allocs, 254,012 frees, 1,410,357,058 bytes 
> allocated.
> ==6261== For counts of detected errors, rerun with: -v
> ==6261== searching for pointers to 99,368 not-freed blocks.

What is it that's running under Valgrind?  Only `prg2lout'?  If so, did
you modify Lout's `prg' to run "valgrind prg2lout" instead of
"prg2lout"?

For easier debugging, I'd recommend running only `prg2lout' (with the
appropriate input file), as in:

  $ valgrind prg2lout -r -lC -ilouti1 -olout1 -elout.err -t8 -T8s

> I killed it at this point.

Which process did you kill?  Why did you have to kill it?

Thanks,
Ludo'.