[m-dev.] "--debug" option and Opium-M

[Erwan Jahier]

Fergus, Tyson or Zoltan probably could give me an hint to where i should look
now.

If a Mercury program is compiled with "--grade asm_fast.gc.tr.debug " (vs
"--trace deep"), executing it within Opium-M provokes a SIGSEGV (and not 
within mdb). 

Looking at the stack (with gdb), it appears that it is due to the following
loop: "MR_trace -> MR_trace_real -> MR_trace_event -> MR_trace_event_external ->
MR_trace -> ..."

As a matter of fact, problems seems to occur each time a call to a function
that is imported from Mercury occurs. Here is the trace produced by mdb (I 
have put a breakpoint on MR_trace_event_external):

(gdb) step
Breakpoint 1, MR_trace_event_external (cmd=0x8347dc8, event_info=0xbfff12e4)
    at mercury_trace_external.c:382
382             Code            *jumpaddr = NULL;
(gdb) 387             Unsigned        seqno = event_info->MR_call_seqno;
(gdb) 388             Unsigned        depth = event_info->MR_call_depth;
(gdb) 389             MR_Trace_Port   port = event_info->MR_trace_port;
(gdb) 390             const char      *path = event_info->MR_event_path;
(gdb) 391             Word            *saved_regs = event_info->MR_saved_regs;
(gdb) 400             ML_DI_init_mercury_string(&MR_mmc_options);
(gdb) Breakpoint 1, MR_trace_event_external (cmd=0x8347dc8, event_info=0xbffed9b4)
    at mercury_trace_external.c:382
382             Code            *jumpaddr = NULL;


i.e., instead of continuing the execution of MR_trace_event_external(), it 
loops by calling again  MR_trace_event_external().



-- 
R1.


--------------------------------------------------------------------------
mercury-developers mailing list
Post messages to:       mercury-developers at cs.mu.oz.au
Administrative Queries: owner-mercury-developers at cs.mu.oz.au
Subscriptions:          mercury-developers-request at cs.mu.oz.au
--------------------------------------------------------------------------