On Sun, 2014-01-26 at 17:38 -0500, Brian Davis wrote: > Brian Drummond wrote: > > > just to be clear : assuming ghdl -a and ghdl -e were run but not > logged? > > Yes, they were run but then assimilated into the preceding SNIP. > > > Is it possible we're not actually getting to static > EXCEPTION_DISPOSITION ghdl_SEH_handler(...) at all? > > That was my first guess, but it'll probably be a few days before I have > time to dig into this further. > > > Perhaps a printf (and an fflush(stdout) !) at its start could answer > that... > > There was already a fprintf, after the switch, that didn't show up on > STDERR. That could have been too late (and would have been, without the patch). And without fflush, I have seen printfs queued for later printing, before an abort with no output.
But your divz experiment appears to answer the main question : the exception we need seems to be trapped/handled before we .... AAAH! do you have anti-virus running? If you (go offline and) turn it off, do we get the printf? Tristan : background question : what raises the overflow? Is it an explicit raise in the JIT mcode? (if so, is there scope to raise an alternative exception?) - Brian _______________________________________________ Ghdl-discuss mailing list [email protected] https://mail.gna.org/listinfo/ghdl-discuss
