blair christensen wrote:
(gdb) bt
#0 0x1a8638 in pobject_lives ()
#1 0x1bcc78 in Parrot_NCI_mark ()
Just one thought:
before my eval patch PerlHash's enum was 12, what now NCI has.
From PMC.pm (now):
NCI => 12,
PerlArray => 13,
PerlHash => 14,
PerlInt => 15,
blair christensen wrote:
and Simon Glover reported:
#0 0x1a8638 in pobject_lives ()
#1 0x1bcc78 in Parrot_NCI_mark ()
I dunno yet, why mark is called with an uninitialized struct_val. The
mark function should only be called, if the corresponding flag is set.
For a first check, what's going
Here's another data-point, for Linux/x86: with a fresh check-out, parrot
is passing all of its tests, but most of the imcc tests are failing.
Specifically, I get:
Failed Test Stat Wstat Total Fail Failed List of Failed
-
On Tue, Jan 21, 2003 at 01:53:26PM +0100, Leopold Toetsch wrote:
> blair christensen wrote:
>
> >On Mon, Jan 20, 2003 at 02:09:10PM -0500, Dan Sugalski wrote:
> >
> >>Rather badly, actually. All the tests that involve parrot currently
> >>segfault in the NCI mark routine. At least on OS X, I'm no
blair christensen wrote:
On Mon, Jan 20, 2003 at 02:09:10PM -0500, Dan Sugalski wrote:
Rather badly, actually. All the tests that involve parrot currently
segfault in the NCI mark routine. At least on OS X, I'm not sure
about other platforms.
(gdb) bt
#0 0x1a8638 in pobject_lives ()
#1 0x
On Mon, Jan 20, 2003 at 02:09:10PM -0500, Dan Sugalski wrote:
> Rather badly, actually. All the tests that involve parrot currently
> segfault in the NCI mark routine. At least on OS X, I'm not sure
> about other platforms.
5.8.0/Sparc64/OpenBSD is also rather unhappy.
I'm getting the following
It appears to be the same thing for tcc and lcc on linux. The nearest I
can tell, the value stuffed into struct_val in Parrot_NCI_set_string_keyed
is somehow not a valid PObj- dereferencing its "obj" member seems to blow
things up.
It stomps all over the stack somehow, and I don't have much
Dan Sugalski wrote:
Rather badly, actually. All the tests that involve parrot currently
segfault in the NCI mark routine. At least on OS X, I'm not sure about
other platforms.
Some questions WRT tinderboxen:
- above error seems currently not to be covered
- which machines are when reported/onl
Dan Sugalski wrote:
Rather badly, actually. All the tests that involve parrot currently
segfault in the NCI mark routine. At least on OS X, I'm not sure about
other platforms.
Not on i386/Linux, anyway ...
Could be a slightly wrong init sequence, which I did introduce and have
hopefully fi
Rather badly, actually. All the tests that involve parrot currently
segfault in the NCI mark routine. At least on OS X, I'm not sure
about other platforms.
--
Dan
--"it's like this"---
Dan Sugalski
10 matches
Mail list logo
