Melvin Smith <[EMAIL PROTECTED]> wrote: > You'll have to edit interpreter.h and set PARROT_CATCH_NULL to 1 > to enable it.
Turned on now by default.
Good.
> The patch adds the Null PMC class, only instantiated once in > system memory.
... which was the main problem with the patch. PMCs have to point to pool memory to make DOD happy. It could have been ok, if the live bit was set - anyway, I use a constant PMC now, which isn't freed.
I figured, but I'm so rusty I can't remember anything.
> I didn't turn it on by default yet as we need to check over the code > to see where assumptions are being made about PMC pointers > just being NULL...
I've fixed all locations, where NULL PMCs are either used or tested, except these I forgot :)
Plain nci code needs still fixing, JIT/i386 nci is done.
I saw these, thanks. I figure there are still some spots hiding that I forgot as well.
Did you really run tests with --gc-debug?
Nope. Just did the default tests.
> We should also integrate the null.pmc generation into pmc2c or pmc2c2.
We should really switch to the latter and only put in there.
I wasn't even sure which one was used. We should remove one of them. I'm also trying to sprinkle some comments here or there as I make my way back through the code but it is slow going.
Thanks for fixing up the patch,
-Melvin
