Breakage in mod_parrot

[Bart Wiegmans]

Hi Jonathan, everybody else,

I've talked with whiteknight++ about this issue. It
get_pointer_keyed_int is used to get different parts of the io
interface (io vtable, write buffer, and read buffer, see src/pmc/handle.pmc).

This poses a bit of a problem, because a pure-parrot object cannot
return pointers so easily, nor can it construct and manage the
appropiate (low-level) objects. It would be possible with some
low-level hackery, but not easily. Anyway, long story short, it is
kind of hard to make a pure winxed / parrot io handle.

Kind regards,
Bart

2012/11/8, Jonathan "Duke" Leto <jonathan at leto.net>:
> Howdy,
>
> Is there a way you can define a "dummy" get_pointer_indexed_key that does
> nothing (or always returns 1 or somesuch) to get around this?
>
> Duke
>
>
> On Thu, Nov 8, 2012 at 12:28 AM, Bart Wiegmans
> <bartwiegmans at gmail.com>wrote:
>
>> Hi everybody,
>>
>> So as a way of doing something fun in the morning I ran make test on
>> mod_parrot again. And sure enough, it found a bug. The apache output
>> handle
>> I wrote doesn't work anymore, because it doesn't have the
>> get_pointer_indexed_key VTABLE method.
>>
>> I have a pretty good guess where this comes from, namely, io_cleanup1 was
>> merged and seems to require this VTABLE method. But is nevertheless a bit
>> disturbing because the thing seems to ask for a pointer, or maybe hand me
>> a
>> pointer; I have no idea how to deal with it.
>>
>> The apache input / output handles are written in pure winxed and simply
>> call out to C using NCI. They do no buffering to speak of, and most
>> especially don't manage pointers. In short, I'd like to fix this, but I'm
>> at loss how.
>>
>> Kind regards,
>> Bart
>>
>> _______________________________________________
>> http://lists.parrot.org/mailman/listinfo/parrot-dev
>>
>>
>
>
> --
> Jonathan "Duke" Leto <jonathan at leto.net>
> Leto Labs LLC http://labs.leto.net
> 209.691.DUKE http://dukeleto.pl
>