iteratee ByteStringLike conversions

Maciej Marcin Piechotka uzytkownik2 at gmail.com
Mon Nov 14 01:32:51 GMT 2011


On Mon, 2011-11-14 at 09:06 +0800, Conrad Parker wrote:
> On 14 November 2011 08:57, Maciej Marcin Piechotka
> <uzytkownik2 at gmail.com> wrote:
> > On Mon, 2011-11-14 at 08:53 +0800, Conrad Parker wrote:
> >> On 12 November 2011 11:15, Maciej Marcin Piechotka
> >> <uzytkownik2 at gmail.com> wrote:
> >> > On Fri, 2011-11-11 at 12:48 +0800, Michael Baikov wrote:
> >> >> Also enumInflate in iteratee-compress *should* be implemented in terms
> >> >> of Enumeratee instread of Enumerator, because it's idea is to
> >> >> transform data stream from one type to another and not the iteratee
> >> >> itself.
> >> >
> >> > I fixed it in current master (along with disabling the debug flag by
> >> > default). I haven't had time to test it.
> >>
> >> Hi Maciej,
> >>
> >> I've just tested the Enumeratee version and it works fine, thanks!
> >> Also the removal of the debug flag has fixed the issues I had with
> >> verbose messages and random segfaults :) It'd be great if you could
> >> upload a new release of iteratee-compress so I can bump version
> >> dependencies in my packages.
> >>

I've just pushed 0.3.0.0.

> >> cheers,
> >>
> >> Conrad.
> >
> > Hello,
> >
> > Just to clarify - have you removed the debug flag or just test the
> > correct version with flag disabled by default?
> 
> I've just tested with your master branch with default flags.
> 
> > I'm a bit worried about segfault but probably I'll remove the debug code
> > if it works without it just fine.
> 
> sure, it could be worth investigating, perhaps it means that some part
> of the bindings can access an invalid pointer. However it should be
> safe as long as that isn't exposed to the user -- perhaps the C
> library also has the same behaviour, with the assumption that you just
> don't access that pointer at particular times.
> 

I'm not dereferening inner pointer - I'm just reading various fields
from pointer, for which I use ForeignPointer.

> cheers,
> 
> Conrad.

Regards




More information about the Iteratee mailing list