On Mon, 24 Jan 2011 13:49 +0100, "Moritz Lenz" <[email protected]>
wrote:
> But notice that even if NQP stays stable, your programs can still break 
> if you call API methods that change.

Well, is that API going to be stable any time soon, or will it continue
to evolve? I'm guessing the latter... I don't have anything against
improving the API when changes are necessary, but since I am aiming to
do a hobby project over the span of maybe a year with about half a
man-year to allocate to it, I do have to be careful in what I sign up
for. 

> Note that significant parts of Rakudo's PIR predate NQP (or at last the 
> NQP-rx variant), and just haven't been ported to a more high level 
> language since then. The presence of PIR code doesn't always imply it's 
> necessary to have that piece of code in PIR.

Interesting. One wouldn't know that by just looking at the current
documentation, though! Or even current HLL implementations. Should I
take the NQP source to be the definitive (de facto) documentation of
what is supported, then?

(Which means we're back at Perl 5's mentality of "if the source code and
documentation disagree, the source code is right". But I digress.)

Regards,
Ville Koskinen
_______________________________________________
http://lists.parrot.org/mailman/listinfo/parrot-dev

Reply via email to