reviewing design for namespaces

[Patrick R. Michaud]

On Mon, Mar 23, 2009 at 06:57:44PM -0700, Allison Randal wrote:
> One of the roadmap tasks for this month is a review of the design for  
> namespaces. I'd like to kick this off with a group braindump.
>
> What works? What doesn't? What would you like to be able to do, even if  
> you aren't sure how it should work?

Here's a brief review of namespace-related issues I've encountered (as 
requested from today's #parrotsketch).

(1) TT #389 describes a place where Parrot deviates from PDD19 --
in particular, subs that have a :method flag should not receive
symbol entries in the namespace unless the :nsentry flag is also
present.

(2) The same is true for subs with :vtable flags -- they should
not be entered in the namespace unless :nsentry is also present.

(3) This isn't strictly about namespace design, but TT #8 describes
some inconsistencies in PDD 15 relating to locating classes by namespaces,
keys, arrays, and the like.  The wording should be made consistent.

(4) A few places in PDD 21 talk about different types of namespaces
for different HLLs, but is a bit short on details (and I'm not sure
how the work).  For example, the mechanism for associating a new
namespace type with an HLL isn't described.  (I now realize that
this would be an excellent use for the .HLL_map directive....
but unfortunately we just got rid of that.)

As best I can remember, these are the only outstanding
namespace-related items that I have at present, and only
the first is "pressing" for the things I'm currently developing.
Most other namespace issues I once had now appear to have 
been resolved.

Pm