Enhance [proc] for anonymous functions and closures

On Apr 2, 4:26=A0pm, EL <eckhardnos...@gmx.de> wrote:

>  Is it possible at all, in a
> certain time frame and what would this time frame be, approximately?

As far as I am aware (and I hope the tcl maintainers can correct this
impression if I am wrong), there isn't a team "out there" somewhere
that is working on the items in the feature request section of the
project. That area is, rather, a "sticky pad" for ideas. If someone in
the community sees a request and decides that it is just what they
needed, and they have time and ability, then they might, some day,
write up a TIP < http://wiki.tcl.tk/TIP > to detail the change, go
through the process of having it approved, write up the code, and
integrate the code, tests, and documentation into the core.

On the other hand, there might not be anyone who likes the idea enough
to actually write any code.

Think of the Feature requests as a "wouldn't it be nice" type of
thing. Think of the TIP as a "hey, I'm working (or about to work) on
some new code - here's what I am planning to do; is this something
that the core team will permit into the core?"

So, you've taken a fine first step - put the idea out there to
discuss. The next step would be to write a prototype implementation -
or perhaps brainstorm here or in some other forum to your liking - on
the way to implement and then implement a first take on the code. And
of course, if you feel your C skills are a bit lacking, then
recruiting help in implementing might be your first step.

On Apr 3, 1:13 pm, "Larry W. Virden" <lvir...@gmail.com> wrote:
>
> [...] go through the process of having it approved,
> write up the code, [...]
>

If I read Donal correctly, this is the converse: first implement, then
discuss. I must admit this is a bit surprising, because it means
wasted efforts if your patch ends up being rejected. But in reality,
there are  possibilities to get feedback before the final vote: TCT
members can't resist forever a question like "would you prefer A or
B" ;-)

-Alex

Alexandre Ferrieux wrote:
> On Apr 3, 1:13 pm, "Larry W. Virden" <lvir...@gmail.com> wrote: 
>
> If I read Donal correctly, this is the converse: first implement, then
> discuss. I must admit this is a bit surprising, because it means
> wasted efforts if your patch ends up being rejected. But in reality,
> there are  possibilities to get feedback before the final vote: TCT
> members can't resist forever a question like "would you prefer A or
> B" ;-)

Hmm, the refined way would be:
- Propose what you want to have
- Write a prototype patch so people can see that it in fact is
possible
- Let TCT vote
- Provide a full patch that fills in the missing pieces from the
prototype

Each of these stages allows some amount of discussion but unless you
can show at least a prototype its still discussing vapourware and gets
not as much attention.

Michael

On 3 Apr., 14:42, schlenk <schl...@uni-oldenburg.de> wrote:
> Hmm, the refined way would be:
> - Propose what you want to have
> - Write a prototype patch so people can see that it in fact is
> possible
> - Let TCT vote
> - Provide a full patch that fills in the missing pieces from the
> prototype

When I have a prototype that works, it is probably the most complete
implementation - because otherwise it wouldn't make sense to create a
proof of concept that fails to show the possibility.

So what Alexandre says is basically right, you put a lot of work and
effort into it and in the end they will likely reject it or put it on
hold forever, and you don't see it in the language anyway. For me this
happened once already (TIP#290). Therefore I am biased towards not
starting any work unless it can be half way confirmed that it is
successful and of interrest.
In the end I think that TIP's and implementations that are created by
TCT members have a much higher chance to make it into the language
than TIP's from non TCT members. I think this makes sense, but there
must be a way for others to file and discuss feature requests. The
sourceforge tracker and this group are good places for this.


Eckhard

On Apr 3, 3:22 pm, Eckhard Lehmann <eckhardnos...@gmx.de> wrote:
> In the end I think that TIP's and implementations that are created by
> TCT members have a much higher chance to make it into the language
> than TIP's from non TCT members. I think this makes sense, but there
> must be a way for others to file and discuss feature requests. The
> sourceforge tracker and this group are good places for this.

Yes, I too think it makes sense. Even if I invent something fabulous
(which I have yet to do ;-), I'll be happier to transfer my enthusisam
to a TCT member so that he can reimplement in agreement with 20 years
of craft, rather than to see my dumbo code accidentally pass the test
suite today and make it into the next Enceladus rover ;-)

Another good place to realize this symbiosis is the tcl-core list.
For example, a few ws ago, after a wild brainstorming session here,
I went to tcl-core proposing a Mutability patch. I had very
interesting discussions on this with several TCT members; then I wrote
roughly 30% of what I set out to do (List, not Dicts), even
distributed the early patch, but then a fundamental objection was
found by Miguel. This is a virtuous process, and I don't regret a
thing. Maybe you can do the same.

-Alex