From: Joe English on
ZB wrote:
> Alexandre Ferrieux wrote:
>
>> Bottom line: new syntax done at script level is very easy unless you
>> have very strict performance requirements; it's extremely flexible,
>> and needs zero consensus.
>
> Of course, the strongest advantage is, that it can be used instantly... :)


Wait, what?

Writing a small pure-tcl command that implements your pet syntax
and using it in your programs is going to be a *lot* less work --
and will certainly take a lot less time -- than getting
your pet syntax into the core.



--JE
From: ZB on
Dnia 31.05.2010 Joe English <jenglish(a)fdip.bad-monkeys.com> napisa�/a:

>>> Bottom line: new syntax done at script level is very easy unless you
>>> have very strict performance requirements; it's extremely flexible,
>>> and needs zero consensus.
>>
>> Of course, the strongest advantage is, that it can be used instantly... :)
>
>
> Wait, what?
>
> Writing a small pure-tcl command that implements your pet syntax
> and using it in your programs is going to be a *lot* less work --
> and will certainly take a lot less time -- than getting
> your pet syntax into the core.

....exactly, as Alexandre wrote: "new syntax done at script level is very
easy...". You can see it quoted above, with my confirmation below. So I can't
understand, what are you referring to(?).
--
Zbigniew
From: Donal K. Fellows on
On 31/05/2010 11:56, ZB wrote:
> Clarify, please. How actually can I meet such problem, while using the
> "arrow-shaped quasi-operators", in the procedure created by Alexandre?
> Any example?

I *definitely* write code like this:

if {[regexp {abc(d*e)f(g+)h} $str -> deVar gVar]} { ... }

I can't remember if I use the other direction of arrow. I'm not sure how
to work around it, or even if it needs working around at all. (More
serious is the fact that I'd like to leave the possibilities open for
doing funky syntax stuff in the future, such as better syntax for
dictionaries; “->” is less entangled than “.” for that, c.f., Tk.)

Donal.