Re: [Karrigell-main] Include() behaviour

[Brian O. <br...@qi...>]

On Wednesday, May 28, 2003, at 04:16 PM, Pierre Quentel wrote:

> As Brian noticed, the last version of Karrigell, with a separate=20
> Template module, has changed the behaviour of Include() : a variable=20=

> defined inside a script which has an Include(otherScript)=A0is no more=20=

> available in otherScript
> This was not intentional, and is probably due to the use of copy.copy=20=

> inside the "render" method in the "Script" class of Template : if I=20
> replace =A0=A0ns =3D copy.copy(globals())
> by ns =3D globals() it works like before
> Now which behaviour do users find the best ? Having the whole=20
> namespace seems more convenient, but are there unwanted side effects=20=

> and should we impose to explicitely pass parameters with=20
> Include(otherScript, **args) ?

Just like to say what I told Pierre through private discussion. I=20
personally like it without having to put arguments within the=20
Include(). I like this because it seems more intuitive, meaning that=20
when you include a document within another document, it is as if you=20
are just extending the first document, which calls the include. Second,=20=

it is easier, if you have a lot of includes running.

Now, the other way, having arguments being passed into the Include()=20
function, has benefits. It is explicit, meaning you have knowledge,=20
through scanning the source, how certain variables end up where they=20
are supposed to be.

I hang more to the first way, but each one has its merits. However, it=20=

seems a bit crazy, at least in my opinion, to have two functions that=20
do the two different things ... you got to pick one and go with it. :-)

Brian=