> 19.03.2002 20:46:04, "Sung-Gu" <[EMAIL PROTECTED]> wrote:
> >> If anyone can test the 1.0 branch in it's current incarnation in CVS,
> >> especially the client-stuff (which I don't use myself at all), that
> >> would be great.
> >
> >Actually, I was going to integrate with commons-http 2.0 in commons
project.
> >However, I'd better to delay it a while.  Because it's gonna be many
changes
> >and need to have tests. ^^;
>
> I agree that moving back to the commons-httpclient would be nice at some
point,
> but I'd do it in the HEAD branch as opposed to the SLIDE_1_0 branch.

Yes, we can't use it before, as it breaks the API (as you pointed out).

> >About the client-stuff, as I glimpse it, there are some minor problems.
> >The command client with antlr doesn't work ok with a lexical command
token
> >parsing  before connection.
> >And antlr doesn't support the refid attribute of ant yet .  so someone
could be
> >some trouble in deploying it.
> >I think it's possible to find a work-around about that.
> >Whatever I'm saying above is not important to use the client library,
just
> >minors, I think.
>
> ANTLR is only used in the HEAD branch, not in SLIDE_1_0.

ANTLR looks like a nice cleanup, but the grammar we use needs a few tweaks
(for example, I could get the client to crash by using invalid tokens).

Remy


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to