[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: ncurses++ and maintenance thereof
From: |
Jürgen Pfeifer |
Subject: |
RE: ncurses++ and maintenance thereof |
Date: |
Fri, 25 Oct 2002 22:06:24 +0200 |
The most reasonable approach is that you create a new style C++
binding using STL etc. etc.
We can put this into our distribution in parallel to the existing
one and do the following:
- For some release cycles the new binding is declared Beta.
- The new binding becomes "released", the old one is deprecated
but still in the distribution
- At some release in the future the old binding is removed.
Cheers
Jürgen
> -----Original Message-----
> From: L. Dee Holtsclaw [mailto:address@hidden
> Sent: Friday, October 25, 2002 4:13 AM
> To: Jürgen Pfeifer
> Cc: Ncurses Mailing List; Thomas Dickey
> Subject: RE: ncurses++ and maintenance thereof
>
>
> I already sent a reply (off list) to Mr. Dickey regarding
> extending these classes to make them usable. As they are now,
> they are too limited to be truly useful in a real-world application.
>
> I don't mind taking these and running with 'em, including
> making them Doxygen compatible so there will be documentation
> available. I'd say, however, that the result will most likely
> be quite unlike they are now.
>
> If, as you say, you'd rather keep them as they are, I'll
> respect that and create a new set of wrappers from scratch. I
> cannot, however, accomplish anything significant by simply
> deriving extension classes since the base is so limiting.
> There are too many missing hooks and I really have a personal
> problem with the naming convention, or rather, the apparent
> lack of same -- makes the code *really* hard to follow.
>
> I have experience in using, maintaining, extending and
> porting a proprietary library [Vermont Views] so I was hoping
> to use that experience and finally contribute to open-source myself.
>
> Just let me know and I'll go from there.
>
> Thanks.
>
> Ciao,
> Dee
>
> On Thu, 2002-10-24 at 19:51, Jürgen Pfeifer wrote:
> > > So... This brings the following questions:
> > > 1. Is anyone maintaining these classes and, if so, how do I
> > > contact?
> > Not really. This was a weekend hack.
> >
> > > 2. Are these classes in enough use to prohibit renaming some
> > > members?
> > >
> > Based on the almost zero feedback I don't think so. But who knows.
> >
> > >3. Is there any reason NOT to use some STL in the implementations?
> > >
> > The time this stuff was written STL was far from being usable with
> > g++. That's different now.
> >
> > > FYI, Some of the functionality I intend to add:
> > > 1. Menu choices displaying details on a "status line" if so
> > > allocated 2. Current field details ditto 3. ESC key
> > > additionally available to exit-cancel a menu or form 4.
> > > Accessing Pad driver outside of Pad's own loop (for an
> > > external loop) 5. Swapping in new SLKs for active forms/fields
> > >
> > All this can easily be done with the classes as they are. I
> prefer not
> > to add all and everything into the classes but only what is
> available
> > in the C APIs.
> >
> > Jürgen
>
>
>