Re: [glade--] error of FC2 and glade--?



Murray Cumming schrieb:
On Fri, 2004-07-02 at 15:10 +0200, Christof Petig wrote:
It is wrong of Anjuta to suggest that glademm is necessary for gtkmm or
gnomemm development. libglademm, however, is very appropriate.

Since you keep telling people to prefer libglademm (your code) without looking at glademm (my code), I want to emphasize that both have their value, you simply have to choose. ;-)


I have to keep telling them because
a) Anjuta makes them think that this is what they should use, and what
everyone else is using.

Try to tell the Anjuta maintainers not me ;-) I believe they chose glademm since it gives a ready to compile project directory to start with. (see below) Sadly none of the anjuta maintainers ever contacted me to improve glademm's interaction with anjuta. I simply know they exist (somewhere). Neither did I receive any feedback for the ability to use glademm to set up a libglademm project.

b) glademm's name and website don't tell them that it's something
unusual, or that it's not part of the gnomemm project.

Correct, I'll change that. But most of the time I have to answer the _only_ question which is answered on the glademm website on the gtkmm mailing list (why can't I access my widgets?). So not many people find and read the site :-(

I don't think that "ability to change the user interface without
recompilation" is wanted by anyone.

I didn't see it either. But I'm very sure this was one of the most cited pros of libglade over glade's C generator (when libglade evolved - 3-4 years ago?).

libglademm is just simpler and it more closely wraps what is done in
GNOME with C - In C, people use libglade but only fools use glade C code
generation. It is not so bad with glade C++ code generation but it's
still an unusual technique that I don't believe is suitable for
beginners.

I still believe that using a tool to set up the project (configure.in, Makefile.am, main.cc, window.hh, window.cc) is a very good idea for beginners. Glademm offers this as a side effect but that might easily get the task of a future external program (which I happily would use for glademm. Configure.in generation has always been an error prone task)

No, I do not step forward to get the maintainer of that program. I might change glademm to do it once the caller interface is designed.

<further thinking>

It would have been a good idea to separate glademm into several pieces:
- generate UI classes from .glade files (foo_glade.??)
>This is glademm's core task<
- generate a user class skeleton (foo.??) [possibly add callback stubs to an existing]
- generate a project directory (configure.in, Makefile.am, po, intl etc.)
>I would happily give away the hassle to maintain such a beast<
- frontend to invoke the latter

Perhaps we (the gtkmm + gnomemm + anjuta people and me=glademm) can combine efforts to maintain the project skeleton generator. Any interest? I'd prepare a cross posting.
    Christof

Attachment: signature.asc
Description: OpenPGP digital signature



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]