Re: [libxml++] Benefit of the dependency on glibmm

[Alexander Vassilev <avasilev voipgate com>]

Thank you for the answer,

I actually tried to do it, and succeeded, just mechanically replacing 
the class name, and the bytes() method, everything works like a charm. I 
guess by "ustring is just used to make the API nicer" you mean it's 
better integrated with gtkmm, and that integration is the reason to have 
ustring istead of std::string. I guess I am looking from the opposite 
angle - to have it standalone, instead of integrated. In case the 
standalone-ness is something that would be beneficial, I think there can 
be a way to do it, without doing a fork - just a mechanism that allows 
the string implementation to be switchable at compile time, something 
involving a stub class that wraps the string api and has switchable 
underlying implementation, or the dreaded DEFINE-s. I would not even 
think about that, but it turns out libxml++ to be one of the very few 
feature rich C++ xml libraries in (the only other comparable I know is 
Xerces). If the community is interested, I could take some time and come 
with a proposal about such a mechanism, which will remove the dependency 
on glibmm, when needed.

Best regards
Alex
> On Wed, 2010-02-10 at 01:23 +0200, Alexander Vassilev wrote:
>   
> > Hi all,
> >
> > I am new to libxml and libxml++ so I am not much familiar with the code
> > of libxml++ (i've just had a quick and superfluous look). Reading the
> > dependency requirements i stumbled upon the line "glibmm-2.4
> > <http://www.gtkmm.org/> (or a subset of it containing at least
> > Glib::ustring)". As I am developing for the windows platform (along with
> > linux and macos), I am a bit picky about dependencies (not easy on
> > windows). What worried me here is the dependency on glibmm, which in
> > turn depends on glib itself, which depends on several other gnu libs.
> > My question is - how heavily does libxml++ depend on glibmm and on
> > GLib::ustring? Is it theoretically possible to use std::string instead
> > and possibly directly libiconv for unicode conversions? Is this a
> > feasible idea, or I have missed something in my quick look?
> >     
>
> It's entirely doable in a (renamed) fork. But that would have to be
> maintained by somebody who wants it enough. So far nobody has made the
> effort:
> https://bugzilla.gnome.org/show_bug.cgi?id=320197
>
> ustring is just used to make the API nicer. No encoding conversion is
> actually done in libxml++. That's taken care of by libxml.
>
> By the way, this has been discussed repeatedly before.
>
> --
> murrayc murrayc com
> www.murrayc.com
> www.openismus.com
>
>
> --------------------------------------------------------
>
> This e-mail and any attached files are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this e-mail by mistake, please notify the sender immediately and delete it from your system. You must not copy the message or disclose its contents to anyone.
>
> Think of the environment: don't print this e-mail unless you really need to.
>
> --------------------------------------------------------
>
>   

begin:vcard
fn:Alexander Vassilev
n:Vassilev;Alexander
org:VoipGATE S.A.
title:Senior Software Engineer
version:2.1
end:vcard