Re: [xml] Adding a build/configure *option* to use ICU converters

[Daniel Veillard <veillard redhat com>]

On Thu, Sep 27, 2007 at 11:04:46AM -0700, Jungshik Shin wrote:
> Hi Daniel,
>
> Thanks a lot for going through my patch.
>
> 2007/9/26, Daniel Veillard <veillard redhat com>:
> > On Tue, Sep 25, 2007 at 04:48:03PM -0700, Jungshik Shin wrote:
>
> > > For those projects, it'd be very nice to have a configuration option
> > > in libxml2 that allow them
> > > to use ICU converters rather than libiconv.
> > >
> > > I made a preliminary patch and uploaded it to bugzilla (
> > > http://bugzilla.gnome.org/show_bug.cgi?id=480323 ).
> > > It works well on Windows, but I haven't yet managed to change
> > > configure.in so that it has an 'with-icu' option.
> > >
> > > I wonder what others think about this.  It'd be also nice if somebody
> > > could help me with patching configure.in and related files to add
> > > '-with-icu' option.
> >
> >   In a nutshell here is my position:
> >     - ICU size is like 5 times libxml2 size, we can't depend on it
> >     - Adding ICU support when the users need it is okay
> >     - This should be disabled by default, it's really a custom option
> >     - your patch looks nicely done
>
> I have  no problem with your position.  Actually, that's exactly my
> intention. I NEVER intended ICU option to be the default. On (desktop)
> Linux (or other POSIX system) where iconv is readily (and by default)
> available, there's no  reason to make ICU the default (or even to
> enable it).

  yeah, I had no doubt about that, it was clear from your patch.

> Do you want me to wait for Mark's patch (with configure.in and other
> related changes for Unix) and submit an updated patch so that you can
> land the whole thing in one shot? Or, do you like to land it as it is
> now and add Mark's change later?

  There is no need for me to put pressure on you. I should commit this
it looks like there is some kind of consensus about your patch. Just bear
with me I'm awfully busy with Real Work ATM :-\

> > I don't see why your patch shouldn't be applied and yes adding the option
> > on configure would be sensible (as long as it's off by default). Also I would
> > really prefer if Unixes didn't switched it on when distributed as part of
> > the OS to maintain a predictable behaviour on all similar platforms.
>
> I also agree with you on this.

  thanks again,

Daniel

-- 
Red Hat Virtualization group http://redhat.com/virtualization/
Daniel Veillard      | virtualization library  http://libvirt.org/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine  http://rpmfind.net/