Re: gEDA-dev: Gschem and Cairo graphics library

[Stuart Brorson <sdb-at-NOSPAMcloud9_dot_net>]

>> I also apologize in advance for pouring cold water on your project. I
>> took a look at the screenshots and the Cairo stuff does look slightly
>> whizzier than the GTK stuff.  I also hate to discourage a developer
>> who just took on a project and did it.  That's the spirit we like
>> around here!
>
> First let me say that this wasn't really a project at all. The patch has
> a little over 200 lines and it took me around 3 hours to make it. As I
> said, this just a hack, a proof of concept - I never said things like
> this should get to the official CVS repository any time soon. The main
> reason I did this was to see how Cairo would perform and to get to know
> Gschem code a bit better.
>
> I sent a mail to the mailing list because I thought my findings would be
> interesting to gEDA developers.

Thanks for not taking offence!  And thanks for your reply!  Your level
headedness is appreciated!

>> However, you've just introduced another external dependency into the
>>  build, which isn't necessarily welcome, IMO.  Indeed, one of our
>> bigger problems with gEDA/gaf is that users frequently can't build
>> the package due to missing dependencies.   This is not only a problem
>>  for the install CD, but apparently also happens for folks building
>> from source, and from users trying to install RPMs by hand.
>
>> ...
>
>> Therefore, I am against any patch which pulls in yet another library,
>>  particularly one which implements just a little graphical
>> improvement, i.e. some eye candy.
>
> I completely agree with you that adding more dependencies is bad.

We agree!

> However, GTK already depends on Cairo since version 2.6. So technically,
> this patch isn't adding any new dependencies (only adding a requirement
> for Gtk >= 2.6). If you compile an unpatched version of Gschem on a
> recent system, you will see that it already gets linked with libcairo.
>
> As far as I know, GTK itself is moving to Cairo as its drawing backend.

Right now gEDA/gaf needs only GTK-2.4.  This is a policy set by Ales.
It can be changed, but first we'd want to have a discussion.  I guess
the question is:  When did the major distros cut into GTK-2.6?  If it
is sufficiently long ago, then contemplating a change is not out of
the question.  However, if the cut-in is recent (only a year or two
ago), then it may be too early to mandate GTK-2.6.

FWIW, I personally know a user who still runs gEDA on a Red Hat 7.2
box.  He hasn't upgraded his gEDA version in years since he doesn't
want to break his carefully-constructed design environment.  This guy
is not crazy.  Rather, he's a typical user.  He uses gEDA at work,
has something working and he doesn't want to lose it for the latest
whizzy features apparent only to developers.  Folks like this are
amongst the users we want to have.

> Besides the eye-candy, I believe this patch also shows that perhaps with
> time it would be wise to make changes to libgeda so that higher
> precision coordinates of objects are accessible. Cairo obviously isn't
> useful for Gschem yet, but at the moment it seems that this is the way
> GTK will evolve in the future, so it wouldn't hurt if Gschem is prepared
> for it.

I don't disagree.  I do think the issue of how recent the GTK-2.6
cut-in is will gate this decision to some degree.

Stuart



_______________________________________________
geda-dev mailing list
geda-dev@moria.seul.org
http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev