gEDA-dev: DBUS binding to PCB actions

[Dan McMahill]

Ales Hvezda wrote:
> [snip]
> 
>>dbus-0.9.3 + some patches has 95,732 lines of code which is 12 lines shy 
>>of matching whats in the clipper branch of pcb right now.
> 
> 
> 	Yikes.  That's pretty non-trivial amount of code.  Gah, I hope
> it's worth it for a "message bus system".

me too.  I really should try and find some time to read about d-bus a 
bit to see what it really offers.

> [snip]
> 
>>of course my recent experiences say that if you want glib, gtk, cairo, 
>>dbus to even compile (much less work) on solaris, you have to fix them 
>>anyway...  Of course I'd claim that cairo doesn't really work even on 
>>linux since it is lacking support for 8-bit psuedo color displays.
>>
> 
> 
> 	Hi Dan, I don't really know what to make if your last statement.
> On one hand, I am a huge fan of using older machine (all my development
> machines are not cutting edge and that's just fine with me), however on
> the other hand, 8-bit pseudo color display (I understand your setup and
> why, but...)?  It's the usual question that the cairo/gtk+ developers
> have to balance: "How many users are actually running on 8-bit pseudo
> color displays?".  I assume not many and there probably aren't many users
> out there using 1 or 4-bit displays either anymore.  However, that is,
> of course, the beauty of free software, if it is important/relavent
> enough, somebody will/should come along and implement it.
> 
> These posts are quite telling (sorry if they have been posted before):
> 
> http://lists.freedesktop.org/archives/cairo/2006-March/006555.html
> http://lists.zerezo.com/gtk-devel/msg01521.html

I suspect that there are more users than one might guess who have an 
8-bit psuedo color display.  In my case it is a Sun ultra/10.  Yeah it's 
not a cutting edge machine by a long shot, but good grief, a 440 MHz 
ultrasparc with 1Gb of RAM is hardly in the category of a vax, dec 
station, or sparc 10.  There are certainly many many useful and current 
pieces of software (like geda and pcb) that don't make this box even 
break a sweat.

It was only in the last few years that a major CAD vendor finally 
started to support displays other than 8-bit psuedo color.

What's telling to me is that users *have* implemented support for psuedo 
color displays in cairo and submitted patches and those patches 
basically get to sit around and bit rot in bugzilla until some other 
user updates them.  There have been complaints from users who are faced 
with replacing scores of machines if they want to use newer gtk.

I guess my feeling is there is also a difference between an end user app 
and a library which could be considered somewhat low level.  In other 
words, I understand if mozilla/seamonkey doesn't support much but the 
latest systems, but I find it very troubling that something which is
used as a building block by _so_ much software takes that attitude.  It 
makes me wonder what else is next.  Maybe something related to compiling 
on MacOS or maybe big endian machines?  Maybe I've just been lucky but 
with the exception of cadence, cairo is the only X client I've used in 
20 years that refused to function at all on some of the displays I was 
using.

This all makes me glad I kept gtk1 support in wcalc....

-Dan