Re: Icon love
- From: Michael Meeks <michael ximian com>
- To: Owen Taylor <otaylor redhat com>
- Cc: Tuomas Kuosmanen <tigert ximian com>, GNOME Hackers <gnome-hackers gnome org>, Jakub Steiner <jimmac ximian com>
- Subject: Re: Icon love
- Date: 18 Jan 2002 18:27:52 +0000
Hi Owen,
On Fri, 2002-01-18 at 17:32, Owen Taylor wrote:
> The one restriction here is that new stock icons are new API
This is rather incredible. I can believe that changing the stock icons
broke some hyfoluted view of ABI, that is if you want to use - as a
value for PI the co-incidental occurance on line 3 of stock pixbuf
gtk-save when scaled to size 13x27.
Is it possible that we can have a concept of reasonableness and hope
that the maintainers might consider reasonable ( in the maintainers view
) ABI change between gtk-2.0 and eg. gtk-2.0.1 ? or does every icon
change have a 6+ month delay attached ? ditto for keynav.
Ultimately, at some level ABI compatibility is just nonsense; if we're
ever going to fix bugs, then me casting my function pointer to guchar *
and memcmping for a bit is never going to work - and yes I only want to
do that on 1 platform, with 1 compiler & 1 optimization level ;-) - or
perhaps more feasibly walking over the end of one structure, to the
adjacent one in memory - maybe a bug in my code, but if you fix yours
you break mine.
It is not possible to provide 100% compatibility against morons; there
has to be some level of understanding and concept of reasonableness in
the whole process - and I'm intrigued as to where you put that line, and
why.
Regards,
Michael.
--
mmeeks gnu org <><, Pseudo Engineer, itinerant idiot
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]