libXaw vs libXaw3d

[Matt Turner]

On Mon, Apr 2, 2012 at 5:13 PM, Alan Coopersmith
<alan.coopersmith at oracle.com> wrote:
> On 04/ 2/12 03:21 AM, Marty Jack wrote:
>> Unknown if you might have meant to add libXaw3d 1.6.2
>
> I did not, as nothing else in the katamari depends on it, and if you're
> building a new system from scratch you may never need it if you're not
> building one of the legacy clients that uses it instead of a modern toolkit.
>
> I'd drop libXaw from the katamari if I could, but we still have clients in
> the katamari that use it.
>
> I would consider replacing libXaw with libXaw3d in the future, if all the
> clients in the katamari were changed to use it (perhaps allowing a choice
> of which Xaw to use in a configure macro?).

Perhaps a new thread is in order..

It seems to me that if we want to replace libXaw with libXaw3d we
should first fix as much of the code as possible to be identical. Now,
libXaw3d's code looks like libXaw's from 20 years ago with some
add-ons. libXaw has seen lots of whitespace clean ups,
s/Boolean/Bool/, and lots of other simple changes that make finding
actual differences annoying.

I'd do this, except I actually don't care a thing about either
library. (The only reason I touched libXaw3d to begin with was to
replace its imake build system with autotools.)

I'd suggest someone else do it that does care about either library,
but it seems that the only reasons projects might use libXaw[3d] is
because (1) no one cares about those projects enough to port them to
something else, and (2) the project authors actually want to avoid
better toolkits.

I think the other alternative, which is ignoring the libXaws whenever
possible, is rather compelling.