Xconq for Windows

[Stan Shebs]

Hans Ronne wrote:

>>The theory of having x11/xconq.c was to provide a way to stitch everything
>>that had to be X11-specific into the tcl/tk interface, without having to
>>conditionalize all the tcl/tk bits internally.  Originally the tcl/tk
>>dir was supposed to contain only platform-independent co
>>
>
>Yes, I can see the logic. But when I added the MacTCL interface, the tcltk
>directory seemed like a more logical place to put tkmac.c and tkmac.r than
>the mac directory, which contains a completely different interface. I'm
>sure that some people trying to build the MacPPC app would have included
>them in their project otherwise.
>
Hmm, good point.

>
>Perhaps the best solution would be what the tcltk people did themselves,
>i.e. to put the platform-specific code in separate mac, win and unix
>directories within the tcltk folder? That would organize things nicely, and
>make it obvious what interface these files belong to. It's pretty much how
>the SDL sources are organized, too.
>
I've always found it convenient to stick to a two-level structure, so
that */* is guaranteed to get all the source files.  Since that's a
practical benefit rather than esthetic, and since the tcltk dir isn't
that large, I'd prefer to just leave them in the tcltk subdir.

(Ironically, SDL is one of those projects whose subdiring I found
annoyingly excessive... but I'm sure it's just the bad influence of
GCC and GDB, where subdirs are the rare evil necessity... :-) )

Stan