[tex-k] Problems of Xdvi in CYGWIN

[Paul Vojta]

> From: "Rodrigo Medina" <rodmedina at cantv.net>
> To: tex-k at tug.org
> Date: Wed, 23 Jul 2003 09:43:45 -0400
> Subject: [tex-k] Problems of Xdvi in CYGWIN
> 
>  Dear Sirs;
> 
> I have just upgraded xdvik to version 22.76.1
> My system is a windows95  box with cygwin version 1.3.22 and 
> XFree86 (X11R6). I compiled with gcc 3.2.
> 
> I had the following problems:
> 
> 1) The linker complianted because:
> /usr/X11R6/lib/libXt.a(NextEvent.o)(.text+0xa84):NextEvent.c:
> multiple definition of `_XtRemoveTimeOut'
> events.o(.text+0x54d0):
> /home/rmedina/xdvi/xdvik-22.76.1/texk/xdvik/events.c:3376
> : first defined here
> /usr/X11R6/lib/libXt.a(NextEvent.o)(.text+0x970):NextEvent.c:
>  multiple definition of `_XtAppAddTimeOut'
> events.o(.text+0x5360):
> /home/rmedina/xdvi/xdvik-22.76.1/texk/xdvik/events.c:3334
> : first defined here
> 
> I commented out the functions in events.c and the compilation
> arrived to the end.

After configuring but before compiling, edit the Makefile in texk/xdvik/
and add:

	-Xlinker --allow-multiple-definition

to the definition of x_link.  That should work for now.

Or, you can leave it as is, if you're using lesstif (as opposed to
Open Motif, which might not even be available for Cygwin).

> 2) The xdvik that was produced failed to find almost any file. So it
> was continuously making fonts like cmr10.600pk.
> Nevertheless  the tetex distribution that comes with cygwin is OK.
> The kpsewhich that comes with that distribution is OK.
> I found that the problem was due to the kpathsea library version 3.4.5
> that comes with xdvik-22.76.1. In config.h  CYGWIN is defined as
> DOSISH, so the path-separator is defined as ";". That IS WRONG,
> for CYGWIN the path separator is ":", exactly as in standard UNIX.
> Compiling with a modified config.h, in which the lines in which
> CYGWIN appears were commented out, yielded a perfectly behaving
> executable. 

Sorry, I can't help you there.

> I hope that this comments may be useful.

Yes, thanks very much for the bug report.

--Paul Vojta, vojta at math.berkeley.edu