Re: Cygwin compilation broken by bin_hook.c

[Hans-Bernhard B. <br...@ph...>]

On Sun, 3 Oct 2004, Daniel J Sebald wrote:

> Hans-Bernhard Broeker wrote:
> 
> >Basically, the problem is that both bin_hook.o and binary.o end up
> >listed in makefile.all, so they will both be compiled and linked.

> For the main gnuplot executable, binary.o should not be linked in. 

Well, then it shouldn't be listed in gnuplot_SOURCES, should it?

>  (Either bin_hook.o is a replica of binary.o or it is empty, depending 
> on a configure switch.)

That's terminally ugly.  We will have to find a better way of doing that.
BTW: it *is* possible to compile the same .c file into more than one .o
file using different switches.  The Windows makefiles have to do that
all the time (because of the stdout/stderr tricks they pull).

> I've looked at this.  I can't seem to generate a representative 
> makefile.all.  (I'm not on windows.)  

You don't have to be.  Just delete it and watch it magically get
recreated, from the contents of Makefile.am.

> BTW, why is makefile.all in the CVS source tree, if it is to be 
> generated by automake?

Because it isn't.  It's generated by Makefile.maint, which is supposed
to be run only by us maintainers (usually by Lars), *before* automake.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.