Re: [Pan-users] Re: Multipart Messages

[David Shochat]

Michael Satterwhite wrote:
> I was going to compile it because 0.14.2 is the latest stable version. 0.131 
> is beta. 
>   
The current version (which is a complete rewrite), although it is 
technically beta, is vastly superior to 0.14.2, particularly in the area 
of binaries, which, based on your earlier question, is relevant to you.
> Here are the compile errors:
>
> gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I../.. -I/usr/local/include/pan -DGNOMELOCALEDIR=\""/usr/local/share/locale"\"  -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12   -pthread -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include   -I/usr/include/libxml2        -g -O2 -I. -c 
> msort.c
> msort.c: In function ‘msort_with_tmp’:
> msort.c:68: error: invalid lvalue in increment
> msort.c:69: error: invalid lvalue in increment
> msort.c:74: error: invalid lvalue in increment
> msort.c:75: error: invalid lvalue in increment
>   
This intrigued me, so as an experiment, I downloaded the source for 
0.14.2 and tried to build it. Well, I get EXACTLY the same compile error 
you do. Now I know that when 0.14.2 was current (which is almost 4 years 
ago), I was able to compile it. I think the code has a subtle dependency 
on whatever version of gcc was current at the time. So if you want to do 
it just for the challenge of it, I can only speculate that you would 
have to get an older gcc (and perhaps build it). Now I see no practical 
reason for doing this (see my previous comment), but I have certainly 
been known to do things like this just for the challenge, even when that 
was the only reason.

The 4 offending lines involve constructs like (lines 68 and 69):
 *((unsigned long int *) tmp)++ = *((unsigned long int *) b1)++;
Maybe there's compile switch that would make the current gcc (4.1.2) 
accept that. I have also converted the offending lines to a more 
simple-minded equivalent such that msort.c (and the rest of the source) 
compiles cleanly with current gcc. I can give you details if you want.
-- David