Re: Some observations building gnuplot for win by gcc-4.9.2 (MinGW64)

[Allin C. <cot...@wf...>]

On Fri, 6 Mar 2015, Tatsuro MATSUOKA reported from gdb:

> Program received signal SIGSEGV, Segmentation fault.
> 0x6862318f in g_utf8_validate (str=0x2888a68 "-1",
>  max_len=-1, end=0xffffffff)
>   at ../../glib-2.42.1/glib/gutf8.c:1634
> 1634     *end = p;

I wonder if this is to do with a mix-up between C and C++ compilers.

The gnuplot code passes NULL as the last argument to g_utf8_validate. 
I gather that C++ compilers often represent NULL as all-bits-one -- 
hence, perhaps, the "0xffffffff" above? But apparently glib is not 
reading 0xffffffff as NULL (probably expects NULL = all-bits-zero).

When I build current gnuplot on Linux, gp_cairo.o is compiled by gcc, 
and then g++ is used as linker in making the gnuplot binary. Could it 
be that gp_cairo.o is being compiled by g++ in Tatsuro's case?

Allin Cottrell