Derek Atkins wrote:
John Ellson <[EMAIL PROTECTED]> writes:
The majority of the casts were fixing the use of unsigned char (xmlChar) for libxml2 routines versus signed char (char and gchar) in routines like strcmp.
I think that restructuring the code to be more table driven could help, but probably won't completely eliminate the problem which originates from libxml2.
Every cast in the patch fixed a warning message from gcc4. I did not
attempt any more generalized changes.
Perhaps there's a gcc4-specific compiler option to turn off those
warnings that are new-in-gcc4?
Perhaps, but ignoring the warnings is no better than hiding them with casts, imho.
The same warning did catch another problem, the use of gint instead of guint.
In my opinion, the real problem is the multiple flavors of char (char/gchar/xmlChar)
for the same thing: a code for a character, where the codes are not intrinsically signed or unsigned;
they are just 8bit values use to differentiate one character from another.
John _______________________________________________ gnucash-devel mailing list [email protected] https://lists.gnucash.org/mailman/listinfo/gnucash-devel
