This is the mail archive of the gdb@sourceware.cygnus.com mailing list for the GDB project. See the GDB home page for more information.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: A gdb 4.17 patch for linux/x86


> 
> The changes to the C expression parser are bad.  They cause any 20 digit hex
> number to be converted to a raw DOUBLEST value.  If you *really* want to
> implement something like that, you have to come up with an unambiguous syntax.
> The current scheme will cause unexpected results if you (reasonably) expect to
> use a 20 digit integer.

It should be a 10 byte integer.

> 				Stu

It is conditional on HEX_LONG_DOUBLE_INPUT which is only defined for
Linux/x86. I don't know gdb can store 10 byte integers on Linux/x86.
Besides, you can specify your own syntax for your OS/CPU. For
Linux/x86, treat a 10 byte integer as a floating point number makes
senses to me. If your OS/CPU suppors 10 byte integers, you can choose
a length your OS/CPU doesn't support for integers, say 20 bytes.


-- 
H.J. Lu (hjl@gnu.org)