>> Allan Rae writes:
AR> On 15 Mar 1999, Lars Gullik Bjønnes wrote:
>> >> Allan Rae writes:
AR> Done. Now a much revised patch is attached. This direct against
AR> latest cvs and includes a couple of other fixes I had to do to
AR> get LyX running or compiling. Two "no nulls allowed" fixes and a
AR> definition for LChar so Asgers Encoding* would compile.
>>
AR> This time I use a #define to get a named variable rather than a
AR> temp. The temp was disappearing too early otherwise as you
AR> suspected.
>> Ok, the patch looks good, apply that.
>>
>> I could also use some help in conformance testing LString against
>> basic_string<char> (string), and fixing the places where it is
>> not conforming.
AR> Do you think we could do some sort of shadowing of all LString
AR> operations with a basic_string<char> ? Then we could make that a
what if you dont have basic_string?
AR> part of the invariant testing. This would require that we add a
AR> basic_string<char> to LString etc. but I fear we'd run into
AR> problems with all the extra function calls and overlapping of
AR> code. Hmmmm...
I thought of implementing LString on top of the non stl string
supplied with gcc. Only a few methods are missing.
Lgb
