map

[Alexandre Oliva]

Jochen Kuepper <jochen@uni-duesseldorf.de> writes:

> On Thu, 01 Jan 1970 Alexandre Oliva wrote:
>> Jochen Kuepper <jochen@uni-duesseldorf.de> writes:

>>> In the class header file (stl_map.h) there is only 

>>> T& operator[](const key_type& k);

>> As the standard mandates.

> So is this considered a bug of listdc++ as of egcs-1.1 ?

No, egcs must follow the standard with respect to this decision.

> As I see it is save to use the non-const function on const objects in the
> current implementation - is that right ?

No, it is never safe to do that.  The underlying map object may be
really const.

> Btw: Will there be something like egcs-1.1.[12] ?

There will probably be 1.1.1; I hope 1.1.2 isn't necessary.

-- 
Alexandre Oliva
mailto:oliva@dcc.unicamp.br mailto:aoliva@acm.org
http://www.dcc.unicamp.br/~oliva
Universidade Estadual de Campinas, SP, Brasil