[Dwarf-discuss] about the static member function's debuginfo

[Wu Zhou]

It seems that G++/GDB also handle static member function this way.

----- Original Message ----- 
From: "Ron 603-884-2088" <brender@gemevn.zko.dec.com>
To: <DWARF-DISCUSS at base3.freestandards.org>
Cc: <BRENDER at gemevn.zko.dec.com>
Sent: Monday, April 25, 2005 9:44 PM
Subject: Re: [Dwarf-discuss] about the static member function's debuginfo


>
> >I am now studying the DWARF output of two different c++ compiler. It
seems
> >that they generate different debuginfo for static member function of a
> >class. Did the latest DWARF spec had any clean way to discern C++ static
and
> >non-static member functions? As far as I read the DWARF3 draft, there is
no
> >such way. If it is really like that, is there any plan for DWARF standard
to
> >specify an attribute to describe the static feature of a member function?
or
> >any other ways?
>
> No, the DWARF document is not clear on this point.
>
> Locally (for hp I64 OpenVMS) we adopted the following convention:
>
>     A non-static member function of a class (or union, struct, etc)
>     must have an initial parameter whose name is 'this' (which is
>     reserved in C++). It should also have the DW_AT_artificial
>     attribute. Any member function that does not have at least one
>     parameter, or whose first parameter does not have a name, or
>     whose first parameters name is other than 'this', is treated as a
>     static member function.
>
>     The type of the first parameter of a non-static member function
>     will necessarily be a pointer type referring to the immediately
>     containing class (or union, struct, etc). However, this is not
>     checked in determining whether or not a member function is
>     static.
>
> It would certainly be appropriate for DWARF to adopt a definitive
> convention.
>
> Ron Brender
>
>
> _______________________________________________
> Dwarf-discuss mailing list
> Dwarf-discuss at mail.freestandards.org
> http://mail.freestandards.org/mailman/listinfo/dwarf-discuss
>
>