[Bug#64801] URGENT: Please evaluate this easy fix enabling wxgtk2.1 to build!
NOTE: the bug in question is filed as Bug#64801; please comment to that
bug report and to -devel (if still appropriate to do so).
> Date: Sat, 27 May 2000 19:28:49 EDT
> To: Jim Lynch <jim@laney.edu>
> cc: debian-devel@lists.debian.org
> From: Mike Bilow <mikebw@colossus.bilow.com>
> Subject: Re: URGENT: Please evaluate this easy fix enabling wxgtk2.1 to build!
>
> No, this is a _feature_ of C++, not a bug. The signature of a function is
> defined by its argument types, and C++ regards two functions with the same
> name but different signatures as two different functions. In standard C,
> you would get the same function, but a silent cast might occur.
I agree with this; the fault is with wxwin. I would have filed against
c++ if I didn't believe that.
> Whether bare 'char' defaults to 'signed char' or to 'unsigned char' is
> implementation dependent, however. Most likely, if this got through from
> upstream, it was because the upstream maintainer's implementation assumed
> 'signed char' in this case.
Which probably should not be, imo. If it's an incoming string that the
called fn will not be changing, then const char * should always be used.
> Most machines are more efficient when handling unsigned values than signed
> values, while a few machines are more efficient when handling signed
> values than unsigned values. If the program is limited to being compiled
> with gcc anyway, then the behavior of the the compiler can be changed with
> the '-fsigned-char' switch. The gcc default is to use whichever mode is
> more efficient for the target platform.
Then wxwin should be self consistant. In any case, the package will not
build from source without this patch or something like it.
-Jim
---
Jim Lynch Finger for pgp key
as Laney College CIS admin: jim@laney.edu http://www.laney.edu/~jim/
as Debian developer: jwl@debian.org http://www.debian.org/~jwl/
Reply to: