Re: gcc 2.95 support

[Owen Taylor <otaylor redhat com>]

On Sun, 2003-05-25 at 15:31, Manuel Clos wrote:
> Bastien Nocera wrote:
> > They're simply bugs. The same kind of thing happens with endian issue
> > (although to a lesser extent). Do like everybody else, fix the bugs...
> 
> The amount of work involved is not proportional to the bug.
> 
> See the bug. Save the original file. Modify the file. Compile to check. 
> Check bonsai to see if the bug is still present or the developer already 
> patched it. Check bugzilla to see if someone already opened a bug and 
> sent a patch. If not, make a diff, open a new bug, attach the simple diff.

A patch for this type of stuff isn't necessary. And as long as you
are using the most recently released version, it isn't necessary
to check CVS. Dup'ing bugs like this when they occur isn't a 
time consuming activity.

Just go to bugzilla.gnome.org/enter_bug.cgi, file a new bug:

 Bad variable declaration at foo.c line 69

 There's a variable not at the beginning of the block at
 line 69 of foo.c

click submit. Shouldn't take more than 45 seconds.

> Isn't it more easy that developers compile using -C97 or whatever?

There is a potential danger with that in that GLib will
in some cases use C99 constructs when available and have fallbacks
for lesser compilers. And other modules may do the same.

If you file the bugs, people will learn what to avoid.

Regards, 
                                  Owen