Re: [PATCH] strict types in code generators.

[Marcin Dalecki <martin at dalecki dot de>]

On 2005-03-02, at 19:41, Zack Weinberg wrote:

> Marcin Dalecki <martin@dalecki.de> writes:
>
> > On 2005-03-02, at 18:47, Zack Weinberg wrote:
> > > Marcin Dalecki <martin@dalecki.de> writes:
> > > > As one would expect C++ simply doesn't give global scope to the
> > > > functions in question:
> > >
> > > See, that's the thing, one does *not* expect that.

Yes indeed. Blame it on my dusty C++.

> > > I tried to reproduce this and ran into too many other problems.  
> > > Would
> > > you please try to construct a small, self-contained test case which
> > > demonstrates the problem you are seeing?
> >
> > Under C there is no formal problem with this. However there are tons
> > of other things that have to be fixed (semi mechanically) before the
> > compiler can be feed through C++:
>
> Yes, I noticed.  No, I'm not interested in looking at those patches.
> I'm interested in figuring out why you seem to need a change that
> should be unnecessary.  Please do try to construct a test case as I 
> asked.

The attached tar should contain the situation as I see it. However
contrary to the GCC code... the test case actually works. Maybe it's 
just that
the linker I'm using is utterly broken. Another explanation would be of 
course
a miss-match between the definition and declaration type signature. 
Anyway I will
have to investigate this a bit further.

Attachment: tc.tar.gz
Description: GNU Zip compressed data