On Wed, Jan 12, 2005 at 07:40:40AM +0100, Andreas Jochens wrote:
> > As I said, I can't check your patch now, but it seems that it only changes
> > ints to longs and similar stuff so it shouldn't break anything, but if you
> > could comment it I will be very thankful.
> 
> The patch indeed only changes some ints to longs. 
> 
> On 32 bit systems like i386 this should not make any difference 
> because ints and longs are both 32 bits. On the amd64 and ppc64
> architectures, an int is 32 bit and a pointer is 64 bit long. So
> the pointer to int conversions in the code will not work for these
> architectures. One could argue that the patch should not replace 
> the (int) casts with (long) casts, but with (intptr_t) casts instead,
> which should be even more portable. However, some older environments 
> do not have 'intptr_t', so 'long' avoids some ugly configure checks
> for these environments.
> 
> I filed a few hundred bug reports with similar patches for other 
> packages. Many of those patches have already been applied either by 
> upstream or by the maintainers. Up to now, there were no problems 
> with theses patches reported for any of the Debian architectures.
> 
> Of course the details of the patches may be discussed, but the 
> casts from pointers to int are definitely not portable and 
> should be changed to something which also works 
> on 64 bit architectures.

Ok here goes upstream comment:

v+
Its just that he changed the coms code, I agree internal code does not
matter but sending a 64 bit type when the client is expecting a 32 bit 
type will cause a problem :)

I have not had time to check the rest of it too carefuly, but I did 
notice this. Rest assured I will port to 64bit or apply the patches given 
time. For now I am sure running in 32bit mode wont cause any problems.

Sorry for any problems, and thanks again,
Kind Regards,
Gavin
v-

And now after looking to these patches there are some network related
changes and we can't assume that all other clients will have fixed version.
I don't want to break compatibility cause these packages are pretty
popular. Maybe we should really wait for changes in *all* clients, I mean
we should wait till new upstream version will be released.

> Thank you for your help!

No no no... I thank you, cause I'm going to buy some AMD64 system really
soon so your work is really useful ;)

regards
fEnIo

-- 
      _  Bartosz Fenski | mailto:[EMAIL PROTECTED] | pgp:0x13fefc40 | IRC:fEnIo
    _|_|_     32-050 Skawina - Glowackiego 3/15 - w. malopolskie - Polska
    (0 0)        phone:+48602383548 | proud Debian maintainer and user
ooO--(_)--Ooo  http://skawina.eu.org | JID:[EMAIL PROTECTED] | RLU:172001

Attachment: signature.asc
Description: Digital signature

Reply via email to