I know you can probably come up with a better argumented response that I can.. Wichert.
--- Begin Message ---
- To: Wichert Akkerman <wichert@cs.leidenuniv.nl>
- Cc: Andi Kleen <ak@muc.de>, g200-dev@lists.openprojects.net
- Subject: Re: [g200-dev]Re: glx problems
- From: Andi Kleen <ak@muc.de>
- Date: Wed, 30 Jun 1999 20:49:59 +0200
- Message-id: <19990630204959.A2928@fred.muc.de>
- In-reply-to: <19990630185104.A3004@soil.nl>; from Wichert Akkerman on Wed, Jun 30, 1999 at 06:51:05PM +0200
- References: <377643AD.8E98735@earthlink.net> <Pine.SGI.3.94.990627164338.31302B-100000@heorot.engr.sgi.com> <19990630013546.A21975@mors.net> <k21zetadiw.fsf@zero.aec.at> <19990630185104.A3004@soil.nl>
On Wed, Jun 30, 1999 at 06:51:05PM +0200, Wichert Akkerman wrote: > Previously Andi Kleen wrote: > > Erm, I think you misunderstood the glibc FAQ. Although there are often > > conflicts between glibc and kernel includes the kernel includes still can > > be used in some cases (and have to in some). The user just has to know > > what he is doing. > > There is a big difference between a user who knows what he is doing > and making those symlinks. Those symlinks *will* break some programs > at some point, and if you want to use the kernel includes you can tell > your compiler to use those directly anyway. I am not sure what if I follow you, but if you think that everything that the kernel has to offer is covered by glibc I tell you that this far from true. Also not that glibc itself uses asm/* and linux/* includes. The kernel is very careful to keep binary compatibility so it is no problem to keep a symlink to your kernel sources. The Debian way of splitting them is obviously misguided, because it assumes that the kernel is a static thing, which it is definitely wrong. -Andi -- This is like TV. I don't like TV.
--- End Message ---
Attachment:
pgpM6zVTIaS6q.pgp
Description: PGP signature