[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: libc5 FAQ



Hello,

> 	My contention is: if we are talking about a program that needs
>  kernel headers, and can't be satisfied with the headers included in
>  the libc5-dev package (which corresponds to 2.0.29, or something),
>  that means we are talking about a program that needs some very
>  specific kernel data structures, and can't do with the definition of
>  that data structure included in the libc headers.

I think libc6 will rebuild the .../sys/ headers with scripts from crrent
kernel source. If even libc6 depends on this, why not use ith with libc5.
Simple Example:

/* -- new Socket ioctl included in kernels 2.1.20++ -- */ 
#define SIOC_BLAFASEL 5

If I want to use new feature BLAFASEL i will have to issue an ioctl 5 to the
kernel. If I hardwire IOCTL 5 in my program i will be able to compile it
independend from kernels versions <= 2.1.20, but i wont be able to be sure
that >= 2.1.21 will redefine the ioctl, or even worse, remove it. Its
impossible to refuse to run on 2.1.21++ Kernels, since I would havbe to
release new tools every day. Best thing i do is:

use BLAFASEL if it is defined, dontuse it if not. Then the program will
compile in <2.1.20 (unable to use the new feature) and  compile with 2.1.20
(enabled to use the feature as long as it is there, even after renumbering).
Thats what the headers are for.

BTW: it seems to be very easy to port net-tools for exapmle to libc6, but
the reason fr this is, that the headers will include all the kernel
dependend stuff in a version depending manner. (only the release cycle of
libc6 has to be more faster to include all new features from new kernels).

Greetings
Bernd
-- 
  (OO)      -- Bernd_Eckenfels@Wittumstrasse13.76646Bruchsal.de --
 ( .. )  ecki@{inka.de,linux.de,debian.org} http://home.pages.de/~eckes/
  o--o     *plush*  2048/93600EFD  eckes@irc  +4972573817  BE5-RIPE
(O____O)       If privacy is outlawed only Outlaws have privacy


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .


Reply to: