Okay, somehow this "error" did not appear ever from numerous builds, until I started cross-compiling everything to i686. (instead of compiling it just for my pentium4)
While compiling GCC 3.4.5, an error occured to the toolchain, the file ctype_noninline.h was checking for glibc version with (__GLIBC__ > 2 || __GLIBC_MINOR__ > 2) to support two different versions of tupper/tolower typecasts. The error occurs when the typecast supported is in fact the one that GCC claims glibc >= 2.3 supports. At first glance, I was just going to modify that line, but looking further, I saw similar declarations all over the place. Using a grep to /usr/include, I found that a number of the newer functions that are available, are being disabled by the __GLIBC_MINOR__ checks in different programs, including Xorg. After some quick research, I found the following: www.uclibc.org/lists/uclibc-cvs/2003-December/007798.html Though, I have not touched xlocale in anyway (it may be enabled by default..) So, is it better to do this change or change every single line in gcc that is effect? Is there any chance of having some sort of option being passed such that something like the following could be used: #if !defined __FORCE_NOGLIBC && (!defined _LIBC || defined __FORCE_GLIBC) #ifndef __USE_NEWER_GLIBC__ # define __GNU_LIBRARY__ 6 # define __GLIBC__ 2 # define __GLIBC_MINOR__ 3 #else # define __GNU_LIBRARY__ 6 # define __GLIBC__ 2 # define __GLIBC_MINOR__ 2 #endif /* __USE_NEWER_GLIBC__ #endif so that individual programs can be quickly hacked to use the newer or older version and prevent other programs needing the older from getting messed up?
uClibc-0.9.28-glibc_compat-1.patch
Description: Binary data
-- http://linuxfromscratch.org/mailman/listinfo/hlfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
