| Summary: | kernel 2.6 : gcc 3.3 : glibc-2.3.2 compile error in sysctl.h | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Boyd Waters <waters.boyd> |
| Component: | [OLD] Core system | Assignee: | x86-kernel (DEPRECATED) <x86-kernel> |
| Status: | RESOLVED DUPLICATE | ||
| Severity: | normal | ||
| Priority: | High | ||
| Version: | 1.4_rc4 | ||
| Hardware: | x86 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
|
Description
Boyd Waters
2003-07-28 23:43:22 UTC
this can be found in a few places ... this patch has been submitted to the kernel people but their answer was 'dont use kernel headers in user space' ... in other words, i dont know if we should really do this :x Use the new kernel test : 2.6.0-test2 "Don't use kernel headers in user space" Wonderful. Now if we could just get everyone in the world to comply... ... and things like glibc *need* these headers. That's "user space". What then? I am in the process of developing a set of "sanitized" headers from the kernel headers. I can use the new -test2 kernel sources, but there will still be errors in compiling things here and there. (I have bugs posted regarding X and some other packages.) In general these errors are very simple tweaks to the kernel headers. But note that I generally do *not* tweak the headers in /usr/src/linux, but rather the headers in /usr/include. Someone has to do this to get us away from kernel headers from 2.4.19 in /usr/include. Some tweaking is necessary. Glibc is a special case. It needs the headers from the running kernel. Linus Torvalds says that the stuff in /usr/include should be the headers from the kernel used to build glibc. So some tweaking to the kernel headers to make sure that glibc can compile is probably a pragmatic thing. Reporting to the glibc maintainers is a good thing, too. Bottom line: The kernel maintainers need to work with the glibc maintainers to get these two to compile. The new ebuild contains a hack^H^H^H^H work-around that fixes this bug. See http://bugs.gentoo.org/show_bug.cgi?id=21326 I suggest marking this bug as duplicate of that one. |