hi list, i just finished building hlfs with glibc 2.6 and gcc 4.2. work (means it did compile ;).
coreutils, tar, gzip needs to be patched because glibc 2.6 defines futimens just like glib, but they are incompatible (as of now). other than that... no problems so far. greets (and keep up the good work !!) michael -----Ursprüngliche Nachricht----- Von: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Im Auftrag von Robert Connolly Gesendet: Freitag, 8. Juni 2007 07:19 An: Hardened LFS Development List Betreff: Re: toolchain stuff On Thursday June 7 2007 06:51:39 pm Kevin Day wrote: > As I was working out problems with uClibc-0.9.29, I discovered that > gcc-4.2.0 is not only more stable than the gcc-4.1.x series, but also > had fewer problems that the gcc-3.4.6. > > It may not be worth rushing it in at this point, but it is very much > worth considering. Last time I looked at a GCC snapshot I noticed they added uClibc support. I'm not sure if it's in gcc-4.2 or gcc-4.3. Binutils, and GDB, should do the same eventually. GCC has the master copy of the top-level makefile (and libiberty), and other packages sync to GCC from time to time. The majority of the gcc-4.2 differences I noticed were better, more specific, warning flags. I didn't see why it's qualified for a minor version number change. From what I see on google, it doesn't look like gcc-4.2 breaks any packages. GCC-4.2 also has improves "combined tree" support, and now it bootstraps by default. Glibc-2.6 does break some packages. Glibc-2.5 doesn't have great linuxthreads support, so I'm strongly considering disabling libc threads for Linux-2.4, and adding gnu's portable threads package. The alternatives are not great... like using linuxthreads-snapshots, which are not maintained well, or patch Linux-2.4 for nptl, which is also not maintained well. Using PTH would be more stable. There are a few architectures that disable threading in Glibc, so it should be reasonably supported. robert -- http://linuxfromscratch.org/mailman/listinfo/hlfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
