4 ms·
> Numerous people are trying to track down the problems with gcc 2.8.0 and the kernel. The problem seems to be an obscure, optimization-related thing. For now,
by nineteen999 7y ago
> Numerous people are trying to track down the problems with gcc 2.8.0 and the kernel. The problem seems to be an obscure, optimization-related thing. For now, compile kernels with a 2.7.2 release of gcc, or use egcs.
That brings back memories. I think the kernel didn't compile correctly with gcc again until at least 2.95.2 or 2.95.3?
- akkartik 7y agoAt which point egcs was gcc.
- nineteen999 7y agoYou are absolutely right. https://slashdot.org/story/99/04/20/1453234/egcs-to-become-gcc https://slashdot.org/story/99/04/20/1453234/egcs-to-become-g... The funny thing was, gcc 2.95.3 (being mostly egcs as you point out) was more stable than both previous versions of egcs and gcc. I distinctly remember gcc 2.8 having problems with the kernel, and egcs having problems with various userland stuff (mostly GNOME). Having to deal with two compiler forks each with their showstopper bugs seemed to go away rather markedly with 2.95.3. EDIT: found a discussion which covers the idea that it was code in both the kernel/userspace which relied on UB or bugs of gcc 2.7.x https://groups.google.com/forum/#!topic/comp.os.linux.development.system/lqaFc2jAKPE%5B1-25%5D https://groups.google.com/forum/#!topic/comp.os.linux.develo...