-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Tue, 5 Nov 2002 00:51:56 -0600, mark wrote:
> I didn't feel like arguing any more, when I knew whoever it was who > accused me of spreading FUD was wrong. Several people. Feel free to add me after reading this message. ;p > Well, I found I needed to recompile, after having tweaked a few config > > changes. Then I went to compiling the kernel, and gcc 2.96: > a) I ran memtest for about 35 or 40 min, and got through pass 0 with > no errors. This suggests no major memory chip errors; It's uncertain whether memtest86 is bullet-proof. When it finds errors, it usually is right. But when it fails to find errors, this doesn't mean you're really 100% free of errors. And btw, other components in your system could still be unstable and would cause problems during stress-tests. Compiling a kernel is a sort of a stress-test for CPU, memory and related components. > b) I rebooted to single user mode, and unmounted everything other > than / > > and started compiling. And compiling. And compiling. Never did get > past IPV4. > > SEGVs. Floating point exceptions. Parse errors. Exactly these symptoms are an indication of faulty hardware. > Just for the fun of it, I looked at one header file it claimed had a > missing semicolon. Nope. Exactly this symptom is an indication of faulty hardware. > And I *think* I know a little bit about C, > having been writing C since 1989. > > FUD my ass. gcc 2.96 is *broken*. It kept telling me to go to > Bugzilla, itself. The compiler error/status messages can be confusing when hardware is faulty and corrupts memory and I/O. It it were a broken compiler, it would stop at always the same source file with always the same error message. Running "make" again immediately after an error would not cause it to compile the next few files without errors. > Now, whoever was claiming claiming I was spreading > FUD, if y'all want to come over to Chicago, and try and run compile > the kernel on my machine, drop me an email offline. Why take a long way to a distant location just to use a broken computer? Many users world-wilde can compile the kernel just fine with the same software, but on good hardware. > Otherwise, when I > suggest to folks that they really *do* need to either upgrade or > downgrade, > ->SHUT UP<-. > unless you think you can *prove* that I'm wrong, by giving *evidence*, > which is defined as demonstrable to an unbiased third party. The described symptoms are an indication of faulty hardware. Period. - -- -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) iD8DBQE9x5ty0iMVcrivHFQRAlOWAKCGeVB3KW7KtTGKSXBLyBBo2vNQSQCeK1L8 reLNNIm1vuTkQzA1zFacUww= =sjYQ -----END PGP SIGNATURE----- -- redhat-list mailing list unsubscribe mailto:redhat-list-request@;redhat.com?subject=unsubscribe https://listman.redhat.com/mailman/listinfo/redhat-list
