Alex Goncharov wrote:
,--- You/Kostik (Wed, 27 Aug 2008 19:04:32 +0300) ----* | cd into /usr/src/gnu/usr.bin/cc/cc1, | and do | make install DEBUG_FLAGS=-gThat simple thing didn't work for me: make install DEBUG_FLAGS=-g install -o root -g wheel -m 555 cc1 /usr/libexec install: cc1: No such file or directory *** Error code 71 I updated and installed world this morning -- and indeed had this gcc: Internal error: Abort trap: 6 (program cc1) while compiling something, as a result. But I've found what may be an extremely easy way out of this hole: I copied over /usr/libexec/cc1 from another machine, which had a code built on Aug 17. Then my `cc1' worked fine and I tripped on a SEGV in `as'. Then I copied over `as', then `ld', then `gcc'. And that was it -- the build began to work. So if anybody wants it, I can just send a package with these four executables, to be deployed in place of the broken ones. Let me know... -- Alex -- [EMAIL PROTECTED] -- /* * The end of the world will occur at 3:00 p.m., this Friday, with * symposium to follow. */
Yes, copying binaries from one machine to another helped for me, too, but I also simply copied /boot/kernel.old to /boot/kernel on one of the defective servers and things obviously ran fine again ... But I was trapped by another pitfall, I do compile on specific machines with specific compiler/arch/cpu option flags, so core2 based boxes compile with 'nocona' while older AMD Athlon64 based machines compile with 'k8'. That resulted in the past in crashes and in this case, too.
Hope this weirdness gets fixed soon ... _______________________________________________ [email protected] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "[EMAIL PROTECTED]"
