This is a quad opteron machine. When trying to compile the postgresql80-server port, using the HIER option, eventually you get this: cc -O -pipe -O3 -funroll-loops -Wall -Wmissing-prototypes -Wpointer-arith -Wdeclaration-after-statement -Wold-style-definition -Wendif-labels -fno-strict-aliasing -I. -I../../../src/include -I/usr/local/include -I/usr/local/include -c -o keywords.o keywords.c keywords.c:84: error: `CONNECT' undeclared here (not in a function) keywords.c:84: error: initializer element is not constant keywords.c:84: error: (near initialization for `ScanKeywords[51].value') <snip repetitive errors> keywords.c:346: error: initializer element is not constant keywords.c:346: error: (near initialization for `ScanKeywords[313]') gmake[1]: *** [keywords.o] Error 1 gmake[1]: Leaving directory `/usr/ports/databases/postgresql80-server/work/postgresql-8.0.2/src/backend/parser' gmake: *** [parser-recursive] Error 2 *** Error code 2 Stop in /usr/ports/databases/postgresql80-server. How-To-Repeat: cd /usr/ports/postgresql80-server; make (make sure to include HIER option)
Responsible Changed From-To: freebsd-ports-bugs->girgen Over to maintainer.
State Changed From-To: open->feedback I cannot reproduce this on a dual Xeon running amd64, with HIER and OPTIMIZED_CFLAGS on. There are known problems with 5.4 on systems with more than four real CPU:s according to http://www.freebsd.org/releases/5.4R/todo.html, although it seems Doug is at it. My experience, at least with amd64, is that two CPU:s running hyperthreading are also not stable. Perhaps this is what's happening? Does it fail on the same sport every time? Can you get it to fail on a another machine?
State Changed From-To: feedback->closed Fixed as suggested in ports/84720