On Sun, Aug 11, 2013 at 07:25:20PM -0500, Bruce Dubbs wrote: > Ken Moffat wrote: > > On Sun, Aug 11, 2013 at 11:03:46PM +0100, Ken Moffat wrote: > >> I haven't built a new system since April, so I thought it was time > >> to test current -svn. In particular, I wanted to see how bison-3.0 > >> impacted other packages. But I can't even build it! > >> > > Looks as if it is breaking bacause of one of my environment > > variables - a plain build is ok. But then it builds ok (by hand) if > > I dump the environment with printenv, export all the obvious > > candidates, and then build by hand. > > > > I'll also note that when it does build by hand, the testsuite fails > > with: > > make[3]: Entering directory `/usr/src/bison-3.0' > > YACC examples/calc++/calc++-parser.stamp > > CXX examples/calc++/examples_calc___calc__-calc++-driver.o > > LEX examples/calc++/calc++-scanner.cc > > CXX examples/calc++/examples_calc___calc__-calc++-scanner.o > > g++: error: ./examples/calc++/calc++-scanner.cc: No such file or > > directory > > Now that's confusing. I have ./examples/calc++/calc++-scanner.cc and > all those other files too. The timestamps do indicate that they are > generated though. calc++-parser.stamp is an empty file. > > -rwxr-xr-x 1 root root 635267 Aug 11 22:09 calc++ > -rw-r--r-- 1 root root 684 Aug 11 22:08 calc++-driver.cc > -rw-r--r-- 1 root root 1195 Aug 11 22:08 calc++-driver.hh > -rw-r--r-- 1 root root 28865 Aug 11 22:09 calc++-parser.cc > -rw-r--r-- 1 root root 24999 Aug 11 22:09 calc++-parser.hh > -rw-r--r-- 1 root root 7785 Aug 11 22:09 calc++-parser.output > -rw-r--r-- 1 root root 0 Aug 11 22:09 calc++-parser.stamp > -rw-r--r-- 1 root root 1812 Aug 11 22:08 calc++-parser.yy > -rw-r--r-- 1 root root 52056 Aug 11 22:09 calc++-scanner.cc > -rw-r--r-- 1 root root 1921 Aug 11 22:08 calc++-scanner.ll > -rw-r--r-- 1 root root 469 Aug 11 22:08 calc++.cc > -rw-r--r-- 1 root root 6045 Aug 11 22:09 calc++.log > -rwxr-xr-x 1 505 20 958 Feb 20 17:33 calc++.test > -rw-r--r-- 1 root root 82 Aug 11 22:09 calc++.trs > -rw-r--r-- 1 root root 234464 Aug 11 22:09 > examples_calc___calc__-calc++-driver.o > -rw-r--r-- 1 root root 1076880 Aug 11 22:09 > examples_calc___calc__-calc++-parser.o > -rw-r--r-- 1 root root 190784 Aug 11 22:09 > examples_calc___calc__-calc++-scanner.o > -rw-r--r-- 1 root root 71032 Aug 11 22:09 examples_calc___calc__-calc++.o > -rw-r--r-- 1 505 20 2900 Jun 11 14:36 local.mk > -rw-r--r-- 1 root root 5126 Aug 11 22:09 location.hh > -rw-r--r-- 1 root root 4894 Aug 11 22:09 position.hh > -rw-r--r-- 1 root root 3530 Aug 11 22:09 stack.hh > > -- Bruce > With the manual build, which got far enough for the tests to fail, all I have is:
root in chroot ~# ls -l /usr/src/bison-3.0/examples/calc++/ total 28 -rw-r--r-- 1 root root 684 Aug 12 00:08 calc++-driver.cc -rw-r--r-- 1 root root 1195 Aug 12 00:08 calc++-driver.hh -rw-r--r-- 1 root root 1812 Aug 12 00:08 calc++-parser.yy -rw-r--r-- 1 root root 1921 Aug 12 00:08 calc++-scanner.ll -rw-r--r-- 1 root root 469 Aug 12 00:08 calc++.cc -rwxr-xr-x 1 505 20 958 Feb 20 17:33 calc++.test -rw-r--r-- 1 505 20 2900 Jun 11 15:36 local.mk root in chroot ~# But that is from the second problem (make check fails), and since I can't build bison with my currrent scripts it isn't my main concerm. The failure to build seems to be a local problem (i.e. specific to my scripts or environment). I was starting to wonder if the currently running kernel (3.11-rc) might be involved, so I tried a manual build on the host system, followed by make check : ## ------------- ## ## Test results. ## ## ------------- ## 409 tests were successful. 23 tests were skipped. make[3]: Leaving directory `/scratch/ken/bison-3.0' That looks good, so I guess that whatever is in the book is probably fine - you maybe recall that I had a lot of problems with automake tests in the past (into the 7.3 release, I think), caused by me setting SCRIPTS in my buildscripts. This is probably something similar (or maybe two similar things, since the tests don't work for me in chroot even when run manually). Perhaps it is time for me to stop pretending that I understand any of this stuff, and just stick with bison-2.7.1 until that is no longer adequate. That way, I won't have to dig out the patches to fix things broken by bison-3.0. Mmm, *almost* sounds attractive :-) If that turns out to be my best plan, "thanks for all the fish". ĸen -- das eine Mal als Tragödie, das andere Mal als Farce -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
