On Sun, 28 Nov 2010 09:51:46 -0700, Alex Rousskov <[email protected]> wrote: > On 11/27/2010 12:00 AM, Amos Jeffries wrote: >> There are some build problems which are not being highlighted by the >> hudson testing. Picking on FreeBSD-6.4 build because of problem #2. > ... >> libtool: compile: g++ -DHAVE_CONFIG_H -I../../.. -I../../../include >> -I../../../lib -I../../../src -I../../include -I/usr/local/include >> -I/usr/include -I/usr/include -I/usr/local/include -I/usr/include >> -I/usr/include -Wall -Wpointer-arith -Wwrite-strings -Wcomments -Werror >> -pipe -D_REENTRANT -D_FILE_OFFSET_BITS=64 -g -O2 -MT gadgets.lo -MD -MP >> -MF .deps/gadgets.Tpo -c ../../../src/ssl/gadgets.cc -fPIC -DPIC -o >> .libs/gadgets.o >> ../../../src/ssl/gadgets.cc: In function `bool >> addCnToRequest(Ssl::X509_REQ_Pointer&, const char*)': >> ../../../src/ssl/gadgets.cc:17: warning: deprecated conversion from >> string constant to `char*' >> *** Error code 1 >> mv -f .deps/context_storage.Tpo .deps/context_storage.Plo >> 1 error >> *** Error code 1 >> 1 error >> *** Error code 2 >> 1 error >> *** Error code 1 >> 1 error >> { test ! -d squid-3.HEAD-BZR || { find squid-3.HEAD-BZR -type d ! -perm >> -200 -exec chmod u+w {} ';' && rm -fr squid-3.HEAD-BZR; }; } >> ================================================== >> squid-3.HEAD-BZR archives ready for distribution: >> squid-3.HEAD-BZR.tar.gz >> squid-3.HEAD-BZR.tar.bz2 >> ================================================== >> buildtest.sh result is 0 >> BUILD: .././test-suite/buildtests/layer-02-maximus.opts >> *** Error code 1 >> *** Error code 1 >> *** Error code 2 >> *** Error code 1 >> Build OK. Global result is 0. >> " >> >> The execution of those test line appears to overwrite the $? magic >> depended on by buildtest.sh. > > Agreed. We should not be building the distribution archives if the > previous build targets fail. > > The "archives ready for distribution" lines appear to come from the > automake-generated distcheck target. Is this an automake bug?
It came into view because a manual run of the testbed on west dies. Yet the hudson run on same box indicates everything is fine. I'm not sure where the bug is. Trying one re-run now after clearing the workspace. > > I assume we bootstrap on FreeBSD. Is this bug specific to automake on > FreeBSD? Can you compare the distcheck target of root Makefile.in on > FreeBSD and Linux and see if there are any important differences? If > yes, perhaps we can upgrade automake on the FreeBSD host? This particular warning is about C-style cast being used. The other OS don't complain about it so we can't see at present how they react. Amos
