Apache OpenOffice (AOO) Bugzilla – Issue 6569
core dump during OO build in module "product/util"
Last modified: 2003-03-16 20:12:53 UTC
Minor nit: During the OO 1.0.1 build process, there's an attempt to run a shared library as a command. On solaris, this results in a "core dump" and a failed command execution (which is ignored, so the build continues). The problem is in module "product/util": % pwd /files/src/oo_1.0.1_src/product/util % dmake tr -d "\015" < /files/src/oo_1.0.1_src/solver/641/unxsoli3.pro/bin/../xml/module-description.dtd > ../unxsoli3.pro/bin/udk3.0.1/xml/module-description.dtd rm -rf ../unxsoli3.pro/bin/udk3.0.1/solia/bin/udkapi.rdb ../unxsoli3.pro/bin/udk3.0.1/solia/lib/librmcxt.so.3.0.1 ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libsal.so.3.0.1 ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libstore.so.3.0.1 ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libreg.so.3.0.1 ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libcppu.so.3.0.1 ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libcppuhelper3C52.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libsalhelper3C52.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libjavaloader.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libsunpro5_uno.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libjava_uno.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libiiop_uno.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/liburp_uno.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libstlport_sunpro.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libacceptor.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libbrdgfctr.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libconnectr.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libcorefl.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libcpld.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libdefreg.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libdynamicloader.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libimpreg.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libinsp.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libinv.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libinvadp.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libjen.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libnamingservice.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libproxyfac.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/librdbtdp.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libremotebridge.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libsimreg.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libsmgr.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libstm.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libtcv.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libtdmgr.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libuuresolver.so ../unxsoli3.pro/bin/udk3.0.1/solia/lib/librmcxt.so.3 ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libsal.so.3 ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libstore.so.3 ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libreg.so.3 ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libcppu.so.3 ../unxsoli3.pro/bin/udk3.0.1/classes/sandbox.jar ../unxsoli3.pro/bin/udk3.0.1/classes/jurt.jar ../unxsoli3.pro/bin/udk3.0.1/classes/ridl.jar ../unxsoli3.pro/bin/udk3.0.1/classes/jut.jar ../unxsoli3.pro/bin/udk3.0.1/classes/juh.jar ../unxsoli3.pro/bin/udk3.0.1/classes/java_uno.jar >& /dev/null Segmentation fault (core dumped) rm -f ../unxsoli3.pro/bin/udk3.0.1_odk.tar.gz cd ../unxsoli3.pro/bin/udk3.0.1/.. && tar cvf - udk3.0.1 | gzip -c > udk3.0.1_odk.tar.gz a udk3.0.1/ 0K a udk3.0.1/solia/ 0K ... Note the "Segmentation fault (core dumped)" message in the dmake out above. % ls -l core* -rw------- 1 jk usr 30200 Jul 25 20:30 core-libacceptor.so-16042 (I've changed my "core" file pattern to record the process name & pid)
Created attachment 2350 [details] Patch to remove extra escaped newline at end of "DLLLIST" make variable
There's an extra escaped "newline" in the definition of the "DLLLIST" variable. Somehow, this results in the ADDITIONALFILELIST beeing split into two lines. The "rm -rf ..." command deletes the part of ADDITIONALFILELIST from the first line, and apparently the rest of ADDITIONALFILELIST - starting with ../unxsoli3.pro/bin/udk3.0.1/solia/lib/libacceptor.so -- starts a new command. Fix is to remove the extra backslash in the DLLLIST definition.
checked in on OOO_STABLE_1 and HEAD. thanks
As discussed with Armin, he thought these bug were closed, he sees all of them as verified.