On Tue, Dec 06, 2005 at 03:33:52PM +0100, Frank Küster wrote: > CRUFT BEGIN > /usr/bin/fakeroot: line 152: 24322 Trace/breakpoint trap > FAKEROOTKEY=$FAKEROOTKEY LD_LIBRARY_PATH="$PATHS" LD_PRELOAD="$LIB" "$@" > Build killed with signal 15 after 150 minutes of inactivity > ****************************************************************************** > Build finished at 20051203-0616 > FAILED [dpkg-buildpackage died] > > Don't know whether this is a problem with fakeroot or with the "CRUFT" > in the Makefile.
blech. well, it could be any of the following commands in our list of CRUFT: @echo "CRUFT BEGIN" @find -type l -print0 | xargs --no-run-if-empty -0 rm -v @find -name .deps -type d -print0 | xargs --no-run-if-empty -0 rm -rfv @rm -vrf ndb/docs/.doxy* ndb/docs/*html ndb/docs/*pdf innobase/autom4te.cache @for i in \ readline/Makefile \ sql-bench/Makefile \ scripts/make_win_binary_distribution \ scripts/mysql_explain_log \ scripts/mysql_tableinfo \ scripts/mysqlbug \ sql/lex_hash.h \ strings/ctype_autoconf.c \ config.log \ config.cache \ ; \ do \ rm -vf $$i; \ done @echo "CRUFT END" none of these should endlessly loop. if faulty hardware/disk is suspect, the find jobs might be taking an extraordinary amount of time as a result... i can see a way to optimise the cleanups on the .deps directory (passing -prune to find) but it seems orthogonal to the problem at hand. is it worth uploading a new version with the "@"'s removed, or is there some other way to investigate what's going on? sean
signature.asc
Description: Digital signature