Yet another build script

Does mostly the same as the rest, but slightly different purpose.

scripts/zmap_one_off_build.sh

+#!/bin/bash
+
+
+trap '' INT
+trap '' TERM
+trap '' QUIT
+
+# ================= README ==================
+
+# Script  to  produce a  binary  for  all  architectures specified  in
+# $ZMAP_BUILD_MACHINES.    This   script   is  almost   identical   to
+# zmap_overnight.sh, but builds no documentation or distribution so is
+# quite a bit quicker.
+
+# Assuming  all builds well,  the $OUTPUT  directory will  contain the
+# results.
+
+# ================== CONFIG ================== 
+# Configuration variables
+
+# SRC_MACHINE= The machine to log into to start everything going
+SRC_MACHINE=tviewsrv
+
+# CVS_CHECKOUT_SCRIPT= The bootstrapping script that starts everything
+CVS_CHECKOUT_SCRIPT=$(dirname $0)/build_bootstrap.sh
+
+# GLOBAL_LOG= The place to hold the log file
+GLOBAL_LOG=~/BUILDS/TODAY.BUILD.LOG
+
+# ERROR_RECIPIENT= Someone to email
+ERROR_RECIPIENT=
+
+# ENSURE_UP_TO_DATE= cvs update the directory where $CVS_CHECKOUT_SCRIPT is [ yes | no ]
+ENSURE_UP_TO_DATE=yes
+
+# OUTPUT dir. If changed, also edit RELEASE_LOCATION= later on...
+# Sadly no expansion of variable is possible there!
+OUTPUT=~/BUILDS/TODAY
+
+# For development make sure these are set
+#CVS_CHECKOUT_SCRIPT=./build_bootstrap.sh
+#GLOBAL_LOG=~/BUILDS/TODAY.BUILD.LOG
+#ERROR_RECIPIENT=
+ENSURE_UP_TO_DATE=no
+
+# ================== MAIN PART ================== 
+MAIL_SUBJECT="ZMap Build Failed (control script)"
+
+if ! echo $GLOBAL_LOG | egrep -q "(^)/" ; then
+    GLOBAL_LOG=$(pwd)/$GLOBAL_LOG
+fi
+
+mkdir -p $OUTPUT
+rm -f $GLOBAL_LOG
+
+# Errors before here only end up in stdout/stderr
+# Errors after here should be mailed to $ERROR_RECIPIENT
+if [ "x$ENSURE_UP_TO_DATE" == "xyes" ]; then
+    old_dir=$(pwd)
+    new_dir=$(dirname  $CVS_CHECKOUT_SCRIPT)
+    up2date=$(basename $CVS_CHECKOUT_SCRIPT)
+    cd $new_dir
+    export CVS_RSH=ssh
+    cvs update -C $up2date || { 
+	echo "Failed to cvs update $CVS_CHECKOUT_SCRIPT"
+	echo "Failed to cvs update $CVS_CHECKOUT_SCRIPT" | \
+	    mailx -s "$MAIL_SUBJECT" $ERROR_RECIPIENT; 
+	exit 1; 
+    }
+    cd $old_dir
+fi
+
+# A one step copy, run, cleanup!
+# The /bin/kill -9 -$$; line is there to make sure no processes are left behind if the
+# root_checkout.sh looses one... In testing, but appears kill returns non-zero in $?
+# actually it's probably the bash process returning the non-zero, but the next test
+# appears to succeed if we enter _rm_exit(), which is what we want. 
+# We use ssh -x so that during testing FowardX11=no is forced so that zmap_test_suite.sh
+# doesn't try to use the forwarded X11 connection, which will be the user's own not zmap's
+cat $CVS_CHECKOUT_SCRIPT | ssh -x $SRC_MACHINE '/bin/bash -c "\
+function _rm_exit                       \
+{                                       \
+   echo Master Script Failed...;        \
+   rm -f root_checkout.sh;              \
+   /bin/kill -9 -$$;                    \
+   exit 1;                              \
+};                                      \
+cd /var/tmp                || exit 1;   \
+rm -f root_checkout.sh     || exit 1;   \
+cat - > root_checkout.sh   || exit 1;   \
+chmod 755 root_checkout.sh || _rm_exit; \
+: Change the variables in next line             ; \
+./root_checkout.sh RELEASE_LOCATION=~/BUILDS/TODAY ZMAP_MASTER_RT_TO_CVS=no ZMAP_MASTER_BUILD_DOCS=no ZMAP_MASTER_BUILD_DOXYGEN_DOCS=no ZMAP_MASTER_BUILD_DIST=no ZMAP_MASTER_RT_RELEASE_NOTES=no ZMAP_MASTER_DOCS2WEB=no ZMAP_CLUSTER_CONFIG_FILE=~/cluster.config.sh || _rm_exit; \
+:                                               ; \
+rm -f root_checkout.sh     || exit 1;   \
+"' > $GLOBAL_LOG 2>&1
+
+
+# ================== ERROR HANDLING ================== 
+
+if [ $? != 0 ]; then
+    # There was an error, email someone about it!
+    TMP_LOG=/tmp/zmap_fail.$$.log
+    echo "ZMap Build Failed"                              > $TMP_LOG
+    echo ""                                              >> $TMP_LOG
+    echo "Tail of log:"                                  >> $TMP_LOG
+    echo ""                                              >> $TMP_LOG
+    tail $GLOBAL_LOG                                     >> $TMP_LOG
+    echo ""                                              >> $TMP_LOG
+    echo "Full log can be found $(hostname):$GLOBAL_LOG" >> $TMP_LOG
+    if [ "x$ERROR_RECIPIENT" != "x" ]; then
+	cat $TMP_LOG | mailx -s "$MAIL_SUBJECT" $ERROR_RECIPIENT
+    fi
+    rm -f $TMP_LOG
+fi
+
+# ================== END OF SCRIPT ==================