[TASK] (util): recode csbucscheck.sh

improve implementation of csbucscheck.sh by
- providing proper command line reading
- providing proper usage information
- providing comments in the code
- providing additional command line options to select an output directory and
  checksum binary executable

util/README

@@ -25,6 +25,8 @@ csbucomparecsfiles.sh
 
 csbucscheck.sh
   Recurse into a read-only subdirectory tree and verify checksums.
+  This provides a complete and independent verification of the consistency,
+  completeness and correctes of csback supervised data archive.
 
 csbusortdirlist.py
   Collect all directories being checksum-supervised in wiki formatted tables.

util/csbucscheck.sh

@@ -18,76 +18,225 @@
 #    29/01/2014   V1.2   Christopher Sanz 
 #                        remove link to script cscomp.sh and include it in
 #                        <csback_cscheck.sh>
+#    13/06/2016   V1.3   Thomas Forbriger
+#                        implement different type of checksums
 #
 # ============================================================================
 #
+VERSION="13/06/2016   V1.3"
+
+# checksum file basename
+CHECKSUMFILEBASE="checksumfile"
+# name of csback checksum files
+CHECKSUMFILE="${CHECKSUMFILEBASE}.cs"
+
+# current working directory
+HERE=$(pwd)
+
+# base name of temporary directory used for intermediate lists and for report
+# files
+TMPBASE=$HERE/report
+
+# default checksum binary executable
 SHABIN=/usr/bin/sha512sum
-TMPDIR=$HOME/tmp/$0.$$
 
-if test $# -lt 1
+# ============================================================================
+# shell functions
+# ===============
+
+# usage function
+function usage {
+cat << HERE
+This is <csbucscheck.sh>
+version: $VERSION
+
+usage: csbucscheck.sh [-v] [-d] [-r dir] dir 
+or:    csbucscheck.sh -h
+
+Recurse into a csback maintained data directory tree and check all files
+against their checksums and test checksum files for completeness.
+HERE
+}
+
+# ----------------------------------------------------------------------------
+# usage function
+function longusage {
+cat << HERE
+
+  -h        print detailed usage information
+  -v        be verbose
+  -d        print debug output
+  -r dir    base directory for report
+            if not provided, a subdirectory in the current directory will
+            be created; the default is
+            $TMPBASE
+  -b bin    binary executable to be used to verify checksums
+            default: $SHABIN
+
+  dir       root directory of the csback supervised tree to be checked
+
+  The name of checksumfiles to be searched for is: $CHECKSUMFILE
+HERE
+}
+
+# ============================================================================
+# read command line
+# =================
+
+verbose=0
+debug=0
+
+# fetch commandline arguments
+while getopts 'hvdr:b:' Option;
+do
+  case ${Option} in
+    h) hflag=1
+      usage; longusage; exit 0;;
+    v) verbose=1;;
+    d) debug=1;;
+    r) TMPBASE=$1; shift;;
+    b) SHABIN=$1; shift;;
+    *) echo -e "ERROR: Unimplemented option chosen.\n"
+      usage; exit 2;;
+  esac
+done
+shift $((${OPTIND} - 1))
+
+MINPARAMS=1
+# check number of commandline arguments
+if [ $# -lt ${MINPARAMS} ]
+then
+  usage 
+  exit 2
+fi
+
+TMPBASEDIR=$(dirname $TMPBASE)
+if test ! ( -d $(dirname $TMPBASE) -a -w $(dirname $TMPBASE) )
 then
-  echo $0 takes a path name as a single argument
-  echo the path must point to the root of csback controlled data structure
+  echo report output base directory $TMPBASEDIR is not a directory
+  echo or is not writable
   exit 2
 fi
 
 DATADIR=$1
+TMPDIR=$TMPBASE/report_$(basename $0 .sh)_$(basename $DATADIR)_$$
+mkdir -p $TMPDIR
+
+if test ! ( -d $TMPDIR -a -w $TMPDIR )
+then
+  echo report output directory $TMPDIR is not a directory
+  echo or is not writable
+  exit 2
+fi
+
+if test ! -x $SHABIN
+then
+  echo $SHABIN is not executable
+  exit2
+fi
+
+# ============================================================================
+# start action
+# ============
+
 LOGLIST=
-HERE=$(pwd)
 nmcs=0 # counter for missing checksums
 nmf=0 # counter for missing files
-mkdir -p $TMPDIR
 
-for FILE in $(find $DATADIR -name checksumfile.cs)
+# ----------------------------------------------------------------------------
+# cycle through all checksum files
+# --------------------------------
+for FILE in $(find $DATADIR -name $CHECKSUMFILE)
 do
   echo check $FILE
   
   DIR=${FILE%/*}
-  cat $DIR/checksumfile.cs | cut -d' ' -f2 > $TMPDIR/chksm.tmp # write only filenames from checksumfile.cs in file
-  NLINES1=$(wc -l $TMPDIR/chksm.tmp | cut -d' ' -f1) # determine number of lines of chksm.tmp
-  #test1############## test if there is an entry in checksumfile.cs for every file
-  find $DIR -maxdepth 1 -type f -printf "%f\n" > $TMPDIR/lsfile.tmp #write the files of the directory in this file
-  NLINES2=$(wc -l $TMPDIR/lsfile.tmp | cut -d' ' -f1) # determine number of lines of lsfile.tmp
+
+  # extract all file names from checksum file
+  cat $DIR/$CHECKSUMFILE | cut -d' ' -f2 > $TMPDIR/chksm.tmp 
+
+  # determine number of files found in checksum file
+  NLINES1=$(wc -l $TMPDIR/chksm.tmp | cut -d' ' -f1)
+
+  # test 1
+  # ------
+  # test if there is an entry in checksumfile.cs for every file in the
+  # directory
+
+  # collect all names of files in the current directory
+  find $DIR -maxdepth 1 -type f -printf "%f\n" > $TMPDIR/lsfile.tmp 
+   
+  # determine number of lines of lsfile.tmp
+  NLINES2=$(wc -l $TMPDIR/lsfile.tmp | cut -d' ' -f1) 
+
+  # cycle through list filename by filename
   for ((line=1;line<=$NLINES2;line++))
   do
+    # extract filename with current index
     FILECHECK=$(sed -ne ''$line'p' $TMPDIR/lsfile.tmp)
+    # ssek file name in corresponding list of file names in checksum file
     CHECK=$(cat $TMPDIR/chksm.tmp | grep -x $FILECHECK )
-    if [[ $CHECK != $FILECHECK ]] && [[ $FILECHECK != checksumfile.* ]]  # test if there is an entry in checksumfile.cs for every file
+    
+    # report missing entry, if not found in checksum file
+    #
+    # the checksum entry is considered as missing, if no corresponding
+    # filename was found and if the file pattern does not match the csback
+    # report file name pattern
+    if [[ "$CHECK" != "$FILECHECK" ]] && [[ $FILECHECK != ${CHECKSUMFILEBASE}.* ]]  
       then
-	  echo $DIR/$FILECHECK has no entry in $DIR/checksumfile.cs
+	  echo $DIR/$FILECHECK has no entry in $DIR/$CHECKSUMFILE
 	  echo $DIR/$FILECHECK >> $TMPDIR/missingCS.log
-	  echo $DIR/$FILECHECK has no entry in $DIR/checksumfile.cs >> $TMPDIR/cscompVerbose.log
+	  echo $DIR/$FILECHECK has no entry in $DIR/$CHECKSUMFILE \
+          >> $TMPDIR/cscompVerbose.log
 	  nmcs=$((nmcs+1))
     fi
     nc=$((nc+1))    
   done
+
+  # summary report
   echo $nc files checked
   echo $DIR checked, total $nc files checked  >> $TMPDIR/cscompVerbose.log
   unset  NLINES2 CHECK FILECHECK DIR
-  #end#of#test1#######
+  # ---- end of test 1 ----
   
+  # test 2
+  # ------
+  # check file contents against checksum
+
+  # cycle through all entries in checksum file and verify checksum
   NLINES=$(wc -l $FILE | cut -f 1 -d ' ')
   if test $NLINES -gt 0
   then
-    echo "  check $NLINES entries"
+    echo "  verify $NLINES entries in $DIR/$CHECKSUMFILE"
     cd $(dirname $FILE)
     LOGFILE=$TMPDIR/$(echo $FILE | tr '/' '_')
     cat $(basename $FILE) | tr -s ' ' | cut -f 1,2 -d ' ' \
       | sed -e 's/ /  /' | $SHABIN -c > $LOGFILE
     LOGLIST+=" "$LOGFILE
   else
-    echo "  checksum file is empty"
+    echo "  checksum file $DIR/$CHECKSUMFILE is empty"
   fi
   cd $HERE
 done
 
+# ----------------------------------------------------------------------------
+# collect and evaluate reports
+# ----------------------------
+
+# delete intermediate files
 rm $TMPDIR/lsfile.tmp $TMPDIR/chksm.tmp
+
+# report checksum verification errors
 echo "deviant results:"
 egrep -v 'OK$' $LOGLIST
 egrep -v 'OK$' $LOGLIST > $TMPDIR/deviant.log
+
+# report missing file names in checksum files
 echo $nmcs missing checksums in checksumfiles >> $TMPDIR/missingCS.log
 echo list of missing checksums: 
 cat $TMPDIR/missingCS.log
+
+# report statistics and report location
 echo "number of checked entries:" $(cat $LOGLIST | wc -l)
 echo results are archived in $TMPDIR