move sorting into check.sh, log to file, add summary

1 files changed, 44 insertions, 18 deletions

diff --git a/check.sh b/check.sh
index 48a8f8a..7de5a92 100755
--- a/check.sh
+++ b/check.sh
@@ -12,35 +12,61 @@ readonly REF_REGEX='^[^:]*:[^:]*::[^:]*:.*$'
 readonly SYNTAX_REGEX='^[^:]*:[^:]*:(sv|debian|parabola|fsf|fedora)?:[^:]*:.*$'
 readonly CSV_CHAR=':'
 readonly SEP_CHAR='!'
+readonly LOG_FILE=./check.log ; rm ${LOG_FILE} 2> /dev/null
 
 exit_status=0
 
-printf "\n\nchecking for entries without reference to detailed description: ... " >&2
-unsourced="$(egrep ${REF_REGEX} *.txt)"
-if   [[ -z "$unsourced" ]]
-then printf "OK\n" >&2
-else printf "\n[citation needed]:\n\n%s\n\n" "$unsourced" >&2
-     echo "entries needing citation: $(grep -c '\n' <<<${unsourced})"
-     exit_status=1
+
+# TODO: the best sorting results are acheived when the field separator ($CSV_CHAR)
+#         precedes any valid package name character in ASCII order
+#       the lowest of which is ASCII 43 '+', and spaces are not allowed;
+#         so ASCII 33 ('!') serves this purpose quite well
+#       someday, we should re-write the tools to use parse on '!' instead of ':'
+#       if that were done, then the `sort` command alone would yeild
+#         the same results as this procedure, except for removing empty lines
+unsortable="$(
+  for blacklist in *.txt
+  do  echo -n "sorting and cleaning: '${blacklist}' ... " >> ${LOG_FILE}
+      if   grep ${SEP_CHAR} ${blacklist}
+      then echo "ERROR: can not sort - contains '${SEP_CHAR}' char" >> ${LOG_FILE}
+           retval=1
+      else echo "OK" >> ${LOG_FILE}
+           cat ${blacklist}        | tr "${CSV_CHAR}" "${SEP_CHAR}" | sort | uniq |     \
+           sed '/^[[:space:]]*$/d' | tr "${SEP_CHAR}" "${CSV_CHAR}" > ${blacklist}.temp
+           mv ${blacklist}.temp ${blacklist}
+      fi
+  done
+)"
+if   [[ -n "$unsortable" ]]
+then printf "\n[Entries containing '%s' char]:\n\n%s\n\n" "${SEP_CHAR}" "$unsortable" >> ${LOG_FILE}
+     echo -n "ERROR: one of the data files is unsortable - check can not continue"
+     echo " - correct the malformed entries, then run this script again"
+     exit 1
 fi
 
-printf "\n\nchecking for entries with syntax errors: ... " >&2
+printf "\n\nchecking for entries with syntax errors: ... " >> ${LOG_FILE}
 invalid="$(egrep -v ${SYNTAX_REGEX} *.txt)"
 if   [[ -z "$invalid" ]]
-then printf "OK\n" >&2
-else printf "\n[Incorrectly formatted entries]:\n\n%s\n\n" "$invalid" >&2
-     echo "entries improperly formatted: $(grep -c '\n' <<<${invalid})"
+then printf "OK\n" >> ${LOG_FILE}
+else printf "\n[Incorrectly formatted entries]:\n\n%s\n\n" "$invalid" >> ${LOG_FILE}
      exit_status=1
 fi
 
-# TODO: this check could be removed someday - see note in 'sort-entries' script
-printf "\n\nchecking for entries that the 'sort-entries' script would mutate: ... " >&2
-unsortable="$(grep ${SEP_CHAR} *.txt)"
-if   [[ -z "$unsortable" ]]
-then printf "OK\n" >&2
-else printf "\n[Entries containing '%s' char]:\n\n%s\n\n" "${SEP_CHAR}" "$unsortable" >&2
-     echo "entries containing $SEP_CHAR('${SEP_CHAR}'): $(grep -c '\n' <<<${unsortable})"
+printf "\n\nchecking for entries without reference to detailed description: ... " >> ${LOG_FILE}
+unsourced="$(egrep ${REF_REGEX} *.txt)"
+if   [[ -z "$unsourced" ]]
+then printf "OK\n" >> ${LOG_FILE}
+else printf "\n[citation needed]:\n\n%s\n\n" "$unsourced" >> ${LOG_FILE}
      exit_status=1
 fi
 
+# summary
+totals=$(wc -l *.txt | sed 's|\(.*\)|\t\1|')
+n_unsourced=$(wc -l <<<${unsourced})
+n_malformed=$(wc -l <<<${invalid}  )
+echo -e "summary:\n\t* number of entries total:\n${totals}"
+(( ${n_malformed} )) && echo -e "\t* number of entries improperly formatted: ${n_malformed}"
+(( ${n_unsourced} )) && echo -e "\t* number of entries needing citation: ${n_unsourced}"
+(( ${exit_status} )) && echo "refer to the file: '${LOG_FILE}' for details"
+
 exit $exit_status