diff options
| -rwxr-xr-x | bin/chkcopyright | 394 |
1 files changed, 332 insertions, 62 deletions
diff --git a/bin/chkcopyright b/bin/chkcopyright index b79b8ff..4347cec 100755 --- a/bin/chkcopyright +++ b/bin/chkcopyright @@ -1,5 +1,6 @@ #! /bin/sh # +# Copyright by The HDF Group. # Copyright by the Board of Trustees of the University of Illinois. # All rights reserved. # @@ -9,8 +10,8 @@ # of the source code distribution tree; Copyright.html can be found at the # root level of an installed copy of the electronic HDF5 document set and # is linked from the top-level documents page. It can also be found at -# http://hdf.ncsa.uiuc.edu/HDF5/doc/Copyright.html. If you do not have -# access to either file, you may request a copy from hdfhelp@ncsa.uiuc.edu. +# http://hdfgroup.org/HDF5/doc/Copyright.html. If you do not have +# access to either file, you may request a copy from help@hdfgroup.org. # # Check Copyright notice. @@ -29,19 +30,25 @@ DIFF="diff" INITFILE=.h5chkright.ini EXCEPTIONS=/tmp/h5chkright.except.$$ tmpfile=/tmp/h5chkright_tmp.$$ -EXCEPTIONDIRS="-name CVS" # at least skip CVS directories. +EXCEPTIONDIRS="-name CVS -o -name .svn" # at least skip CVS directories. EXTRACTEDFILE=/tmp/h5chkright.extracted.$$ VERBOSE= # default no +FIXIT= # default no DIRS=. # default current directory NFAILEDFILES=0 # Number of failed files found. -NPASSEDFILES=0 # Number of failed files found. +NPASSEDFILES=0 # Number of passed files found. +NFIXEDFILES=0 # Number of files fixed. +NFIXFAILEDFILES=0 # Number of files fix failed. NUMBEGINLINES=60 # Copyright notice should be located within the # this number of lines at the beginning of the file. -COPYRIGHTSTR="Copyright by the Board of Trustees of the University of Illinois" +UICOPYRIGHTSTR="Copyright by the Board of Trustees of the University of Illinois" +THGCOPYRIGHTSTR="Copyright by The HDF Group." PASSEDLOG=/tmp/h5chkright_passed.$$ SKIPPEDLOG=/tmp/h5chkright_skipped.$$ FAILEDLOG=/tmp/h5chkright_failed.$$ +FIXEDLOG=/tmp/h5chkright_fixed.$$ +FIXFAILEDLOG=/tmp/h5chkright_fixfailed.$$ C_COPYRIGHT=/tmp/h5chkright_C.$$ # C style copyright FTN_COPYRIGHT=/tmp/h5chkright_FTN.$$ # Fortran style copyright @@ -49,6 +56,9 @@ HTM_COPYRIGHT=/tmp/h5chkright_HTM.$$ # HTML style copyright SH_COPYRIGHT=/tmp/h5chkright_SH.$$ # SHELL style copyright SH_COPYRIGHT2=/tmp/h5chkright_SH2.$$ # SHELL style copyright, 2nd style. WINBAT_COPYRIGHT=/tmp/h5chkright_WINBAT.$$ # Windows Batch file Copyright notice +VMSCMD_COPYRIGHT=/tmp/h5chkright_VMSCMD.$$ # VMS command file Copyright notice + +tmpfixfile=/tmp/h5chkright_fix.$$ # Temporary fixed copy of file # Help page @@ -67,16 +77,31 @@ Usage: $PROGNAME [-h | -help] [-fname name-patter] [-v | -v9] [dir1 dir2 ...] verbose mode -v9 highly verbose + -fix + fix failed files if possible EOF } +# Print Debug output +# +PRINTDEBUG() +{ + if [ -n "$VERBOSE" ]; then + echo $* + else + : # noop + fi +} + + # Generate various styles of Copyright notices # BUILDCOPYRIGHT() { # C and C++ source Copyright notice cat > ${C_COPYRIGHT} << \EOF + * Copyright by The HDF Group. * * Copyright by the Board of Trustees of the University of Illinois. * * All rights reserved. * * * @@ -86,12 +111,13 @@ BUILDCOPYRIGHT() * of the source code distribution tree; Copyright.html can be found at the * * root level of an installed copy of the electronic HDF5 document set and * * is linked from the top-level documents page. It can also be found at * - * http://hdf.ncsa.uiuc.edu/HDF5/doc/Copyright.html. If you do not have * - * access to either file, you may request a copy from hdfhelp@ncsa.uiuc.edu. * + * http://hdfgroup.org/HDF5/doc/Copyright.html. If you do not have * + * access to either file, you may request a copy from help@hdfgroup.org. * EOF # Fortran9X source Copyright notice cat > ${FTN_COPYRIGHT} << \EOF +! Copyright by The HDF Group. * ! Copyright by the Board of Trustees of the University of Illinois. * ! All rights reserved. * ! * @@ -101,12 +127,13 @@ EOF ! of the source code distribution tree; Copyright.html can be found at the * ! root level of an installed copy of the electronic HDF5 document set and * ! is linked from the top-level documents page. It can also be found at * -! http://hdf.ncsa.uiuc.edu/HDF5/doc/Copyright.html. If you do not have * -! access to either file, you may request a copy from hdfhelp@ncsa.uiuc.edu. * +! http://hdfgroup.org/HDF5/doc/Copyright.html. If you do not have * +! access to either file, you may request a copy from help@hdfgroup.org. * EOF # HTML file Copyright notice cat > ${HTM_COPYRIGHT} << \EOF + * Copyright by The HDF Group. * * Copyright by the Board of Trustees of the University of Illinois. * * All rights reserved. * * * @@ -116,12 +143,13 @@ EOF * of the source code distribution tree; Copyright.html can be found at the * * root level of an installed copy of the electronic HDF5 document set and * * is linked from the top-level documents page. It can also be found at * - * http://hdf.ncsa.uiuc.edu/HDF5/doc/Copyright.html. If you do not have * - * access to either file, you may request a copy from hdfhelp@ncsa.uiuc.edu. * + * http://hdfgroup.org/HDF5/doc/Copyright.html. If you do not have * + * access to either file, you may request a copy from help@hdfgroup.org. * EOF # Shell style Copyright notice cat > ${SH_COPYRIGHT} << \EOF +# Copyright by The HDF Group. # Copyright by the Board of Trustees of the University of Illinois. # All rights reserved. # @@ -131,12 +159,13 @@ EOF # of the source code distribution tree; Copyright.html can be found at the # root level of an installed copy of the electronic HDF5 document set and # is linked from the top-level documents page. It can also be found at -# http://hdf.ncsa.uiuc.edu/HDF5/doc/Copyright.html. If you do not have -# access to either file, you may request a copy from hdfhelp@ncsa.uiuc.edu. +# http://hdfgroup.org/HDF5/doc/Copyright.html. If you do not have +# access to either file, you may request a copy from help@hdfgroup.org. EOF # Shell style Copyright notice (2nd type) cat > ${SH_COPYRIGHT2} << \EOF +## Copyright by The HDF Group. ## Copyright by the Board of Trustees of the University of Illinois. ## All rights reserved. ## @@ -146,12 +175,13 @@ EOF ## of the source code distribution tree; Copyright.html can be found at the ## root level of an installed copy of the electronic HDF5 document set and ## is linked from the top-level documents page. It can also be found at -## http://hdf.ncsa.uiuc.edu/HDF5/doc/Copyright.html. If you do not have -## access to either file, you may request a copy from hdfhelp@ncsa.uiuc.edu. +## http://hdfgroup.org/HDF5/doc/Copyright.html. If you do not have +## access to either file, you may request a copy from help@hdfgroup.org. EOF # Windows Batch file Copyright notice cat > ${WINBAT_COPYRIGHT} << \EOF +@REM Copyright by The HDF Group. @REM Copyright by the Board of Trustees of the University of Illinois. @REM All rights reserved. @REM @@ -161,8 +191,24 @@ EOF @REM of the source code distribution tree; Copyright.html can be found at the @REM root level of an installed copy of the electronic HDF5 document set and @REM is linked from the top-level documents page. It can also be found at -@REM http://hdf.ncsa.uiuc.edu/HDF5/doc/Copyright.html. If you do not have -@REM access to either file, you may request a copy from hdfhelp@ncsa.uiuc.edu. +@REM http://hdfgroup.org/HDF5/doc/Copyright.html. If you do not have +@REM access to either file, you may request a copy from help@hdfgroup.org. +EOF + + # VMS command file Copyright notice + cat > ${VMSCMD_COPYRIGHT} << \EOF +$!# Copyright by The HDF Group. +$!# Copyright by the Board of Trustees of the University of Illinois. +$!# All rights reserved. +$!# +$!# This file is part of HDF5. The full HDF5 copyright notice, including +$!# terms governing use, modification, and redistribution, is contained in +$!# the files COPYING and Copyright.html. COPYING can be found at the root +$!# of the source code distribution tree; Copyright.html can be found at the +$!# root level of an installed copy of the electronic HDF5 document set and +$!# is linked from the top-level documents page. It can also be found at +$!# http://hdfgroup.org/HDF5/doc/Copyright.html. If you do not have +$!# access to either file, you may request a copy from help@hdfgroup.org. EOF } @@ -188,7 +234,7 @@ EOF INITIALIZATION() { # clean up log files - rm -f $PASSEDLOG $SKIPPEDLOG $FAILEDLOG + rm -f $PASSEDLOG $SKIPPEDLOG $FAILEDLOG $FIXEDLOG $FIXFAILEDLOG # Generate various styles of copyright notice. BUILDCOPYRIGHT @@ -244,6 +290,9 @@ PARSE_OPTION() shift FNAME="$1" ;; + -fix ) + FIXIT=yes + ;; -v* ) VERBOSE=yes if test X$1 = X-v9; then @@ -277,12 +326,36 @@ RINSE() dos2unix < $tmpfile | expand | sed -e 's/ *$//' > $rf } +# Locate a line in the file and print the line number. +# Print 0 if not found; -1 if error. +# $1 The line. +# $2 The file. +# +FindLineInFile() +{ + if [ $# -ne 2 ]; then + # expect two arguments + echo -1 + return + fi + xline=$1 + xf=$2 + xpos=`grep -n "${xline}" $xf` + if [ "$?" -ne 0 ] ; then + # Not found, return 0 + xpos=0 + else + xpos=`echo $xpos | cut -f1 -d:` + fi + echo $xpos +} + # Match Copyright notice. # $1 file which contains the expected copyright notice. # $2 file in which to look for the copyright notice. # Copyright notice must be found within the beginning $NUMBEGINLINES of lines. -# Hunt for the particular string $COPYRIGHTSTR which signifies the beginning +# Hunt for the particular string $UICOPYRIGHTSTR which signifies the beginning # of the copyright notice. # MATCH_COPYRIGHT() @@ -294,12 +367,23 @@ MATCH_COPYRIGHT() fi COPYRIGHTFILE=$1 f=$2 - # Must use stdin for wc to prevent filename from popping up. - nlines=`wc -l < ${COPYRIGHTFILE}| tr -d ' '` - head -$NUMBEGINLINES < $f | sed -n -e "/${COPYRIGHTSTR}/"',$p' | \ - head -${nlines} > ${EXTRACTEDFILE} - RINSE ${EXTRACTEDFILE} - $DIFF ${EXTRACTEDFILE} ${COPYRIGHTFILE} >/dev/null 2>&1 + nlines=`wc -l ${COPYRIGHTFILE} | cut -f1 -d' '` + # Find a line that contains the copyright string and its line number in + # the file. + begin=`FindLineInFile "${UICOPYRIGHTSTR}" $f` + if [ "$begin" -le 0 ] ; then + # Not found, generate an empty dummy file + cp /dev/null ${EXTRACTEDFILE} + false + else + if [ $begin -gt 1 ]; then + begin=`expr $begin - 1` + fi + end=`expr $begin + $nlines - 1` + sed -n -e "${begin},${end}p" < $f > ${EXTRACTEDFILE} + RINSE ${EXTRACTEDFILE} + $DIFF ${EXTRACTEDFILE} ${COPYRIGHTFILE} >/dev/null 2>&1 + fi if test $? -eq 0; then echo PASSED else @@ -307,6 +391,64 @@ MATCH_COPYRIGHT() fi } +# Fix Copyright notice. +# $1 file which contains the expected copyright notice. +# $2 file in which to look for the copyright notice. +# Copyright notice must be found within the beginning $NUMBEGINLINES of lines. +# Hunt for the particular string $UICOPYRIGHTSTR which signifies the beginning +# of the copyright notice. +# +FIX_COPYRIGHT() +{ + if [ $# -ne 2 ]; then + # expect two arguments + echo FAILED + return + fi + COPYRIGHTFILE=$1 + f=$2 + nlines=`wc -l ${COPYRIGHTFILE} | cut -f1 -d' '` + # If the file has UICOPYRIGHTSTR but not THGCOPYRIGHTSTR, then replace the lines + # starting at UICOPYRIGHTSTR and down. + # If the file has THGCOPYRIGHTSTR, then replace the lines starting at the + # THGCOPYRIGHTSTR and down. + # If neither found, abort it. + + # Find a line that contains the THG copyright string and its line number in + # the file. + insertbegin=`FindLineInFile "${THGCOPYRIGHTSTR}" $f` + if [ $insertbegin -gt 0 ]; then + insertend=`expr $insertbegin + $nlines` # no need to -1. See below. + else + insertbegin=`FindLineInFile "${UICOPYRIGHTSTR}" $f` + if [ $insertbegin -gt 0 ]; then + insertend=`expr $insertbegin + $nlines - 1` # no need to -2. See below. + else + FIXFAILED + return + fi + fi + + # Copy line 1 up to insertbegin from original file + xbegin=`expr $insertbegin - 1` + if [ $xbegin -gt 0 ]; then + sed -n -e "1,${xbegin}p" $f > $tmpfixfile + else + cp /dev/null $tmpfixfile # make it empty. + fi + + # now the correct copyright file + cat $COPYRIGHTFILE >> $tmpfixfile + + # the rest of the original file + sed -n -e "${insertend},"'$p' $f >> $tmpfixfile + + # copy them all back to the original file + cp $tmpfixfile $f + FIXED + rm -f $tmpfixfile +} + # Check C and C++ source files # C_SOURCE() @@ -321,6 +463,9 @@ C_SOURCE() # show the difference FAILED $f $DIFF ${EXTRACTEDFILE} ${C_COPYRIGHT} + if [ -n "$FIXIT" ]; then + FIX_COPYRIGHT $C_COPYRIGHT $f + fi ;; esac } @@ -340,6 +485,9 @@ FORTRAN_SOURCE() # show the difference FAILED $f $DIFF ${EXTRACTEDFILE} ${FTN_COPYRIGHT} + if [ -n "$FIXIT" ]; then + FIX_COPYRIGHT $FTN_COPYRIGHT $f + fi ;; esac } @@ -359,28 +507,57 @@ HTML_FILE() # show the difference FAILED $f $DIFF ${EXTRACTEDFILE} ${HTM_COPYRIGHT} + if [ -n "$FIXIT" ]; then + FIX_COPYRIGHT $HTM_COPYRIGHT $f + fi ;; esac } -# Check Shell script files and other files (e.g., Makefile) that use -# the style of copyright notice of leading #'s. -# There is a preferred style (single leading #) vs the alternate style (double -# leading #'s). The double leading #'s style is used in Makefile.am, for -# example. +# Check Shell script files which use the style of copyright notice of leading #'s. # SHELL_FILE() { f=$1 - if [ `MATCH_COPYRIGHT $SH_COPYRIGHT $f` = FAILED -a \ - `MATCH_COPYRIGHT $SH_COPYRIGHT2 $f` = FAILED ]; then - # show the differences with the preferred style. + case `MATCH_COPYRIGHT $SH_COPYRIGHT $f` in + PASSED) + PASSED $f + return + ;; + FAILED) + # show the difference FAILED $f $DIFF ${EXTRACTEDFILE} ${SH_COPYRIGHT} - else + if [ -n "$FIXIT" ]; then + FIX_COPYRIGHT $SH_COPYRIGHT $f + fi + ;; + esac +} + + +# Check files that use #'s as comments such as Makefile. +# The Copyright body of text happen to be the same as used by Shell script +# files. +# +MAKE_FILE() +{ + f=$1 + case `MATCH_COPYRIGHT $SH_COPYRIGHT $f` in + PASSED) PASSED $f - fi + return + ;; + FAILED) + # show the difference + FAILED $f + $DIFF ${EXTRACTEDFILE} ${SH_COPYRIGHT} + if [ -n "$FIXIT" ]; then + FIX_COPYRIGHT $SH_COPYRIGHT $f + fi + ;; + esac } @@ -398,11 +575,73 @@ BATCH_FILE() # show the difference FAILED $f $DIFF ${EXTRACTEDFILE} ${WINBAT_COPYRIGHT} + if [ -n "$FIXIT" ]; then + FIX_COPYRIGHT $WINBAT_COPYRIGHT $f + fi ;; esac } +# Check Windows Batch files +# +VMSCMD_FILE() +{ + f=$1 + case `MATCH_COPYRIGHT $VMSCMD_COPYRIGHT $f` in + PASSED) + PASSED $f + return + ;; + FAILED) + # show the difference + FAILED $f + $DIFF ${EXTRACTEDFILE} ${VMSCMD_COPYRIGHT} + if [ -n "$FIXIT" ]; then + FIX_COPYRIGHT $VMSCMD_COPYRIGHT $f + fi + ;; + esac +} + + +# Guess the type of file. +# Inspect the first 5 lines to guess what type of file it is. +# +GUESS_File_Type() +{ + if [ $# -ne 1 ]; then + echo "wrong number of arguments($#)" + return + fi + f=$1 + # Now guess the file type. + head -5 < $f > $tmpfile + if head -1 < $tmpfile | grep '^#!' > /dev/null; then + # First line is "#!". It is likely a shell script or similar type. + echo SHELL_FILE + elif grep '\/\*' < $tmpfile > /dev/null; then + # Found some lines containing '/*'. It may be a C/C++ style file. + echo C_SOURCE + elif grep '^!' < $tmpfile > /dev/null; then + # Some lines start with a "!". It may be a Fortran 9X style file. + echo FORTRAN_SOURCE + elif grep '^#' < $tmpfile > /dev/null; then + # Some lines start with a "#". It may be a shell like type. + # Put this after C_SOURCE which may have #define and such lines. + echo SHELL_FILE + elif grep -i '^<html>' < $tmpfile > /dev/null || \ + grep '^<!--' < $tmpfile > /dev/null ; then + # Some lines start with a "<html>" or having an html comment tag. + # It may be an HTML file. + echo HTML_FILE + else + # Unknown type. + echo UNKNOWN_TYPE + fi +} + + # Check Unknown type file. # First check if there is something that resemble a copyright notice in # the first "page". If so, then inspect the first 5 lines to guess what @@ -413,30 +652,16 @@ UNKNOWN_FILE() { f=$1 if head -$NUMBEGINLINES < $f | grep "${COPYRIGHTSTR}" > /dev/null; then - # Now guess the file type and try match it. - head -5 < $f > $tmpfile - if head -1 < $tmpfile | grep '^#!' > /dev/null; then - # First line is "#!". It is likely a shell script or similar type. - SHELL_FILE $f - elif grep '\/\*' < $tmpfile > /dev/null; then - # Found some lines containing '/*'. It may be a C/C++ style file. - C_SOURCE $f - elif grep '^!' < $tmpfile > /dev/null; then - # Some lines start with a "!". It may be a Fortran 9X style file. - FORTRAN_SOURCE $f - elif grep '^#' < $tmpfile > /dev/null; then - # Some lines start with a "#". It may be a shell like type. - # Put this after C_SOURCE which may have #define and such lines. - SHELL_FILE $f - elif grep -i '^<html>' < $tmpfile > /dev/null || \ - grep '^<!--' < $tmpfile > /dev/null ; then - # Some lines start with a "<html>" or having an html comment tag. - # It may be an HTML file. - HTML_FILE $f - else - # Unknown type. - UNKNOWN_TYPE $f - fi + xftype=`GUESS_File_Type $f` + PRINTDEBUG f=$f xftype=$xftype > /dev/tty + case $xftype in + SHELL_FILE) SHELL_FILE $f;; + C_SOURCE) C_SOURCE $f;; + FORTRAN_SOURCE) FORTRAN_SOURCE $f;; + SHELL_FILE) SHELL_FILE $f;; + HTML_FILE) HTML_FILE $f;; + UNKNOWN_TYPE) UNKNOWN_TYPE $f;; + esac else # Unknown type. UNKNOWN_TYPE $f @@ -487,6 +712,28 @@ FAILED() } +# Copyright fixed. +# $1 file that has been fixed. +# +FIXED() +{ + if test X-$VERBOSE = X-yes; then + echo " FIXED" + fi + echo $1 >> $FIXEDLOG +} + + +# Copyright fix failed. +# $1 file that has failed. +# +FIXFAILED() +{ + echo "FIX FAILED: $1" + echo $1 >> $FIXFAILEDLOG +} + + # # Main body @@ -517,13 +764,24 @@ while read file; do *.htm | *.html ) HTML_FILE ${file} ;; - *.sh | *.sh.in | *Makefile | *Makefile.in | *Makefile.am ) + *.sh | *.sh.in ) SHELL_FILE ${file} ;; + *.pl ) + # Perl script files are similar to Shell files + SHELL_FILE ${file} + ;; + *Makefile | *Makefile.in | *Makefile.am | Makefile.dist ) + MAKE_FILE ${file} + ;; *.bat | *.BAT ) # Windows Batch files BATCH_FILE ${file} ;; + *.com ) + # VMS Command files + VMSCMD_FILE ${file} + ;; *.h5 | *.hdf5 ) # Ignore HDF5 data files continue @@ -544,6 +802,10 @@ while read file; do # Ignore CVS control files. continue ;; + *.txt | *.TXT ) + # Ignore text files. + continue + ;; *) UNKNOWN_FILE $file ;; @@ -558,14 +820,20 @@ fi if [ -f $FAILEDLOG ]; then NFAILEDFILES=`wc -l < $FAILEDLOG` fi +if [ -f $FIXEDLOG ]; then + NFIXEDFILES=`wc -l < $FIXEDLOG` +fi +if [ -f $FIXFAILEDLOG ]; then + NFIXFAILEDFILES=`wc -l < $FIXFAILEDLOG` +fi # Cleanup rm -f $C_COPYRIGHT $FTN_COPYRIGHT $HTM_COPYRIGHT $SH_COPYRIGHT $SH_COPYRIGHT2 rm -f $EXCEPTIONS $EXTRACTEDFILE -rm -f $PASSEDLOG $SKIPPEDLOG $FAILEDLOG +rm -f $PASSEDLOG $SKIPPEDLOG $FAILEDLOG $FIXEDLOG $FIXFAILEDLOG # Report Results -# Results are not total accurate--e.g., Passed are not counted, thus not +# Results are not total accurate--e.g., Skipped are not counted, thus not # reported. # echo "*******************" @@ -573,6 +841,8 @@ echo " REPORT" echo "*******************" echo Number of passed files = $NPASSEDFILES echo Number of failed files = $NFAILEDFILES +echo Number of fixed files = $NFIXEDFILES +echo Number of fix failed files = $NFIXFAILEDFILES if [ $NFAILEDFILES -gt 0 ]; then exitcode=1 |