source: other-projects/gti/remove_extra_lines_glihelp.sh@ 32022

#! /bin/bash

cp $1 $1.bak

CLEANVERSION_NAME="$1.clean"

# echo $1

# Removing lines that will become completely empty after sed operation
# http://soft.zoneo.net/Linux/remove_empty_lines.php
# http://www.grymoire.com/Unix/Sed.html#uh-30
sed '/ *# -- Missing translation/d' $1 > $CLEANVERSION_NAME
mv $CLEANVERSION_NAME $1

# Now we got rid of all the empty lines created by removing "Missing translation" lines
# We remove any lines containing "Missing translation" but also containing other stuff

cat $1 | sed -e 's/# -- Missing translation:.*$//' > $CLEANVERSION_NAME

# Remove comment of form: #Updated 21-Nov-2011 by <name>
# The date can be either 1 or 2 digits. It need not be followed by "by <name>"
#cat $CLEANVERSION_NAME | sed -e 's/# Updated \d\d-[a-zA-Z]{3}-\d{4} by.*$//' > $1
#cat $CLEANVERSION_NAME | sed -e 's/ *# Updated .* (by.*)?$//' > $1
cat $CLEANVERSION_NAME | sed -e 's/ *# Updated .[0-9]*\-...\-.....*$//' > $1


# Process gli/help/<lang>/help.xml files by removing Missing and Updated xml elements:
# <Updated date="dd-mmm-yyyy by someone"/>
# <!-- Missing translation: \d\d\d -->
cat $1 | sed -e 's@<Updated date=\"[^\"]*\"\/>\s*@@' > $CLEANVERSION_NAME
# remove empty lines introduced into the file
# http://stackoverflow.com/questions/14570489/how-to-remove-blank-lines-from-a-unix-file
# Note: need to escape the plus sign (the regex symbol for one or more chars)
cat $CLEANVERSION_NAME | sed -e 's@<!\-\-\s*Missing translation:\s*[^ ]\+ \+\-\->@@' | sed '/^$/d' > $1


# To test if there were ultimately no relevant changes, work on the temporary *.clean file:
# remove all lines that are merely empty/spaces
# and any lines that are purely comments (though these may start with spaces)
# Then test if the *.clean file is non-empty at the end of it
# double sed operation to remove all comments and then all empty lines:
#   http://stackoverflow.com/questions/22080937/bash-skip-blank-lines-when-iterating-through-file-line-by-line
#cat $1 | sed -e 's/^\s*(#.*)$//' > $CLEANVERSION_NAME
sed -e '/\s*#.*$/d' -e '/^\s*$/d' $1 > $CLEANVERSION_NAME

# if the *.clean file is actually empty, print a message
# http://www.unix.com/shell-programming-and-scripting/21753-check-if-file-empty.html
if [[ ! -s $CLEANVERSION_NAME ]] ; then
    echo "@@@ $1 only has comments and empty lines/spaces."
fi ;
rm $CLEANVERSION_NAME


echo "**** cleaned version $1"
echo "**** pre-cleaned original is in $1.bak"