I am trying to write a bash script to scan for authorized_keys files and remove the keys of a couple previous employees if found. I am having one heck of a time figuring out the escaping for the sed command at the end. I am using commas instead of / since / can show up in the ssh-key. Any help would be appreciated
#!/bin/bash
declare -A keys
keys["employee1"]='AAAAB3NzaC1yc2EAAAABJQAAAIEAxoZ7ZdpJkL98n8cSTkFBwaAeSNK0m/tOWtF1mu5NAzMM/+1SDO6rJH/ruyyqBJo9s+AHWZLGRHfXT2XBg2SRaUnubAKp0w6qNIbej0MsA/ifAs8AfVGdj0pUPLtKpo6XVZkB8vEZSIQ+xNk1n5HJrGJnFGWKWeY3z1/KOLxcLHU='
keys["employee2"]='AAAAB3NzaC1yc2EAAAABIwAAAQEAwHYNAVhb319OBVXPhYF8cSTkFBwaAekr7UcKjfLPCHMpz19W0L/C0g+75Hn8COxOQILDUhIPhYHXOduQjGD/6NXgJDWxgyT00Azg5BREUnBd58WqZPlEvTZYlAgmdMIbnWPPGdJwzqKH/k7/STK6vTKxL6rxBo4lSNK0m/tOWtF1mu5NAzMM/+1SDO6rJH/ruyyqBJo9s+NIbej0MsA/ifAs8AfAkfO2JjgeQpJMyZ7B02XVN5iSLAyC3Cb5FXIjJuk4LPhcApuVyszH2lgve0r5bt/nFgVujJTvJTHPlGrqkYDcDJVUtfbjoLqGPrnpijp6rGIC7aFDDe7bk0ygHYMXDFWcjJBerfLGUWTYWFFLY3bfiO/h/9oEycmQHyB2co4a0IyyDnaYn9OY6xsRRATVlk4Q=='
files=`find / -name authorized_keys`
echo "Checking Authorized_Keys files on: " `hostname`
echo ""
echo "Located files: "
for file in $files; do
echo " $file"
done
echo""
for file in $files; do
for key in "${!keys[#]}"; do
if grep -q ${keys[$key]} $file; then
echo " *** Removing $key from $file"
sed "s,${keys[$key]},d" $file
fi
done
done
You've made it a bit complicated I think.
You can do this using grep -vf and process substitution:
# array to hold the value you want to remove
keys=(
'AAAAB3NzaC1yc2EAAAABJQAAAIEAxoZ7ZdpJkL98n8cSTkFBwaAeSNK0m/tOWtF1mu5NAzMM/+1SDO6rJH/ruyyqBJo9s+AHWZLGRHfXT2XBg2SRaUnubAKp0w6qNIbej0MsA/ifAs8AfVGdj0pUPLtKpo6XVZkB8vEZSIQ+xNk1n5HJrGJnFGWKWeY3z1/KOLxcLHU=',
'AAAAB3NzaC1yc2EAAAABIwAAAQEAwHYNAVhb319OBVXPhYF8cSTkFBwaAekr7UcKjfLPCHMpz19W0L/C0g+75Hn8COxOQILDUhIPhYHXOduQjGD/6NXgJDWxgyT00Azg5BREUnBd58WqZPlEvTZYlAgmdMIbnWPPGdJwzqKH/k7/STK6vTKxL6rxBo4lSNK0m/tOWtF1mu5NAzMM/+1SDO6rJH/ruyyqBJo9s+NIbej0MsA/ifAs8AfAkfO2JjgeQpJMyZ7B02XVN5iSLAyC3Cb5FXIjJuk4LPhcApuVyszH2lgve0r5bt/nFgVujJTvJTHPlGrqkYDcDJVUtfbjoLqGPrnpijp6rGIC7aFDDe7bk0ygHYMXDFWcjJBerfLGUWTYWFFLY3bfiO/h/9oEycmQHyB2co4a0IyyDnaYn9OY6xsRRATVlk4Q=='
)
while IFS= read -d '' -r file; do
grep -vf <(printf "%s\n" "${keys[#]}") "$file" > "$file.tmp"
mv "$file.tmp" "$file"
done < <(find / -name authorized_keys -print0)
In your case, it's easy, just need to use a sign which not contained in base64 code as the delimiter, eg |:
sed "\|${keys[$key]}|d" $file
Explanation in the sed manual:
\%regexp%
(The % may be replaced by any other single character.)
This also matches the regular expression regexp, but allows one to use a different delimiter than /.
Related
For files like :
_aaa.txt
_bbb.txt
_ccc.txt
I want to convert them to :
_aAa.txt
_bBb.txt
Any idea how to do this ?
In plain bash, using only shell parameter expansions to perform the conversion:
#!/bin/bash
n=3
for file in *; do
[[ -f $file ]] || continue
suffix=${file:n-1}
mv -i "$file" "${file:0:n-1}${suffix^}"
done
Check the output of the following
#!/bin/bash
for filename in *; do
newname=$(sed 's/./\U&/3' <<< "$filename")
echo "$filename --> $newname"
# mv $filename $newname
done
Then remove the #, if the filename printed is correct
If rename command is available, please try:
rename 's/^(..)(.)/$1\U$2/' *.txt
By default, vim spell checker is code aware, so it doesn't spell-check code parts of the file. In effect, in markdown it considers (pandoc multiline) tables to be codes and thus doesn't spell-check their contents.
Is it possible to override this? Or enable spell-check for the entire file including code.
As far as I'm able to determine, there is no way to tell Vim to ignore the
spellcheck suggestions in the syntax file and to just "check everything".
A fairly heavy-handed workaround is to disable syntax entirely with :syn off;
you can re-enable this with :syn on.
Specifically for Markdown, you can disable highlighting of code blocks with
:syn clear markdownCodeBlock; you can reset this with :syn on as well.
Use syntax spell
:syntax spell toplevel
See:
http://usevim.com/2013/05/10/synspell/
In that case I would contact the maintainer of the markdown syntax file and ask him/she if (s)he could fix this issue.
I created bash script fixing syntax files. IT IS NOT PERFECT BUT IT IS GOOD. It can be reversed by running it again. It adds contains=#Spell to syn match and syn region definitions in all files in given directory.
To use it:
Save the script as fix_syntax_files.sh
Give it permissions
Change path at the bottom of the script to one corresponding to your vim plugins location
Run the script
(OPTIONAL) Run script again to revert the changes
The script makes backup of all files before modification so you can assume it is safe to run it. I anyway do not take any responsibility for potential problems caused by the script.
Edit:
You can leave feedback to the script in the following repository:
https://github.com/dominikduda/config_files/blob/master/bin/fix_vim_syntax_files.sh
#!/bin/bash
function fix_file {
sed -i -e '/exe/! {s/contains=/contains=#Spell,/g}' $1
sed -i -e 's/contains=#Spell,ALL/contains=ALL/g' $1
sed -i -e 's/contains=#Spell,ALLBUT/contains=ALLBUT/g' $1
sed -i -e 's/contains=#Spell,TOP/contains=TOP/g' $1
sed -i -e 's/contains=#Spell,CONTAINED/contains=CONTAINED/g' $1
sed -i -e 's/contains=#Spell,NONE/contains=#Spell/g' $1
sed -i -e '/^ *syn match/ {/contains=/! s/$/ contains=#Spell/g}' $1
sed -i -e '/^ *syn region/ {/contains=/! s/$/ contains=#Spell/g}' $1
return 0
}
function revert_file {
mv "$1/$2.spellfix-backup" "$1/$2"
return 0
}
function fix_recursively_in_catalog {
syntax_catalogs_paths="$(find $1 -type d ! -name '*.*' -not -path '*git*' -print)"
syntax_catalogs_count="$(echo "${syntax_catalogs_paths}" | wc -l)"
echo "${syntax_catalogs_count} syntax catalogs found and will be scanned for files"
echo "${syntax_catalogs_paths}" | while read -r catalog_path ; do
echo " Scanning $catalog_path"
ls -p "${catalog_path}" | grep -v / | grep -v .spellfix-backup | grep .vim | while read -r file_name ; do
cp "${catalog_path}/${file_name}" "${catalog_path}/${file_name}.spellfix-backup"
fix_file "${catalog_path}/${file_name}"
echo " Fixing ${file_name} (backup created as ${file_name}.spellfix-backup)"
done
done
echo 'Fix done.'
echo 'Remember to REVERT FIX before updating vim plugins'
return 0
}
function revert_recursively_in_catalog {
syntax_catalogs_paths="$(find $1 -type d ! -name '*.*' -not -path '*git*' -print)"
syntax_catalogs_count="$(echo "${syntax_catalogs_paths}" | wc -l)"
echo "${syntax_catalogs_count} syntax catalogs found and will be scanned for spellfix-backup files"
echo "${syntax_catalogs_paths}" | while read -r catalog_path ; do
echo " Scanning $catalog_path"
ls -p "${catalog_path}" | grep -v / | grep -v .spellfix-backup | grep .vim | while read -r file_name ; do
revert_file "${catalog_path}" "${file_name}"
echo " Reverting ${file_name} (from file ${file_name}.spellfix-backup)"
done
done
echo 'Revert done.'
echo 'Remember to FIX AGAIN after plugins update (or set it as a post update hook)'
return 0
}
function main {
syntax_catalogs_paths="$(find $1 -type d ! -name '*.*' -not -path '*git*' -print)"
while read -r catalog_path ; do
if ls -p "${catalog_path}" | grep -v / | grep .spellfix-backup; then
echo ".spellfix-backup files found, reverting fix!"
echo "--------------------------------------------"
revert_recursively_in_catalog $1
return 0
fi
done < <(echo "${syntax_catalogs_paths}")
echo ".spellfix-backup files NOT found, fixing!"
echo "-----------------------------------------"
fix_recursively_in_catalog $1
}
main ~/PATH/TO/VIM/PLUGINS/
I apologize if this is a trivial question. I am learning how to use linux bash and this little task is giving me a headache...
So I need to write a script, let's call it count.sh. I want that: for each file in the working directory, prints the filename, the number of lines, and the number of words to the console:
test.txt 100 1023
someOtherfiles 10 233
So far, I know that the following gives me all the files names in the directory. And thanks for all who helped me, I get this working version:
for f in *; do
echo -n "$f"
cat "$f" | wc -wl
done
I would really appreciate your help! Thanks ahead!
P.s. If you know great resources (links for tutorials) for learning about script and you are willing to share it with me. I think I really need to know these basics. Thanks again!
If you must have the file name as the first field in your output, try this:
for f in *; do
if [ -f "$f" ]; then
echo -n "$f"
cat "$f" | wc -wl
fi
done
for f in *; do
if [[ -f $f ]]; then
echo "$f $(wc -wl < "$f")"
fi
done
[[ -f $f ]] processes only files (excludes subdirectories) and also handles the case where the directory is empty (in which case * is (by default) left unexpanded, i.e. assigned to $f as is).
echo "$f $(wc -wl < "$f")" uses command substitution ($( ... )) to directly include the output from the enclosed command in the output string passed to echo.
Note that the reason that < is used to direct the content of file $f to wc via stdin is that wc would otherwise append the name of the input file to its output (thanks, #R Sahu).
I am writing a shell script in which I will read a file and will modify it.
there will be occurrence of some string "ABC_1" in multiple lines.
I need to replace it with "XYZ_1" only when there is "OPQ_3" also present in the line else there should be no modification in the line.
please help how can I do replacement if I read a file liken by line.
for FILE in $FILES
do
echo $FILE
while read line
do
if grep -n "OPQ_3" $line
then
sed -i 's/ABC_1/XYZ_2/'
fi
done < $FILE
done
You can use this sed:
sed -i '/OPQ_3\|OPQ_4/s/ABC_1/XYZ_2/' file
Anubhava has the better answer. Here's how you'd write it in bash
for file in $FILES; do
echo "$file"
tmpfile=$(mktemp)
while IFS= read -r line; do
[[ $line == *OPQ_3* ]] && line=${line/ABC_1/XYZ_2/}
echo "$line"
done < "$file" > "$tmpfile"
mv "$tmpfile" "$file"
done
Note IFS= read -r line is the only way to read a line from stdin exactly, without losing any whitespace or special characters.
Below line has printed in my all php project pages because of malicious attacks.Now think is how i can find and delete this lines using shell script
function_exists('date_default_timezone') ?
date_default_timezone_set('America/Los_Angeles') :
($_REQUEST['c_id']));
I tried with below script but i getting error.I mean to say I not able to match above line with sed commend.Please help me to correct this script..
#!/bin/sh
search='^function_exists\(\'date_default_timezone\'\)\ \?\ date_default_timezone_set\(\'America/Los_Angeles\'\)\ \:\ \(\$_REQUEST\[\'c_id\'\]\)\)\;'
for file in `find /root/test1 -name "*.php"`; do grep "$search" $file &> /dev/null if [ $? -ne 0 ]; then echo "Search string not found in $file!" else sed -i '/$search/d' $file
Try sed with : seperators rather than / since in your pattern America/La conflicts with / ir add a backslash so its America/la
You're not escaping the regex correctly. Try the following:
while IFS= read -r -d '' file; do
if grep -qF "function_exists('date_default_timezone') ? date_default_timezone_set('America/Los_Angeles') : (\$_REQUEST['c_id']));" "$file"
then
sed -i "s|function_exists('date_default_timezone') ? date_default_timezone_set('America/Los_Angeles') : (\$_REQUEST\['c_id'\]));|FOO|g" "$file"
fi
done < <(find /root/test1 -type f -name "*.php" -print0)
This might work for you (GNU sed)
pattern1='function_exists('\''date_default_timezone'\'''
pattern2='.*date_default_timezone_set('\''America\/Los_Angeles'\'') :'
pattern3='.*($_REQUEST\['\''c_id'\''\]));'
sed '/^'"$pattern1"'/{N;N;/^'"$pattern1$pattern2$pattern3"'/d}' file