The code:
first=true
mountpoint=" "
partlist=`df -h | grep "^/dev"` # get partition info
for i in $partlist # loop through info
do
if [[ ${i:0:1} = "/" ]] # items starting with / are what I'm interested in
then
if [[ $first ]] # The first instance in each pair is the device name
then
mountdev=$i
mountpoint=" "
first=false
else # The second instance is the device mount point
mountpoint=$i
first=true
fi
if [[ $mountpoint != " " ]] # If the mountpoint was just set (last to be set
# before printing config)
then
# Print config
echo "${mountpoint} \${fs_size ${mountpoint}"
echo "USED \${fs_used ${mountpoint}}\${alignr}\${fs_free ${mountpoint}} FREE"
echo "\${fs_bar 3,300 ${mountpoint}}"
echo "READ \${diskio_read ${mountdev}}\${alignr}\${diskio_write ${mountdev}} WRITE"
echo ""
fi
fi
done
The problem:
The goal is to create a script that will generate my preferred conky config for each of my computers without having to edit the file for each computer. The above is a snippet that generates the section which reports on my disk usage information. Unfortunately I'm having trouble having it output ANYTHING. I throw various debug echos into it and it seems to all be working right except the if statement that actually output the config. I can't figure out what's wrong with it though. Any suggestions or assistance?
Advance thank you :)
This is the wrong line:
if [[ $first ]] # The first instance in each pair is the device name
That would always evaluate to true. [[ true ]] or [[ false ]] is synonymous to [[ -n true ]] or [[ -n false ]] which is always correct.
What you probably meant was
if "$first"
Or
if [[ $first == true ]]
Related
I am attempting to use the [[ ]] operator in Bash and failing badly.
My script is:
#!/bin/bash
# Test driver for the ProtectDatae script
set -o xtrace
# [[ -z "$1" ]] || source="$1" && source=""
if [[ -z "%1" ]]
then
source=""
else
source=$1
fi
[[ -z "$2" ]] || target="$2" && target="" # This does not work
[[ -z "$3" ]] || sourceBackup="$3" && sourceBackup="/tmp/sourcebackup" # This does not work
source cmd.sh # Get a function to run a Linux command
if [[ -e "$sourceBackup" && "$sourceBackup" -ne "" ]]
then
# If a source backup directory is specified then get rid of any old directory and make a new backup
# and verify it. If OK, make the source directory be the source backup directory
# otherwise work directly in the source directory
if [[ -e "$sourceBackup" ]]
then
cmd "sudo rm -R $sourceBackup" "empty backup directory $sourceBackup failed"
cmd "cp -apu $source $sourceBackup" "backup home directory"
cmd "diff -aprN ~/$source/* $sourceBackup" "bad backup in $sourceBackup"
source="$sourceBackup"
fi
fi
exit 0
My command invocation is ./TestProtectData.sh "~" /tmp/jfghome /tmp/jfgbackup
The result of xtrace is:
+ source='~'
+ [[ -z /tmp/jfghome ]]
+ target=/tmp/jfghome
+ target=""
+ [[ -z /tmp/jfgbackup ]]
+ sourceBackup=/tmp/jfgbackup
+ sourceBackup=/tmp/sourcebackup
+ source cmd.sh
+ [[ -e /tmp/sourcebackup ]]
+ exit 0
What happens with the following line is the error. Both alternatives appear to be executed and the variable winds up being set incorrectly:
[[ -z "$2" ]] || target="$2" && target=""
I have tried both orders for && and || and they both give the same result with the variable target set to blank. On the next line, a similar thing happens with the variable sourceBackup set to the second alternative, and both alternatives appear to be executed. If I use the if then construct it works. What am I doing wrong?
What am I doing wrong?
Your intended logic doesn't match the bash constructs you're using. This line:
[[ -z "$2" ]] || target="$2" && target="" # This does not work
Breaks down to mean if 2 is not empty set target to $2. If that command succeeds, set target to "". The command to the left of && will always succeed - either the -z test succeeds or the target="$2" succeeds. Thus target="" always runs at the end.
You can use if ... ; then ...; else ...; fi or you can look at these ways to effect a ternary operator in bash, including:
#!/bin/bash -e
[[ -z "$3" ]] && sourceBackup="/tmp/sourcebackup" || sourceBackup="$3"
echo $sourceBackup
% ./t.sh 1 2 3
3
Here, if -z test succeeds we set sourceBackup to the default. If the test fails, $3 is not null and we set sourceBackup to $3.
To reiterate, this is not exactly the same as a tertiary operator. But if you get the order correct, it will work.
A plain assignement (foo=bar) always sets the status code to 0, so after target has been set to $2, it is immediately after set to empty. I would start by turning the logic inside out:
target= # Set to empty string
[[ -z $2 ]] && target=$2
However, this is redundant. You could easier simply just write
target=$2
without doing any [[...]]-test. If the second parameter is empty, target will be empty as well. If it is not empty, target will get that value.
There is one point to consider: In case you decide to turn on set -u, to catch uninitialized variables, a target=$2 would abort the script if there is no second parameter. Therefore, you could also write
target=${2:-}
which tells bash that a missing parameter is OK and should be treated as a null string.
Even though it is redundant, if you do not turn on -u, using ${2:-} shows your intent explicitly, and makes your program more maintainable.
I know there is a way to do it but I can't seem to remember now.
I have a small shell script which has a simple while loop and it iterates through a CSV file and has an if else loop inside it. The problem is for every line the while loop iterates the if else statement prints a no match found until it finds a match or it keeps printing no match until end of file is reached. I wanted to check how can I print no-match found only once after iterating through the entire file. Here is my sample code below
while IFS=, read -r path servicename app text1 text2 text3;
do
text1="$key1"
text2="$key2"
text3="$key3"
if [[ $servicename = $SERVICE_NAME ]];
then
if [[ $ENV = "sb" ]]; then
ID="$key1"
elif [[ $ENV = "lb" ]]; then
ID="$key2"
elif [[ $ENV = "dev" ]]; then
ID="$key3"
fi
else
echo "no match found"
fi
done < $input_file
sample data that is used in the iteration
path,service,app,text1,text2,text3
path/of/service1,servicename1,appname,key1,key2,key3
path/of/service2,servicename2,appname,key1,key2,key3
path/of/service3,servicename3,appname,key1,key2,key3
path/of/service4,servicename4,appname,key1,key2,key3
path/of/service5,servicename5,appname,key1,key2,key3
Use a variable that keeps track if a match was found or not, and display the message at the end only.
Like this:
#!/bin/bash
SERVICE_NAME="blablabla"
ENV="blablabla"
input_file="input.txt"
match_found="no"
while IFS=, read -r path servicename app text1 text2 text3
do
text1="$key1"
text2="$key2"
text3="$key3"
if [[ "$servicename" == "$SERVICE_NAME" ]]
then
match_found="yes"
if [[ "$ENV" == "sb" ]]
then
ID="$key1"
elif [[ "$ENV" == "lb" ]]
then
ID="$key2"
elif [[ "$ENV" == "dev" ]]
then
ID="$key3"
fi
fi
done < "$input_file"
if [[ "$match_found" == "no" ]]
then
printf "No match found\n"
fi
Another couple modifications:
used == in the if conditions to check if text matches
all variables use is double quoted
I had to set a couple global variables to some bogus values to test my answer. When you ask a question, you should provide a complete minimal sample.
When I execute below script it works fine:
if [[ "[1,2,3]" =~ .*1.* ]]; then
techStatus=1
else
techStatus=0;
fi
echo $techStatus
Output is 1
But when we changes it to variable it does not work.
var1=[1,2,3]
var2=1
if [[ "$var1" =~ .*"$var2".* ]]; then
techStatus=1
else
techStatus=0;
fi
echo $techStatus
Output is 0.
Please help me figure out what is wrong here.
A better & readable approach would be to convert var1 to array and loop through var1.
var1=(1 2 3)
var2=1
for elem in "${var1[#]}"; do
if [[ "$elem" -eq "$var2" ]]; then
techStatus=1
break
else
techStatus=0
fi
done
echo "$techStatus"
I would like to ask you and request for the help with this particular part of the script - I would like to make it more compact as I feel it is too long, or there might me a shorter syntax alternatives...
This part is cut from my script for creation of X amount of files with descending date (1 day jumps, so each file is 1 day older that previous), which is already 3x longer than the whole "implementation" part.
This part's purpose is to maintain, check and correct (appropriately) the arguments defined with the script execution, which looks like
./script.sh X X
whereas script expects two arguments (number indicating amount of files to create and the path to the folder, where it is supposed to create those0:
- if there is just 1 it assumes the user wants to create files in the present folder, therefore it checks, if the one included argument is a number and if it is it will add output from pwd to the path (in this case to not have some sort of "cross-mount" accident with the system variable PATH conveniently named PATHX), so there will be two arguments at the end anyway
- if there are more than 2, it will terminate the script
If there are 2 arguments, it performs additional "tasks:
1. checks if there are just two numbers, if so, script ends
2. checks if there are just two words/letters, if so, script ends
3. if there are, let's say, just two dots as arguments, it will end anyway as there is one if, which always will need one argument to be just number
4. after first steps this will save the arguments to the variables (it is necessary as I have had some problems in the implementation part), this step even coves any possible combination of the positions of the arguments (it does not matter now, if the amount is first or second; same for the path)
5. this is a very last step, which just (for any case) covers the case, the path included has not been inserted with the slash in front (as this is mandatory for bash to recognise it indicates an absolute path) or at the end (important for the touch command in the implementation path as the command looks like "touch -t *TIMESTAMP* "$PATHXfile$i""); together with it it does provide an appropriate action, if the path is defined with ".","..","./","../" - if there would be a full path, not relative one
if [[ $# -eq 2 ]]
then
if [[ $1 == ?(-)+([0-9]) ]] && [[ $2 == ?(-)+([0-9]) ]]
then
echo "Invalid arguments. Did you include a slash at the start of the path (if the file name consists only of the numbers)? Well, try it again. Terminating script..."
exit
elif [[ $1 == ?(-)+([a-z]) ]] && [[ $2 == ?(-)+([a-z]) ]]
then
echo "Only characters in the arguments. Is this some sort of a joke? If you have tried some sick hexadecimal format of number (just A-F), it will not work, mate. Terminating script..."
exit
fi
if [[ $1 == ?(-)+([0-9]) ]]
then
AMOUNT=$1
PATHX=$2
else
AMOUNT=$2
if [[ $AMOUNT != ?(-)+([0-9]) ]]
then
echo "No argument with number/numbers-only as an amount of files to create. Terminating script..."
exit
fi
PATHX=$1
fi
else
if [[ $# -eq 1 ]]
then
if [[ $1 == ?(-)+([0-9]) ]]
then
AMOUNT=$1
PATHX=$(pwd)
else
echo "Please, include the argument with an amount of files to create. Terminating script..."
exit
fi
else
echo "2 Arguments are expected. Terminating script..."
exit
fi
fi
if [[ $PATHX != /* ]]
then
if [[ "$PATHX" == "." ]] || [[ "$PATHX" == ".." ]] || [[ "$PATHX" == "./"* ]] || [[ "$PATHX" == "../"* ]]
then
true
else
PATHX=$(echo "/$PATHX")
fi
fi
if [[ $PATHX != */ ]]
then
PATHX=$(echo "$PATHX/")
fi
Excuse this formatting, but without adding the description to the code sample it is just a mess of text...
Anyway, thank you, Guys, for any input.
It may be reduced to:
[[ $# -gt 2 ]] && { echo "Incorrect number of arguments"; exit 1; }
[[ $1 == ?(-)+([0-9]) ]] || { echo "First argument is not a number"; exit 2; }
[[ $2 == ?(-)+([0-9]) ]] && { echo "Both arguments are numbers"; exit 3; }
[[ $2 == +([a-z]) ]] || { echo "File name must be only letters"; exit 4; }
[[ ! -d $2 ]] && { echo "Path does not exist"; exit 5; }
n=$1
p=${2:-$PWD} # Use the PWD if path is missing.
if [[ $p != */ ]]; then
echo "Missing trailing slash; correcting";
p+=/
fi
if [[ $p != /* ]]; then
echo "Missing leading slash; correcting";
[[ $p == #(.|..|./*|../*) ]] || p=/"$p"
fi
Aside from running every code path that has an NSLocalizedString in it, is there a way to verify that all NSLocalizedStrings have a key that actually exists in all your Localizable.strings files of all your bundles?
E.g. there wasn't a typo in one key such that NSLocalizedString won't find the key it's looking for?
OK I wrote a bash script to accomplish the above. Here it is. It took me hours so don't forget to up-vote me if you like. Feel free to make improvements, etc. I added a few comments suggesting potential improvements.
#!/bin/sh
# VerNSLocalizedStrings
while getopts "vsl:" arg; do
case $arg in
v)
verbose="yes"
;;
s)
stopOnMissing="yes"
;;
l)
lang=$OPTARG
;;
esac
done
if [[ -z $lang ]]
then
lang="en"
fi
searchDir=$lang.lproj
fileFound=`ls . | grep $searchDir`
if [[ -z $fileFound ]]
then
echo "dir "$searchDir" not found."
exit
fi
fileFound=`ls $searchDir/ | grep strings`
if [[ -z $fileFound ]]
then
echo "No .strings files found in dir "$searchDir"."
exit
fi
echo "Verifying NSLocalizationStrings in "$searchDir
# Get all the NSLocalizedString Commands
output=$(grep -R NSLocalizedString . --include="*.m")
# Go thru the NSLocalizedString commands line for line
count=$(( 0 ))
missing=$(( 0 ))
unusable=$(( 0 ))
OIFS="${IFS}"
NIFS=$'\n'
IFS="${NIFS}"
for LINE in ${output} ; do
IFS="${OIFS}"
# Now extract the key from it
# admittedly this only works if there are no line breaks between
# NSLocalizedStrings and the entire key,
# but it accounts for the keys it couldn't identify.
quotes=`echo $LINE | awk -F\" '{ for(i=2; i<=NF; i=i+2){ a = a"\""$i"\"""^";} {print a; a="";}}'`
key=`echo $quotes | cut -f1 -d"^"`
# If we couldn't find the key then flag problem
if [[ -z $key ]]
then
(( unusable += 1 ))
echo "Couldn't extract key: " $LINE
if [ -n "$stopOnMissing" ]
then
break
else
continue
fi
fi
# I don't know how grep works regarding length of string, only that
# if the string is too long then it doesn't find it in the file
keyLength=$(echo ${#key})
if [ $keyLength -gt 79 ]
then
(( unusable += 1 ))
echo "Key too long ("$keyLength"): " $key
if [ -n "$stopOnMissing" ]
then
break
else
continue
fi
fi
# It would be nice if this were a regular expression that allowed as many
# spaces as you want, even a line break then forced the quotes on the
# other side of the equal sign.
keyString=$key" ="
# Search for the key
found=$(iconv -sc -f utf-16 -t utf8 $searchDir/*.strings | grep "$keyString")
# damned if I know why some strings files are utf-16 and others are utf8
if [[ -z $found ]]
then
found=$(grep -r "$keyString" $searchDir/ --include=*.strings)
fi
# analyze the result
if [[ -z $found ]]
then
(( missing += 1 ))
echo "Missing: " $key "\n from: " $LINE
if [ -n "$stopOnMissing" ]
then
break
fi
else
if [ -n "$verbose" ]
then
echo "found: " $key
fi
fi
(( count += 1 ))
IFS="${NIFS}"
done
IFS="${OIFS}"
# It would also be nice if it went the other way and identified
# extraneous unused items in the strings files. But
# I've spent enough time on this for now
echo $count " keys analyzed"
echo $unusable " keys could not be determined"
echo $missing " keys missing"
To verify that all NSLocalizedStrings have a key that actually exists in all your Localizable.strings files or you missed localised you can enable the Localization enable "Show non-localized strings" option in the your project scheme editor.
Now run application you will see console logs for the missing localised string.