shell scripting topic

cyperghost

@meleu I think that's just a note to me ... BashPitfalls

function1(){
    local status=$(false)
    echo $?
}

Will return 0 which is obviously wrong
So the return code 0 just indicates the correct setting of a local setted value, which was correctly done ;)

So to get out of this make following

function1(){
    local status
    status=$(false)
    echo $?
}

This will put out correct value for "error" 1

That's all folks

meleu

@cyperghost yeah, that's a thing to be careful. I learned it while reading the RetroPie's Shell Style Guide: https://retropie.org.uk/docs/Shell-Style-Guide/#use-local-variables

meleu

I would like to share a little trick I learned today and also ask for some help...

First the short story

I was needing to check if the current hour is after 18h, then I tried this:

hour=$(date +%H)
if [[ $hour -gt 18 ]]; then
    echo "do something..."
fi

And then I got this error (please, forgive the non-english):

-bash: [[: 08: valor muito grande para esta base de numeração (token de erro é "08")

As you can see, the problem is that date +%H returns 08, and when I try to compare it, bash doesn't see 08 as a decimal number.

The solution is obviously getting rid of that leading zero. I decided that using sed would be overkill for such a simple task, then I've found a pure bash solution using a feature of $(( )).

hour=$(date +%H)
hour=$(( 10#$hour ))  # could also be an oneliner: $(( 10#$(date +%H) ))
if [[ $hour -gt 18 ]]; then
    echo "do something..."
fi

And now my script is working perfectly!

Now the help I mentioned earlier on the beginning of this post...

On that stackoverflow answer I see this:

The $(( )) sets up an arithmetic context and the 10# converts the number from base 10 to base 10 causing any leading zeros to be dropped.

Alright, but I like to see stuff on the official documentation in a hope to learn more tricks. The $(( )) is a bash builtin feature, but in the official documentation there's no mention to the 10# operand.

Any thoughts on where to get info about it?

Sano

@meleu That I can answer !
Actually it's in the very doc you mention :
https://www.gnu.org/software/bash/manual/bash.html#Shell-Arithmetic

Constants with a leading 0 are interpreted as octal numbers. A leading ‘0x’ or ‘0X’ denotes hexadecimal. Otherwise, numbers take the form [base#]n, where the optional base is a decimal number between 2 and 64 representing the arithmetic base, and n is a number in that base. If base# is omitted, then base 10 is used. When specifying n, the digits greater than 9 are represented by the lowercase letters, the uppercase letters, ‘@’, and ‘_’, in that order. If base is less than or equal to 36, lowercase and uppercase letters may be used interchangeably to represent numbers between 10 and 35.

meleu

@sano whoops! Didn't catch those little words 😅

Thanks Sano-san. And yeah, I learned more tricks!

Sano

@meleu Just FYI, you could just have used date +%-H ;)
Glad to see you here again BTW !

meleu

@sano ouch! Looks like I need RTFM some more...

From the date man page:

       By default, date  pads  numeric  fields  with  zeroes.   The  following
       optional flags may follow '%':

       -      (hyphen) do not pad the field

Thanks again, Sanso-sensei!

Sano

@meleu I really deserve no special credit for this, I just remembered something like this existed, probably had to use it in the past, too :)

cyperghost

A common mistake is often made by using grep.
It is often told that grep finds a total of all search strings within a text file due the -c switch. That is total nonsense .... -c just counts lines

So we miss

hello hello hello hello
hello

grep -c hello
2

We use a small hack ;)
Better is to use the -o option is will show occourences of search string listed in newline and now we pipe to wc -l ... and count lines again - now we catched them all ;)

hello hello hello hello
hello

grep -o hello | wc -l
5

cyperghost

@meleu

EDIT BEGIN
I think I found a good solution with pure bash
But maybe some knows a better one ;)

val=${#Unix[@]}
for ((i=0; i<$val; i+=1)); do

EDIT END

Is there something that can be made easier (more efficient) than this script?
I want to merge two arrys
A1=(1 2 3)
A2=(one two there)
the result should be A3
A3=(1 one 2 two 3 three)

It works with this script.
But first... I want to avoid any counters (if possible)
Do you know a nice trick to count up values {0..6} works only with fixed characters not as variable. So {0..6} is fine {0..$arraysize} not

#!/bin/bash
# A small script to show how to merge two arrays
# with alternating values (exp. for creating arrays for dialogs)

# Example Array
Unix=("Debian" "Red hat" "Ubuntu" "Suse" "Fedora" "UTS" "OpenLinux")
Shell=("bash" "csh" "jsh" "rsh" "ksh" "rc" "tcsh")

# Check if both arrays got some size
[[ ${#Unix[@]} -eq ${#Shell[@]} ]] || exit 1
echo "Both arrays got same size -- Proceed"

val=$((${#Unix[@]}-1))
echo $val
for i in $(seq 0 $val); do
    echo "Merging ${Unix[$i]} and ${Shell[$i]}"
    UnixShell+=("${Unix[$i]}" "${Shell[$i]}")
done

echo "${UnixShell[@]}"

meleu

@cyperghost both methods are perfectly valid (using a for to iterate through all items).

cyperghost

@meleu thanks for the feedback -- I appreciate it ;)
Come on ... take a look in the MAME RoW now - if you have time ;)

hiulit

FYI, I'll release a new version of RetroPie Shell Script Boilerplate soon-ish.

This is a template for building shell scripts for RetroPie, with some helper functions, dialogs, etc.

hiulit

Well, here it is! A new version of RetroPie Shell Script Boilerplate.

This is a template for building shell scripts for RetroPie.

I've added examples of the following dialog boxes:

--infobox
--yesno
--msgbox
--menu

All the dialogs can be found at utils/dialogs.sh

I've also moved some functions to utils/base.sh.