All we get in Makefiles is just one %/$* [pattern rule](https://www.gnu.org/software/make/manual/html_node/Pattern-Rules.html#Pattern-Rules) pair.

$ cat Makefile
%.bla:; echo $*
$ make -s m.bla
m

How shortsighted of our Unix™ fathers.
How can I achieve something like %{0}.%{1}.bla that will give me

$ make -s a.b.bla
a b

Sure, one could use $(basename), $(suffix), $(subst), but I'm talking about two actual independent $*-like variables.

I'm trying to setup a script that will fetch changes from github where a script has a command that's not executable in Mac OS X. What I'm aiming for is to use sed etc to replace the command within that script. These are the commands: Original command (to be replaced in script): DIR=$(dirname "$(readlink -f $0)") New command (I want to replace with): DIR="$(cd "$(dirname "$0")" && pwd -P)" What I've tried so far: StartStopScript="/path/to/script.sh" DIRnew=""$(cd "$(dirname "$0")" && pwd -P)"" DIRold=""$(dirname "$(readlink -f $0)")"" echo "$StartStopScript" | sed -e 's/"$DIRold"/"$DIRnew"/g' This fails and the commands seems to be executed instead of interpreted as a string. Figure it comes down to quoting in sed. Grateful for any help I can get with this.

For a school project, I am tasked with making my own (simplified) shell, with bash being the reference point. This includes replicating heredoc behavior which was fun until I stumbled upon variable substitution INSIDE of the delimiter, as such:

TESTVAR=hello

followed by cat hello > $TESTVAR hello As demonstrated here, typing the value of the variable manually does not end the heredoc, whereas typing $TESTVAR does. Does this mean that variables are not substituted if they're inside of the delimiter? Disabling variable substitution results in the same thing: cat hello > $TESTVAR hello And getting wild with the quotes also results in the same thing: cat hello > TVAR > $TESTVAR hello TVAR All evidence seems to point towards $TESTVAR does not get substituted **because** it is part of a heredoc delimiter, but I can't find a definitive answer to this! Does anyone know?

If we do: VAR=100:200:300:400 We can do: echo ${VAR%%:*} 100 and echo ${VAR##*:} 400 Are there any equivalents I could use to get the values of 200 and 300? For instance, a way to get only what's between the second and third colon specifically in case there are more than just 3 colons?

For the shell fans (bash) : I want a variable to get a value resulting from an operation output which contains an globbing character (such as *), and the shell expansion always happens, even if I quote or doublequote the expression. To reduce the problem scope, all the three following commands return the current directory file list (which is quite strange for me, as I never met this issue before after 30 years of shell practice !!! Still time to learn) :

-bash
var=*; echo ${var}
var="*"; echo ${var}
var='*'; echo ${var}

but I would like the last command to print simply the * character. I've searched thru the shopt options but didn't find any one to solve the issue. Has anyone an idea?

- On ubuntu 24.04 - GNU bash, version 5.2.21(1)-release (x86_64-pc-linux-gnu) - automake (GNU automake) 1.16.5 - autoconf (GNU Autoconf) 2.71 From [automake doc](https://www.gnu.org/software/autoconf/manual/autoconf-2.71/autoconf.pdf) >${var:+value} Old BSD shells, including the Ultrix sh, don’t accept the colon for any shell substitution, and complain and die. Similarly for ${var:=value}, ${var:?value}, etc. However, all shells that support functions allow the use of colon in shell substitution, and since m4sh requires functions, you can portably use null variable substitution patterns in configure scripts. The substituion does not happen in the following example Given : SOURCE_DATE_EPOCH = 333 src/Makefile.am

bin_PROGRAMS = myproject
myproject_SOURCES = main.c

param6=123
a=${param6:+xyz}
$(info ####### a = $(a)#######)

$(info ####### $(SOURCE_DATE_EPOCH) #######)
$(info ##########  ${SOURCE_DATE_EPOCH:+ -d @${SOURCE_DATE_EPOCH}})

output :

####### a = #######
####### 333 #######
##########

The third info log isn't printing SOURCE_DATE_EPOCH value despite it being defined Is it possible to use variable substitution like ":+" in automake files ?

I'd like to have standard parameters, i.e., my minimum reproducible example is:

#!/usr/bin/zsh
a=${1:a}
printf 'a: "%s"\n' "${a}"
b=${2:./build}
printf 'b: "%s"\n' "${b}"

I'd expect to see, when running ./demo.zsh without arguments:

a: "a"
b: "./build"

Instead I get

a: ""
./demo.zsh:4: bad floating point constant

- why does ${1:a} seem to assume $1 to be set and non-zero, but the $a expanded is empty string? - where does the floating point misparsing come from? If I replace ./build with build, it complains about b being an unknown modifier. If I run ./demo.zsh asdf bar, I get

a: "/tmp/fasf"
./demo.zsh:4: bad floating point constant

which, frankly is even more confusing; where does /tmp/ come from (it's the cwd).

I'm running a series of experiments with bash and want to store the log files in a directory whose name is based on the experiment configuration. Some items in the configuration are boolean (true/false). Take the following configuration as an example:

batch_size=16
fp16=false
bf16=true
checkpoint_activations=true

I'd like to store the log file from the experiment with above configuration as input in a directory with the following name:

output_dir="experiment_bs${batch_size}_dt${fp16 if fp16=true else bf16}_${cp if checkpoint_activations=true else empty}"

Of course, I could declare helper variables:

data_type=""
"${fp16}" && data_type=fp16
"${bf16}" && data_type=bf16
"${cp}" && cp="_cp" || cp=""
output_dir="experiment_bs${batch_size}_dt${data_type}${cp}"

But I find this somewhat clunky and hope that [parameter substitutions](https://tldp.org/LDP/abs/html/parameter-substitution.html#PARAMSUBREF) could be useful here. "${bf16:+bf16}" won't help in my case because this will always print "bf16" regardless of its boolean value as long as it's defined. Are there any parameter substitutions that could be applied to this use-case? Or is there an even better in-line solution to this problem? Note: there's an application-specific reason why I don't directly use data_type in my configuration.

I need to assign a command to a variable and execute it with variable in linux bash. The command executes fine from command line but not from the variable because of multiple quotes. I hope backslash should be used for multiple quotes to avoid the error. Please help with the correct format. **Code:** parm="cat file|awk '$1>0 && $1="abc" {print f $0} {f=$0 ORS}'" eval "$parm" **Contents:** abc 123 efg 456 **Error:** awk: cmd. line:1: >0 && =abc {print f bash} {f=bash ORS} awk: cmd. line:1: ^ syntax error Thanks in advance!

I have a lot of webpages saved in my Download folder and I want to move all html files together with its folder (for every "NAME.html" exists "NAME_files/" folder). The big problem I can't deal is with the space in the name. I thinkered around this and after some attempts I did this:

export OLDIFS="$IFS"
export IFS=$'\n'

then:

for FILE in $(find . -type d -iname "*_files" ) ; do DIR=$(basename ${FILE/_files/}); WEB="${DIR}.html" ;  mv "${DIR}" 000 ;   mv "${WEB}" 000 ; done

I think i've quoted everything, but it doesn't works. Or better, the command above move the html files but I receive the "No such file or directory" for the folders. For example, I have this files : When I do my command only the html file go in 000 folder, but not the folder! Yes, I know, the name, "[FreeFileSync] SyncSettings-Contabilità ⚠️.html" is a very **bad** name... 😅 I tried also with less exotic names with same results. Please help me, where is my mistake?

following example:

string=" 'f o o' 'f oo' 'fo o' "
array=($string)
echo "${array}"

outputs:

'f

while the expected output is:

'f o o'

The only solution I came with is by changing the delimiter to newline (rather than space):

IFS=$'\n'
string="'f o o' # 'f oo' # 'fo o' #"
array=($(echo "$string" | tr '#' '\n'))
echo "${array}"

outputs:

'f o o'

which outputs the expected result, however this requires to set a custom delimiter in string and it's will invoke a new subprocess on each call. I would like to see a better solution than this, thanks you! **Edit:** The above example string should be enough, but since some has asked what the file contents looks like, it's something similar to this:

'magic1' 'some text' 'another simple text' 'some t£xt'
'magic2' 'another line with simple text' 'foo bar' 'yeah another substring'
'magic3' 'so@m{e$%& te+=*&' 'another substring' 'here is the substring'
'magic4' 'why not another line !' 'yeah I like that' 'the substring yeah !!!'

Even more details (as some users requested): there is a custom function which uses grep to return the line from the above file by using a magic substring. for example my find function similar to this:

_find()
{
   string="$(grep "$1" "$file")"
}

then I call it like this:

_find "magic2"
echo "$string"

which outputs:

'magic2' 'another line with simple text' 'foo bar' 'yeah another substring'

P/S: each substring is delimited by space, and each substring are quoted with single quotes. editing the _find is not possible since it's used by other functions. please see the comments for more.

I can remove a pattern in a bash variable using ${variable##pattern} (leading) or ${variable%%pattern} (trailing). But I can't find a bash-only way to keep the pattern and throw the rest. I know there are solutions using sed, awk, or grep, but I want to know if there is a reasonably efficient bash-only solution that I am overlooking? PS: This isn't just an idle question. The initial problem is that I want to process files where the name contains a pattern (technically, '[A-Z]+([A-Z])-[0-9][0-9]+([0-9])': capital letters followed by a dash and digits) and I want to use the same pattern to list the files and extract the matching string for further processing.

Say I have a bash script that takes some environment variables and parameters and substitutes those variables in a script. Is there a tool that can show how the script with the substituted variables will look like? Take this short script for instance. echo user = $USER host = $HOSTNAME The program would output echo user = vfclists host = unixse So the script would be along the lines of expandscript script.sh --param1=xxxx --param2=xxxx --param3=xxxx and in addition to the environment variables filling the values of the params where they are set to the programs variables.

I looked through some of the posted threads and none of them cover my query. I have a simple line of code below which prints all the ASCII characters: echo {' '..'~'} I want to be able to use a variable to substitute this array like below list="{' '..'~'}" and then do the following echo $(print ${list}) To clarify what I am trying to do: passwordLength="2" for (( i = 0; i < $passwordLength; i++ )) do # concatenate random characters samplePasswordRegex="$samplePasswordRegex{' '..'~'}" done echo $^samplePasswordRegex

I recently added a variable to a script process that will be used as a replacement value via sed for an output script (not bash). This is one of many variables set within an if statement and (for reasons I cannot seem to deduce) the variable loses its value once outside of the if statement. This is not occurring with any of the other variables that are set within the if statement. Code and expanded results follow. if [ ${ModeTyp} == "P" ] ; then PC_DB=idpacleanelig PS2_DB=idpaprodserv2 PV_DB=idpaprodviews PV_OUT=idpaprodviews P_DB=idpaprod P_DBIN=idpaprod CERNH="CE_RunHierarchyControl_T" | CECONTROL=CE_LegacyControlTable_RHMv5 echo "CERNH value is ${CERNH} after Default Population in if prod statement" else PC_DB=hfsdevsandbox PS2_DB=hfsdevsandbox PV_DB=idpaprodviews PV_OUT=hfsdevsandbox P_DB=hfsdevsandbox P_DBIN=idpaprod CERNH="CE_RunHierarchyControl_T" | CECONTROL=CE_LegacyControlTable_RHMv5 GOTOLBL_1="" GOTOLBL_3="" echo "CERNH value is ${CERNH} after Default Population in if test statement" fi echo "CERNH value is ${CERNH} after Default Population" RUN RESULTS: + '[' T == P ']' + PC_DB=hfsdevsandbox + PS2_DB=hfsdevsandbox + PV_DB=idpaprodviews + PV_OUT=hfsdevsandbox + P_DB=hfsdevsandbox + P_DBIN=idpaprod + CECONTROL=CE_LegacyControlTable_RHMv5 + CERNH=CE_RunHierarchyControl_T + GOTOLBL_1= + GOTOLBL_3= + echo 'CERNH value is CE_RunHierarchyControl_T after Default Population in if test statement' CERNH value is CE_RunHierarchyControl_T after Default Population in if test statement + echo 'CERNH value is after Default Population' CERNH value is after Default Population

Does Bash have an option to diagnose (and optionally abort execution) "expanded to empty value" variables? Example (hypothetical):

$ bash -c 'echo $x' --xxx
bash: line 1: variable 'x' expanded to empty value

Reason for the question: such an option may be useful when debugging scripts. For example, in some script all variables are expected to expand to a non-empty values. Hence, using an option to detect (and optionally abort execution) "expanded to empty value" variables may be useful for such case. I've already searched for this option in [set builtin](https://www.gnu.org/software/bash/manual/html_node/The-Set-Builtin.html) , but have found nothing.

This question is *not* about how to write a properly escaped string literal. I couldn't find any related question that isn't about how to escape variables for direct consumption within a script or by other programs. My goal is to enable a script to generate other scripts. This is because the tasks in the generated scripts will run anywhere from 0 to *n* times on another machine, and the data from which they are generated may change before they're run (again), so doing the operations directly, over a network will not work. Given a known variable that may contain special characters such as single quotes, I need to write that out as a fully escaped string literal, e.g. a variable foo containing bar'baz should appear in the generated script as: qux='bar'\''baz' which would be written by appending "qux=$foo_esc" to the other lines of script. I did it using Perl like this: foo_esc="'perl -pe 's/('\'')/\\1\\\\\\1\\1/g' <<<"$foo"'" but this seems like overkill. I have had no success in doing it with bash alone. I have tried many variations of these: foo_esc="'${file//\'/\'\\\'\'}'" foo_esc="'${file//\'/'\\''}'" but either extra slashes appear in the output (when I do echo "$foo"), or they cause a syntax error (expecting further input if done from the shell).

I source these variables from a txt file in CentOS bash shell:

NAME01=dns1
HOST01=1.1.1.1
NAME02=dns2
HOST02=1.1.2.2
NAME03=dns3
HOST03=1.1.2.3
## many more lines of similar

NUM=02

How can I use $HOST$NUM to give me 1.1.2.2, such as:

echo $HOST$NUM

I'm seeing an issue with zsh where a glob character within a variable is not expanding as I would expect. The following example does a better job of explaining it. $ echo $0 -bash $ echo $HOME/Downloads/zsh-test/* /Users/bruce/Downloads/zsh-test/file1 /Users/bruce/Downloads/zsh-test/file2 /Users/bruce/Downloads/zsh-test/file3 /Users/bruce/Downloads/zsh-test/file4 $ file=*; echo $HOME/Downloads/zsh-test/$file /Users/bruce/Downloads/zsh-test/file1 /Users/bruce/Downloads/zsh-test/file2 /Users/bruce/Downloads/zsh-test/file3 /Users/bruce/Downloads/zsh-test/file4 Macbook% echo $0 zsh Macbook% echo $HOME/Downloads/zsh-test/* /Users/bruce/Downloads/zsh-test/file1 /Users/bruce/Downloads/zsh-test/file2 /Users/bruce/Downloads/zsh-test/file3 /Users/bruce/Downloads/zsh-test/file4 Macbook% file=*; echo $HOME/Downloads/zsh-test/$file /Users/bruce/Downloads/zsh-test/* I would have expected the last command to expand like it does in bash. Any idea what I'm doing wrong?

Instead of the following printf '%s\t%s\t%s\t%s\n' 'index' 'podcast' 'website' 'YouTube' I want to store the printf output in a Results variable, how can I do this?