Let's suppose that you have a super python 2&3 one-liner that you need to use in a shell script. On system A the python command works, but on system B you have to use python3, on system C you need to use python3.12 or python2, etc... What would be a sensible way to check some, if not every, "python*" command name? I want to avoid using a construct like this:

#!/bin/bash

if command -v python
then
    python_exe=python
elif command -v python3
then
    python_exe=python3
elif command -v python2
then
     python_exe=python2
else
    echo "python executable not found" >&2
    exit 1
fi > /dev/null

"$python_exe" -c 'print("Hello world!");'

I need to setup a portable kali-linux environment that I can boot on any available computer. I don't want to use a virtual environment because I would need to download VMware on the host computer, plus it's not practical since I will need to boot my OS on friends/customers computers. I just want to boot it from the BIOS, do what I have to do and leave without changing or downloading any files/software on machines that aren't mine. I want to leave the hosts like I was never there after I'm done. Of course, I need data persistence so that I can access, change and keep my datas across reboots on different machines. So, if I'm not wrong, what I need is: Adding Persistence to a Kali Linux Live USB Drive Making a Kali Bootable USB Drive (Windows) I have two questions: - Wouldn't something like that run too slow on a classic USB flash drive ? - I kind of want to use a portable SSD for this. I found one that I like and I'd like your opinions on it and if what I want to do with it is possible. Since this SSD has built-in features and software, is it going to be a problem to make it a kali-linux portable environment? For example, what if I need to format the disk to a specific file system type?

On Linux the included program returns from select and exits: $ gcc -Wall -Wextra select_test.c -o select_test $ ./select_test reading from read end closing write end first read returned 0 second read returned 0 selecting with read fd in fdset select returned On OS X, the select blocks forever and the program does not exit. The Linux behavior matches my expectation and appears to conform to the following bit of the POSIX manual page for select: > A descriptor shall be considered ready for reading when a call to an input function with O_NONBLOCK clear would not block, whether or not the function would transfer data successfully. (The function might return data, an end-of-file indication, or an error other than one indicating that it is blocked, and in each of these cases the descriptor shall be considered ready for reading.) Since read(2) on the read end of the fifo will always return EOF, my reading says that it should always be considered ready by select. Is macOS's behavior here well-known or expected? Is there something else in this example that leads to the behavior difference? A further note is that if I remove the read calls then macOS's select returns. This and some other experiments seem to indicate that once an EOF has been read from the file, it will no longer be marked as ready if select is called on it later. ## Example Program #include #include #include #include #include #include #include #define FILENAME "select_test_tmp.fifo" int main() { pid_t pid; int r_fd, w_fd; unsigned char buffer; fd_set readfds; mkfifo(FILENAME, S_IRWXU); pid = fork(); if (pid == -1) { perror("fork"); exit(1); } if (pid == 0) { w_fd = open(FILENAME, O_WRONLY); if (w_fd == -1) { perror("open"); exit(1); } printf("closing write end\n"); close(w_fd); exit(0); } r_fd = open(FILENAME, O_RDONLY); if (r_fd == -1) { perror("open"); exit(1); } printf("reading from read end\n"); if (read(r_fd, &buffer, 10) == 0) { printf("first read returned 0\n"); } else { printf("first read returned non-zero\n"); } if (read(r_fd, &buffer, 10) == 0) { printf("second read returned 0\n"); } else { printf("second read returned non-zero\n"); } FD_ZERO(&readfds); FD_SET(r_fd, &readfds); printf("selecting with read fd in fdset\n"); if (select(r_fd + 1, &readfds, NULL, NULL, NULL) == -1) { perror("select"); exit(1); } printf("select returned\n"); unlink(FILENAME); exit(0); }

After an IP address is assigned to this network interface, with any of the the following commands: ip addr add 10.0.0.0 dev eth1 valid_lft forever preferred_lft forever ip addr replace 10.0.0.0 dev eth1 valid_lft forever preferred_lft forever ip addr add 10.0.0.0 dev eth1 ip addr replace 10.0.0.0 dev eth1 I can verify with ip addr that the IP address for eth1 is set to 10.0.0.0/32 which is excellent (I think): ... 3: eth1: mtu 1500 qdisc pfifo_fast state UNKNOWN group default qlen 1000 link/ether 08:00:27:4d:1e:43 brd ff:ff:ff:ff:ff:ff inet 10.0.0.0/32 scope global eth1 valid_lft forever preferred_lft forever inet6 fe80::a00:27ff:fe4d:1e43/64 scope link tentative dadfailed valid_lft forever preferred_lft forever ... I begin to ping myself...the ping command hangs on the 31st ping: username@computer:~$ ping 10.0.0.0 PING 10.0.0.0 (10.0.0.0) 56(84) bytes of data. 64 bytes from 10.0.0.0: icmp_seq=1 ttl=64 time=0.043 ms 64 bytes from 10.0.0.0: icmp_seq=2 ttl=64 time=0.034 ms ... 64 bytes from 10.0.0.0: icmp_seq=30 ttl=64 time=0.038 ms 64 bytes from 10.0.0.0: icmp_seq=31 ttl=64 time=0.041 ms Once the ping hangs, I can verify with ip addr that the IP address for eth1 is has disappeared: ... 3: eth1: mtu 1500 qdisc pfifo_fast state UNKNOWN group default qlen 1000 link/ether 08:00:27:4d:1e:43 brd ff:ff:ff:ff:ff:ff inet6 fe80::a00:27ff:fe4d:1e43/64 scope link tentative dadfailed valid_lft forever preferred_lft forever ... **How can one assign a static IP address to a network interface using ip(8), and not let it disappear?** (disappearing after system restart is OK) I am running Ubuntu 14.04. From researching on the Internet about my problem, it seems that modifying the file /etc/network/interfaces is the solution, but this is undesirable, because this solution is not as portable as the ip(8) command.

This may have more to do with detecting operating systems, but I specifically need the init system currently in use on the system. Fedora 15 and Ubuntu now use systemd, Ubuntu used to use Upstart (long time default until 15.04), while others use variations of System V. I have an application that I am writing to be a cross-platform daemon. The init scripts are being dynamically generated based on parameters that can be passed in on configure. What I'd like to do is only generate the script for the particular init system that they are using. This way the install script can be run reasonably without parameters as root and the daemon can be "installed" automagically. This is what I've come up with: * Search for systemd, upstart, etc in /bin * Compare /proc/1/comm to the systemd, upstart, etc * Ask the user **What would be the best cross/platform way of doing this?** Kind of related, **Can I depend on bash to be on the majority of *nix or is it distribution/OS dependent?** Target platforms: * Mac OS * Linux (all distributions) * BSD (all versions) * Solaris, Minix, and other *nix

I want strict mode in my scripts. I would also appreciate portability. [set -o pipefail](https://www.translucentcomputing.com/2020/11/unofficial-bash-strict-mode-pipefail/) seems compulsory. Yet shellcheck(a static linter) is unhappy that "In POSIX sh, set option pipefail is undefined". Is it correct? If so, is this a bash solely feature or is it rather prolific?

When looking for the path to an executable or checking what would happen if you enter a command name in a Unix shell, there's a plethora of different utilities (which, type, command, whence, where, whereis, whatis, hash, etc). We often hear that which should be avoided. Why? What should we use instead?

MSYS2-built Linux applications internally rewrite Linux pathes to Windows pathes. This can be problematic if those are supposed to interface with other, proper-Linux applications (e.g. in my case, MSYS2 rsync connecting to a Debian 12 rsync server). It can be configured but neither are the means particularly elegant nor are they particularly dependable , it would seem. There is however a very appealing shorthand : Use an absolute path prefixed with a second /. e.g. /home/me/something will become C:\msys64\home\me\something but //home/me/something will remain unchanged (and the rsync server treats it as single / at least on Debian 10 and 12). I am using MSYS2 to make a single Windows machine work mostly-seamlessly in an environment otherwise dominated by Debian 12 machines, wanting to use the same tools/scripts there, if possible, so I wonder: **Can I presume a path beginning with // equals a path starting with / on all or most Linux distributions / for all or most common CLI tools?** As per the Unix standard , interpretation of the first slash in // is implementation-defined. But what do the implementations define? (This question deals with the native Linux side, but if you just happen to know whether this is a stable solution on the MSYS2 side, I would be grateful for mentioning it.)

If I include a shebang line in my script like following #! /bin/sh Would it prevent from execution on other operating systems like Windows which may not have sh in bin folder? Is this better to omit shebang line to make a portable script? How to write a operating system agnostic, shell agnostic, portable script which would execute in the same way in all operating systems or shells? If the script has to be executed the same way on Mac OS, Windows, Ubuntu, any other Linux or any other operating systems? --- Assumptions. I assume to prepare a script for a workshop where workshop participants may use different operating systems and may have different shells; I share preparation notes to install all needed tools like jq, sed, sfdx, curl and I assume the script would run the same way on all machines and all the operating systems. I assume workshop participants mainly use Mac OS, Windows, Ubuntu. Also I assume that participants may install Git Bash tool on Windows if there is no posix compatible shell. --- The script itself unwrap2.sh verbose=$1 # deploy data cloud Data Kit sf project deploy start -d dc # unwrap Data Kit components x=$(sf apex run -f scripts/unwrap.apex --json | jq '.result.logs' -r) x1=${x#*EXECUTION_STARTED} x2=${x1#*Result is \(success): } if [[ "$verbose" = "verbose" ]]; then echo "x2: $x2" fi # Actually this was my mistake here which I have posted in another question # this line should be x3=$(echo "$x2" | head -n 1) to make it work # However, original version of this script with this error didn't have the doublequotes here x3=$(echo $x2 | head -n 1) if [[ "$verbose" = "verbose" ]]; then echo "x3: $x3" fi status=$(sf data query -q "select Id,Status FROM BackgroundOperation WHERE Id = '$x3'" --json | jq '.result.records.Status' -r) if [[ "$verbose" = "verbose" ]]; then echo "select Id,Status FROM BackgroundOperation WHERE Id = '$x3'" fi echo "Status: $status" while [[ "$status" != "Complete" ]]; do sleep 10 status=$(sf data query -q "select Id,Status FROM BackgroundOperation WHERE Id = '$x3'" --json | jq '.result.records.Status' -r) echo "Status: $status" done Also, I have other scripts, like setup.sh email=$1 integrationUsername=$2 defaultFolder="force-app/main/default" mkdir "$defaultFolder/connectedApps" cp "template/connectedApps/Integration.connectedApp-meta.xml" "$defaultFolder/connectedApps" cp "template/data/int.csv" "data" SEDOPTION= if [[ "$OSTYPE" == "darwin"* ]]; then SEDOPTION="-i ''" fi # transform connected app from the template folder into default folder key=${integrationUsername//@/..} echo "Key: $key" sed -e "s/{{email}}/$email/g" -e "s/{{integration.username}}/$integrationUsername/g" -e "s/{{key}}/$key/g" "template/connectedApps/Integration.connectedApp-meta.xml" > "$defaultFolder/connectedApps/Integration.connectedApp-meta.xml" # transform integration user sed -e "s/{{email}}/$email/g" -e "s/{{integration.username}}/$integrationUsername/g" "template/data/int.csv" > "data/int.csv" # create integration user ./insertUsers.sh data/int.csv # deploy connected app sf project deploy start -d force-app # assign perm set sf org assign permset -n GenieAdmin sf org assign permset -n GenieAdmin -b $integrationUsername sf org open -p "lightning/setup/SetupOneHome/home?setupApp=audience360" which invokes insertUsers.sh data=$(cat $1) username=$(sfdx force:org:display --json | jq '.result.username' -r) echo "Username = $username" echo "Last part is ${username##*.}" sandbox=${username##*.} echo "sandbox ${sandbox}" if [[ $sandbox == 'com' ]] then sandbox=${username%%@*} echo "First part is ${sandbox}" fi data=${data//.org/.$sandbox} echo "$data" > us.csv sf data upsert bulk -s User -f us.csv -w 500 -i Id --json > upsert.json echo $(cat upsert.json) users=$(cat upsert.json | jq '.result.records.successfulResults[].sf__Id' -r) echo "Users before transformation $users" users="${users//$'\n'/','}" echo "Users after transformation $users" code=$(cat scripts/updateUser.apex) echo "${code//users/$users}" > temp.apex sf apex run --file temp.apex rm temp.apex rm us.csv rm upsert.json IFS="$old_ifs" Also, I have deploy script ./deploy.sh verbose verbose=$1 # deploy data cloud Data Kit sf project deploy start -d dc # unwrap Data Kit components scDomain=$(sf org display --json | jq '.result.instanceUrl' -r) if [[ "$verbose" = "verbose" ]]; then echo "Sales Cloud domain: $scDomain" fi scToken=$(sf org display --json | jq '.result.accessToken' -r) if [[ "$verbose" = "verbose" ]]; then echo $scToken fi version=$(cat sfdx-project.json | jq '.sourceApiVersion' -r) result=$(curl --location --request POST "$scDomain/services/data/v$version/actions/custom/flow/sfdatakit__DeployDataKitComponents" \ --header "Authorization: Bearer $scToken" \ --header 'Content-Type: application/json' \ --data @data/unwrap.json ) figuid=$(echo $result | jq '..outputValues.Flow__InterviewGuid' -r) if [[ "$verbose" = "verbose" ]]; then echo "result: $result" echo "figuid: $figuid" fi sf data query -q "SELECT Id, BundleName, ComponentName, ComponentTemplateId, ComponentType, DataKitName, DataSpaceName, DeployJob, DeploymentError, DeploymentStatus , FlowInterviewIdentifier, Name, PublisherOrgComponentId, SubscriberOrgComponentId, TemplateVersion FROM DataKitDeploymentLog WHERE FlowInterviewIdentifier = '$figuid'" --json > soql.json count=$(cat soql.json | jq '.result.records | length') if [[ "$verbose" = "verbose" ]]; then echo "Count: $count" fi if [ "$count" = "1" ]; then x="count is one" else x="count is not one" fi if [[ "$verbose" = "verbose" ]]; then echo "x: $x" fi while [ "$count" = "1" ]; do echo "Unwrapping Data Kit components in progress, waiting another 5 seconds..." sleep 5 sf data query -q "SELECT Id, BundleName, ComponentName, ComponentTemplateId, ComponentType, DataKitName, DataSpaceName, DeployJob, DeploymentError, DeploymentStatus , FlowInterviewIdentifier, Name, PublisherOrgComponentId, SubscriberOrgComponentId, TemplateVersion FROM DataKitDeploymentLog WHERE FlowInterviewIdentifier = '$figuid'" --json > soql.json count=$(cat soql.json | jq '.result.records | length') done deploymentStatus1=$(cat soql.json | jq '.result.records.DeploymentStatus' -r) deploymentStatus2=$(cat soql.json | jq '.result.records.DeploymentStatus' -r) if [[ "$verbose" = "verbose" ]]; then echo "deploymentStatus1: $deploymentStatus1" echo "deploymentStatus2: $deploymentStatus2" fi echo "Unwrapping Data Kit components completed, with statuses $deploymentStatus1 and $deploymentStatus2" aid=$(sf data query -q "SELECT Id FROM ConnectedApplication WHERE Name = 'Integration'" --json | jq '.result.records.Id' -r) res=$(curl --location --request GET "$scDomain/$aid" --header "Cookie: sid=$scToken") r1=${res#*applicationId=} r2=${r1%\'); *} echo "r2: $r2" echo "Open the Connected App View page" sf org open -p "/app/mgmt/forceconnectedapps/forceAppDetail.apexp?applicationId=$r2" echo "Trying to open Connected App so that you can click Manage and Copy Consumer Details: Key and Secret" sf org open -p "/app/mgmt/forceconnectedapps/forceAppManageConsumer.apexp?applicationId=$r2" # Potentially we could try to follow the Manage Consumer Detail link to parse Key and Secret # However, it is been redirected, so we would have to fetch the redirection link first # res=$(curl --location --request GET "$iu/app/mgmt/forceconnectedapps/forceAppManageConsumer.apexp?applicationId=$r2" --header "Cookie: sid=$sid") # echo $res # r1=${res#*replace\(\'} # l1=${r1%\');*} # l2=${l1#*replace\(\'} # echo "link: $l2" # res2=$(curl --location --request GET "$iu/$l2" --header "Cookie: sid=$sid") # # which is in $l2 variable, then we would have to figure out how to avoid insufficient permission error # Currently if you follow this link from UI, it requires MFA for confirmation. IngestS.sh script scDomain=$(sf org display --json | jq '.result.instanceUrl' -r) echo "Sales Cloud domain: $scDomain" key=$1 secret=$2 verbose=$3 result=$( curl --location --request POST "$scDomain/services/oauth2/token" \ --header 'Content-Type: application/x-www-form-urlencoded' \ --data-urlencode 'grant_type=client_credentials' --data-urlencode "client_id=$key" --data-urlencode "client_secret=$secret" ) if [[ "$verbose" = "verbose" ]]; then echo $result fi scToken=$(echo $result | jq '.access_token' -r) if [[ "$verbose" = "verbose" ]]; then echo $scToken fi result=$(curl --location --request POST "$scDomain//services/a360/token" \ --header 'Content-Type: application/x-www-form-urlencoded'\ --data-urlencode 'grant_type=urn:salesforce:grant-type:external:cdp'\ --data-urlencode "subject_token=$scToken"\ --data-urlencode "subject_token_type=urn:ietf:params:oauth:token-type:access_token") if [[ "$verbose" = "verbose" ]]; then echo $result fi dcToken=$(echo $result | jq '.access_token' -r) dcDomain=$(echo $result | jq '.instance_url' -r) if [[ "$verbose" = "verbose" ]]; then echo $dcToken echo $dcDomain fi result=$(curl --location --request POST "https://$dcDomain/api/v1/ingest/sources/Looker/treatments_statuses/actions/test " \ --header "Authorization: Bearer $dcToken" \ --header 'Content-Type: application/json' \ --data @data/dc.json ) if [[ "$verbose" = "verbose" ]]; then echo "Ingest results: $result" fi result=$(curl --location --request POST "https://$dcDomain/api/v1/ingest/sources/Looker/treatments_statuses " \ --header "Authorization: Bearer $dcToken" \ --header 'Content-Type: application/json' \ --data @data/dc.json ) echo "Ingest results: $result" IngestB.sh script set -e set -o pipefail scDomain=$(sf org display --json | jq '.result.instanceUrl' -r) if [[ "$verbose" = "verbose" ]]; then echo "Sales Cloud domain: $scDomain" fi key=$1 secret=$2 verbose=$3 result=$( curl --location --request POST "$scDomain/services/oauth2/token" \ --header 'Content-Type: application/x-www-form-urlencoded' \ --data-urlencode 'grant_type=client_credentials' --data-urlencode "client_id=$key" --data-urlencode "client_secret=$secret" ) if [[ "$verbose" = "verbose" ]]; then echo $result fi scToken=$(echo $result | jq '.access_token' -r) if [[ "$verbose" = "verbose" ]]; then echo $scToken fi result=$(curl --location --request POST "$scDomain//services/a360/token" \ --header 'Content-Type: application/x-www-form-urlencoded'\ --data-urlencode 'grant_type=urn:salesforce:grant-type:external:cdp'\ --data-urlencode "subject_token=$scToken"\ --data-urlencode "subject_token_type=urn:ietf:params:oauth:token-type:access_token") if [[ "$verbose" = "verbose" ]]; then echo $result fi dcToken=$(echo $result | jq '.access_token' -r) dcDomain=$(echo $result | jq '.instance_url' -r) if [[ "$verbose" = "verbose" ]]; then echo $dcToken echo $dcDomain fi result=$(curl --location --request POST "https://$dcDomain/api/v1/ingest/jobs " \ --header "Authorization: Bearer $dcToken" \ --header 'Content-Type: application/json' \ --data '{"object":"patients", "sourceName":"Looker", "operation":"upsert"}' ) if [[ "$verbose" = "verbose" ]]; then echo "Job create request: $result" fi jid=$(echo $result | jq '.id' -r) if [[ "$verbose" = "verbose" ]]; then echo "jid: $jid" fi if [[ "$jid" = "null" ]]; then error=$(echo $result | jq '.error' -r) message=$(echo $result | jq '.message' -r) echo "Process halted because of error: $error with message: $message" exit 1 fi result=$(curl --location --request PUT "https://$dcDomain/api/v1/ingest/jobs/$jid/batches " \ --header 'Content-Type: text/csv' \ --header "Authorization: Bearer $dcToken" \ --data-binary @data/dcPat.csv ) if [[ "$verbose" = "verbose" ]]; then echo "Data upload results: $result" fi result=$(curl --location --request PATCH "https://$dcDomain/api/v1/ingest/jobs/$jid " \ --header 'Content-Type: application/json' \ --header "Authorization: Bearer $dcToken" \ --data '{"state" : "UploadComplete"}' ) if [[ "$verbose" = "verbose" ]]; then echo "Patch job results: $result" fi result=$(curl --location --request GET "https://$dcDomain/api/v1/ingest/jobs/$jid " \ --header 'Content-Type: application/json' \ --header "Authorization: Bearer $dcToken" ) if [[ "$verbose" = "verbose" ]]; then echo "get job status results: $result" fi state=$(echo $result | jq '.state' -r) echo "state: $state" while [[ "$state" != "JobComplete" ]]; do sleep 10 state=$(echo $result | jq '.state' -r) echo "state: $state" done echo "Final state: $state"

This code works fine on Linux but not Mac OS:

#!/usr/bin/env bash
foo=$(cat <

It fails on Mac with

./test.sh: line 6: unexpected EOF while looking for matching `"'
./test.sh: line 7: syntax error: unexpected end of file

If I do cat < instead of foo=$(cat <, it works fine. Is there a portable way to get heredocs (or multiline strings) into variables without using a file as an intermediate? **Edit**: I want to use a heredoc becuase I have a multiline string with " and '. My actual example looks like:

EXPECTED_ERROR=$(cat <