I have stored configurations for some jobs in JSON format in the managed file section of Jenkins and it works as expected.
I want to expand this by changing these jobs to parametrized one, where all available managed files (not the global ones) could be selected from a combo box, but I have problems getting all available managed files.
Plugins like the configFileProvider are capable to get both the global ones and more workspace/namespace specific ones, so I thought there would be an easy way to achieve this.
Adding a more job-specific not global config file
def files = org.jenkinsci.plugins.configfiles.GlobalConfigFiles.get().getConfigs()
works fine for the global ones, but I want to get workspace/namespace specific ones and thought it would go something like this
def files = org.jenkinsci.plugins.configfiles.getConfigsInContext()
It would be amazing if somebody could point me in the right direction how to achieve this.
Thanks
I don't know any direct solution to achieve it as these managed files aren't stored as files in the disk.
My way to achieve it's running a script that parse the ${jenkins_home}/org.jenkinsci.plugins.configfiles.GlobalConfigFiles.xml file and create the managed files in the disk.
Having these files in the disk, you can use the "Filesystem List Parameter Plug-in" to list they in a job parameter.
This way you are
Script
#!/bin/bash
filenames=$(xmllint --xpath '//name/text()' GlobalConfigFiles.xml 2>/dev/null)
while read -r filename
do
content=$(xmllint --xpath "//name[text()='$filename']/../content/text()" GlobalConfigFiles.xml 2>/dev/null)
if [[ "$filename" == *"xml"* ]]; then
echo $content | sed -nr 's/</</gp' | sed -nr 's/>/>/gp' > $filename
else
echo $content > $filename
fi
done <<< $filenames
Check the following script
// name of the folde you want get configfiles from
def folderName = "Folder2"
def folderItem = Jenkins.instance.getAllItems(com.cloudbees.hudson.plugins.folder.AbstractFolder.class).find{ (it.name == folderName) }
// If you want to filter a specific configFile type, example Json configs here
def descriptor = org.jenkinsci.plugins.configfiles.json.JsonConfig.JsonConfigProvider.class
def allConfigFiles = folderItem.getProperties().get(org.jenkinsci.plugins.configfiles.folder.FolderConfigFileProperty.class).getConfigs();
def jonConfigFiles = folderItem.getProperties().get(org.jenkinsci.plugins.configfiles.folder.FolderConfigFileProperty.class).getConfigs(descriptor);
echo "$allConfigFiles"
echo "$jonConfigFiles"
Related
To create a binary shell script I'm using writeShellScriptBin. This function takes a string that represents the shell script to be produced, e.g.:
my-script = writeShellScriptBin "my-script" ''
# Some script
''
Using writeShellScriptBin works great for small scripts, but once the script starts growing the string that produces it becomes hard to maintain. So I was wondering what are the alternatives to writeShellScriptBin.
There is substituteAll. Example:
substituteAll {
src = ./foo.in;
name = "foo";
dir = "/bin"; # If you omit it, you will get substitution result in $out
isExecutable = true; # it is script, so you are likely want this.
inherit curl jq execline;
}
In my Docusaurus project my internal links work on my local environment, but when I push to GitLab they no longer work. Instead of replacing the original doc title with the new one it adds it to the url at the end ('https://username.io/test-site/docs/overview/add-a-category.html'). I looked over my config file, but I do not understand why this is happening.
I tried updating the id in the front matter for the page, and making sure it matches the id in the sidebars.json file. I have also added customDocsPath and set it to 'docs/' in the config file, though that is supposed to be the default.
---
id: "process-designer-overview"
title: "Process Designer Overview"
sidebar_label: "Overview"
---
# Process Designer
The Process Designer is a collaborative business process modeling and
design workspace for the business processes, scenarios, roles and tasks
that make up governed data processes.
Use the Process Designer to:
- [Add a Category](add-a-category.html)
- [Add a Process or Scenario](Add%20a%20Process%20or%20Scenario.html)
- [Edit a Process or Scenario](Edit%20a%20Process%20or%20Scenario.html)
I updated the add a category link in parenthesis to an md extension, but that broke the link on my local and it still didn't work on GitLab. I would expect that when a user clicks on the link it would replace the doc title in the url with the new doc title ('https://username.gitlab.io/docs/add-a-category.html') but instead it just tacks it on to the end ('https://username.gitlab.io/docs/process-designer-overview/add-a-category.html') and so the link is broken as that is not where the doc is located.
There were several issues with my links. First, I converted these files from html to markdown using Pandoc and did not add front matter - relying instead on the file name to connect my files to the sidebars. This was fine, except almost all of the file names had spaces in them, which you can see in my code example above. This was causing real issues, so I found a Bash script to replace all of the spaces in my file names with underscores, but now all of my links were broken. I updated all of the links in my files with a search and replace in my code editor, replacing "%20" with "_". I also needed to replace the ".html" extension with ".md" or my project would no longer work locally. Again, I did this with a search and replace in my code editor.
Finally, I ended up adding the front matter because otherwise my sidebar titles were all covered in underscores. Since I was working with 90 files, I didn't want to do this manually. I looked for a while and found a great gist by thebearJew and adjusted it so that it would take the file name and add it as the id, and the first heading and add it as the title and sidebar_label, since as it happens that works for our project. Here is the Bash script I found online to convert the spaces in my file names to underscores if interested:
find $1 -name "* *.md" -type f -print0 | \
while read -d $'\0' f; do mv -v "$f" "${f// /_}"; done
Here is the script I ended up with if anyone else has a similar setup and doesn't want to update a huge amount of files with front matter:
# Given a file path as an argument
# 1. get the file name
# 2. prepend template string to the top of the source file
# 3. resave original source file
# command: find . -name "*.md" -print0 | xargs -0 -I file ./prepend.sh file
filepath="$1"
file_name=$("basename" -a "$filepath")
# Getting the file name (title)
md='.md'
title=${file_name%$md}
heading=$(grep -r "^# \b" ~/Documents/docs/$title.md)
heading1=${heading#*\#}
# Prepend front-matter to files
TEMPLATE="---
id: $title
title: $heading1
sidebar_label: $heading1
---
"
echo "$TEMPLATE" | cat - "$filepath" > temp && mv temp "$filepath"
I am using jenkins for automated deployment.
I needs to modify xml tag value in xml file using groovy script. I am using below groovy code. When I try to edit xml tag value I am receiving error unclassified field xml.uti.node error.
Node xml = xmlParser.parse(new File("c:/abc/test.xml"))
xml.DeployerServer.host[0] = '172.20.204.49:7100'
FileWriter fileWriter = new FileWriter("c:/abc/test.xml")
XmlNodePrinter nodePrinter = new XmlNodePrinter(new PrintWriter(fileWriter))
nodePrinter.setPreserveWhitespace(true)
nodePrinter.print(xml)
I need to modify host tag value and host is available inside DeployerServer tag.
Any help will be much appreciated.
Here is the script, comments in-line:
//Create file object
def file = new File('c:/abc/test.xml')
//Parse it with XmlSlurper
def xml = new XmlSlurper().parse(file)
//Update the node value using replaceBody
xml.DeployerServer.host[0].replaceBody '172.20.204.49:7100'
//Create the update xml string
def updatedXml = groovy.xml.XmlUtil.serialize(xml)
//Write the content back
file.write(updatedXml)
I was wanting to read / manipulate the CSProj file and NUSPEC files in a Pipeline script. I could not get passed the parseText() without the dreaded "SAXParseException: Content is not allowed in prolog".
There are quite a few threads about this error message. What wasn't clear is that both CSProj and NUSPEC files are UTF-8 with BOM - BUT this is invisible!
To make it worse I've been trying to automate the NUSPEC file creation, and there is no way I can tell the tools to change file encoding.
The answers above helped solve my issue, and once I added code to look for 65279 as the first character (and deleted it). I could then parse the XML and carry out the above.
There didn't seem to be good thread to put this summary on, so added it to a thread about Jenkins, Groovy & XML files which is where I found this "known Java" issue.
I used powershell to do this change in app.config file.
My problem was with passwords. So, I created a Credential, in jenkins, to store the password.
If you do not need to work with credential, just remove the withCredentials section
Here is part of my jenkinsfile:
def appConfigPath = "\\server\folder\app.config"
stage('Change App.Config'){
steps{
withCredentials([string(credentialsId: 'CREDENTIAL_NAME', variable: 'PWD')]) {
powershell(returnStdout: true, script: '''
Function swapAppSetting {
param([string]$key,[string]$value )
$obj = $doc.configuration.appSettings.add | where {$_.Key -eq $key }
$obj.value = $value
}
$webConfig = "'''+appConfigPath+'''"
$doc = [Xml](Get-Content $webConfig)
swapAppSetting 'TAG_TO_MODIFY' 'VALUE_TO_CHANGE'
$doc.Save($webConfig)
''')
}
}
}
Don`t forget to update your powershell. (minimum version 3)
I want to read a file a extract information from it based on certain tag. For example :
SCRIPT_NAME:mySimpleShell.sh
This is a simple shell. I would like to have this as
Description. I also want to create a txt file our of this.
SCRIPT_NAME:myComplexShell.sh
This is a complex shell. I would like to have this as
Description. I also want to create a txt file our of this.
So when I pass in this file to my shell script, my shell will read it line by line and
when it gets to SCRIPT_NAME, It extract it and save it in $FILE_NAME, then starts writing
the description to a file on disk with $FILE_NAME.txt name. And It does it until It reaches the end of file. If there is 3 SCRIPT_NAME tag, then it creates 3 description file.
Thanks for helping me in advance :)
Read the lines using a while loop. Use a regex to check if a line has SCRIPT_NAME and if so, extract the filename. This is shown below:
#! /bin/bash
while IFS= read -r line
do
if [[ $line =~ SCRIPT_NAME:(.*$) ]]
then
FILENAME="${BASH_REMATCH[1]}"
echo "Writing to $FILENAME.txt"
else
echo "$line" >> "$FILENAME.txt"
fi
done < inputFile
#!/bin/sh
awk '/^SCRIPT_NAME:/ { split( $0, a, ":" ); name=a[2]; next }
name { print > name ".txt" }' ${1?No input file specified}
Is it possible to get the path to the php.ini file with a php script and save this path to a variable? I know I can call phpinfo() to find out the path, but it prints a lot of info and I only need the path to php.ini file and no output at all. Thank you in advance.
Sure, there are two functions related to what you would like to do. The first one is exactly what you're looking for, the second one shows the bigger picture that there can be more than one ini file:
php_ini_loaded_fileDocs - Retrieve a path to the loaded php.ini file.
php_ini_scanned_filesDocs - Return a list of .ini files parsed from the additional ini dir.
Next to that, mind the gap with .user.ini files, they don't show up in php_ini_scanned_files nor phpinfo.
You could exec("php -i | grep php.ini"), and grab that output.
Or you could use outputbuffering (ob_start()), run phpinfo(), get the contents of the outputbuffer (ob_get_contents()) and search trough that (preg_match) to find the php.ini file...
This works on the command line, might also work for the CGI:
ob_start(); phpinfo();
if ( preg_match( '#>(.+)/php.ini#', ob_get_clean(), $matches ) ) {
echo 'php.ini location: ' . trim( $matches[1] ) . '/php.ini';
}