How I switched to Lando to develop using Docker!

A few months ago I tested Docker to manage my development environments. The major advantage of Docker is to be able to develop locally on the same technical basis of the server. Admit that it’s tempting when you have sites hosted by different providers (even if most of them for my case are at O2Switch ) with different software versions of PHP, Mysql, Mariadb. The major drawback of Docker, in my humble opinion, is the complexity of configuration! I had never managed to run several vhosts at the same time!

With Lando, it’s easy!

While chatting with a colleague on the French WordPress Slack, I discovered Lando.

To use Lando, you first need to install Docker, then Lando comes as an overlay. It will greatly facilitate the configuration of the different Docker images to use to run your environment.

WordPress Recipes!

Lando offers “Recipes”, literally recipes for the main CMS and Frameworks on the market, including WordPress.

With the WordPress recipe, Lando will create an environment including PHP, and Mariadb to run your site but also wp-cli!

In short, in a few seconds of launching, you are ready to develop on your favorite CMS.

A simplified configuration for Docker!

Here is an example of the .lando.yml configuration file that you will need to have at the root of your project to launch your Lando instance!

name: test
recipe: wordpress
config:
  webroot: .
  php: '7.0'
  xdebug: true
  database: mariadb
services:
  pma:
    type: phpmyadmin
    hosts:
      - database
      - old-test
  mailhog:
    type: mailhog
    hogfrom:
    - appserver
  old-test:
    type: mariadb
    portforward: true
    creds:
      user: wordpress
      password: wordpress
      database: old-test
tooling:
  compose:
    service: appserver

This is a simple text file in .yml format that will be read by Lando to create your environment.

Note a “name”, this one will be used to create your local development url! as well as the services you want to create and in which version! Here I chose a PHP version 7.0, Mariadb in its current stable version, Phpmyadmin, Mailhog to “catch the mails” without having a mail server to configure. There is also the configuration of a second ‘old-test’ database server, the activation of Xdebug and Composer! Lando on Github

Migrate a site from a multisite to a classic installation

Introducing the WordPress multisite.

The Multisite technique allows you to run several WordPress sites together within a single installation.

This technique allows you to share the WordPress core, themes and extensions.

This results in significant time savings when maintaining your sites since you only update once.

Read more about this:

Extracting a site from a multisite step by step.

First of all, this tutorial is not meant to be THE method but a method that works and that I put into action today as part of my job as a WordPress developer.

Step -1 – Backup

Saaaaauuuuuvvvvvveeeeegggggaaaaaarrrrrrdddddeeeeee

Backup your files (WP, extensions, themes, mu-plugins and Mysql database)

Step 0 – Local installation

Having the multisite locally, it will be much easier to manage and in case of accident, it will generate only you!

Step 1 – Updates

The first step will be to update your multisite:

  • WordPress
  • Themes
  • Extensions
  • Languages

If this step is not fundamental, I think that:

  1. it can’t hurt
  2. it can avoid conflicts and worries later on when you import themes and extensions into your “Classic” WordPress site.

Step 2 – Installing WordPress

Install a brand new WordPress… For this I suggest you do it in 3 minutes thanks to the wp-cli script I created for you…

OK I admit, I created it for myself… and I thought you might be interested!

Step 3 – Transferring the files

So now, let’s start a file transfer. That’s why I asked you in step 0 to work locally. It will be much easier to move files in the explorer of your machine than remotely via sFTP or whatever.

Remember to transfer:

  • Themes (parent and children)
  • Extensions
  • Mu-plugins
  • the “uploads” folder… but I reserve him the step 4!

Of course, on the multisite maybe not all the extensions or all the themes were useful for the site you are extracting.

Upload only what you need and do not activate anything

Step 4 – the “Uploads” folder

The Uploads folder located in `wp-content` contains all the media uploaded to your site and available in the media library.

Admit that it would be a shame not to find them in your new installation.

The Upload folder is a little different to transfer since its folder and subfolder architecture is different from a classic installation.

Logiciel
Architecture of the Uploads folder in Multi-site

Notice the difference with a classical architecture. In “uploads”, we have a “sites” folder that does not exist in a regular WordPress installation.

Inside this “sites” folder we have a folder named by the id of the site in the multisite infrastructure.

Then we find the classic WordPress media classification with a year folder and a month folder.

Find the site ID to extract from the multisite

I believe that if you do this manipulation then you have the privileges of super-administrator.

This step is quite simple, you have to go to the network administration of your multisite then in the submenu “Sites” at the end of the line concerning the site to migrate you will find the ID of this one.

Microsoft Azure
Find the ID of the site to migrate

The ID recovered you can move all the contents of the folder with the id as name in wp-content/uploads to find the classic tree structure of a WP.

Step 5 – export then import the database

It’s all very well but for the moment, you have everything except your content…. and in a site, it’s the most important thing, the database.

For this tutorial, I will assume that you have the default prefix “wp_”.

Just like media, WordPress multisite adds the id of the site on some tables.

You will have to export all the tables with the ID as well as the wp_user and wp_usermeta tables which do not have an id because in a multisite, the users are common to all the sites.

Personally I do it with PHPmyAdmin if the tables are not too big… we select the desired tables (example: the one of the site “4” wp_4_ as well as the 2 users tables).

Before importing, delete the tables wp_user and `wp_usermeta` from the classic installation to avoid conflicts with the tables you are going to import.

For the import in the classic WP, also use the tool you are used to.

You will end up with duplicate tables: those of your classic installation and the one you just imported that have the prefix `wp_4_`. Delete the duplicate tables from the classic installation.

Import your “multi-site” tables

At this stage, if you look at your database in PHPmyadmin or other, you should notice some tables in wp_ and others in wp_4_

In PhpMyadmin, check all the tables in `wp_4_` then at the bottom of the page in the dropdown menu select “Replace table prefix”

If you decide to change the table prefix, don’t forget to correct the `wp-config.php` accordingly.

Document

We’re almost there!

Step 6 – Replace the media URLs in the Database.

For this I use the SRDB script from Interconnectit. You can find a guide to this script from my friend Grégoire Noyelle.

Here we are going to modify the urls to the medias and thus make a search and replace which would look like this:

Search: https://example.com/wp-content/uploads/site/4/2019/02/mon-image.jpg

Replace: https://example.com/wp-content/uploads/2019/02/mon-image.jpg

Step 7 – the user base

When importing the database, we imported all the users of the multisite… but some of them have no relation with the site you just extracted.

Unfortunately, I haven’t yet found an sql query that fits … if you have an idea, submit it to me in comments, I’ll be happy to improve this tutorial.

Step 8 – Checks

Normally, your classic site is an exact copy of your site when it was in the multisite.

Depending on your setup, there may be 2/3 things to adjust. For example, replace urls to images written in hardcopy in css files created by Elementor page builder…

You may have transferred extensions that are only useful for a multisite…

Step 10 – Rest!

Yeah, there’s no step 9 because it’s good to stop on a round number! a dozen or so!

Normally you have just extracted a site from a multisite to a classic WordPress site… you have earned a little foam, a coffee or anything else that will make you happy!

And if not, there’s the easy way! (added on April 20, 2019)

If like me you are adept of WP-CLI then this script will allow you to extract a site from the multisite in a few seconds…

trepmal/blog-extractor

WP-CLI command. Extract single blog from multisite – trepmal/blog-extractor

https://github.com/trepmal/blog-extractor

How to fix the Gravity Form error “Call to a member function can_be_used() on bool”

This article is more for my future self, but it may help others.

Following the upgrade to Gravity Form 2.6.2, I found myself on several sites I maintain with a fatal error looking like:

[06-May-2022 20:30:52 UTC] PHP Fatal error: Uncaught Error: Call to a member function can_be_used() on bool in /var/www/my-great-website/wp-content/plugins/gravityforms/common.php:2813
Stack trace:

I didn’t dig too much to know the cause… I’ll do that later… when a client site is down… we correct it quickly 😉

To correct this error I just added

define( 'GF_CACHE_DEBUG', true );

in the `wp-config.php` and then go to resave the form settings (without changing anything)

This error occurs when the license verification fails. The developers of Gravity Forms are studying the issue and will soon publish a new version.

The added constant allows to validate the license without taking into account the cached data.

Tell me in comments if you had this error and this tip helped you!

Deploy your website with Github Actions

I already told you about Github Actions in a previous article to easily publish your WordPress extension on the official extensions repository

Today, the idea is to automatically deploy your developments when you push your changes to GitHub

State of the art

Currently, I publish my changes to my clients’ WordPress sites through DeployHQ.com

. It’s a very simple service to set up:

  • Link DeployHQ to our Github profile
  • Select the GH repository to deploy
  • Set the SFTP connection on our hosting

At this point, DeployHQ adds a webhook that will be triggered on each push to a defined branch of our Github repo that will start the deployment.

A few seconds later (or it counts in seconds), your changes are online.

Principle of Github Actions

In simple terms, Github Actions allows you to automate tasks according to events occurring on the repository on which it is active.

There can be hundreds or even thousands of tasks and in order to help developers, Github has set up a

marketplace where anyone can publish their Action in order to share it.

The principle:

  • Create a configuration file in yml format.
  • This file is stored in your GitHub repository
  • When the action defined in the configuration file is executed, GitHub Actions comes into play and executes the tasks requested.

Let’s get down to business

In the previous paragraph, I told you about a marketplace. In this marketplace, I came across https://github.com/marketplace/actions/ssh-deploy which according to the description, does exactly what I want it to do.

In practice, you need to create a .github folder in your project that will itself contain a workflows folder. Inside this folder you will add a file with any name you want, but in yml format. For example deploy.yml

.

GitHub actions

In this deploy.yml file we will add this code:

name: Deploy

on:
  push:
    branches:
      - master

jobs:
  build:

    runs-on: ubuntu-latest

    steps:
      - uses: actions/checkout@v1
      - name: Install Node.js
        uses: actions/setup-node@v1
        with:
          node-version: '10.x'
      - name: Install npm dependencies
        run: npm install
      - name: npm init
        run: npm init -y
      - name: Run build task
        run: npm run build --if-present
      - name: Deploy to Server
        uses: easingthemes/ssh-deploy@v2.1.7
        env:
          SSH_PRIVATE_KEY: ${{ secrets.SERVER_SSH_KEY }}
          ARGS: "-avz --delete"
          REMOTE_HOST: ${{ secrets.REMOTE_HOST }}
          REMOTE_PORT: ${{ secrets.REMOTE_PORT }}
          REMOTE_USER: ${{ secrets.REMOTE_USER }}
          TARGET: ${{ secrets.REMOTE_TARGET }}
          EXCLUDE: "/.git, /.github, /dist/, /node_modules/"

And what does that mean?

name: Deploy

Is simply the name of this action, because of course you can have as many actions as you want.

on:
push:
branches:
- master

When do we get this action? easy! “The “push” of the “master” branch.

So we can imagine having another action with :

on:
  push:
    branches:
      - staging

The next block we are interested in is:

        env:
SSH_PRIVATE_KEY: ${{ secrets.SERVER_SSH_KEY }}
ARGS: "-avz --delete"
REMOTE_HOST: ${{ secrets.REMOTE_HOST }}
REMOTE_PORT: ${{ secrets.REMOTE_PORT }}
REMOTE_USER: ${{ secrets.REMOTE_USER }}
TARGET: ${{ secrets.REMOTE_TARGET }}
EXCLUDE: "/.git, /.github, /dist/, /node_modules/"

Because it defines the information related to Rsync

SSH_PRIVATE_KEY

The private SSH key used for deployment. According to the Action doc, you can create one with the command:

ssh-keygen -m PEM -t rsa -b 4096

This command will create a private key, to be kept secret, and a public key.

The public key is to be added in the file /home/user/.ssh/authorized_key of your server / hosting. The path will be adapted depending on whether you are on a shared hosting allowing SSH or on a dedicated server or other case.

The private key must be stored on your Github account in the settings of the repository you want to deploy. The principle will be the same for each of the following instructions.

guthub actions secrets

By clicking on “New repository secret” you will have 2 fields.

  • The “Name” field will take the capitalized part of your settings, for example SERVER_SSH_KEY
  • The “Value” field will take the value to send during the transfer (here the value of the private key)
ARGS: "-avz --delete"

The arguments of RSYNC to be found in its doc

I invite you to look at the very simple and clear Action doc for the other parameters.

Tadam!

Normally… if everything went well, your modifications are uploaded automatically when you push them on your main branch (or on the branch you have chosen)

Conclusion

You now know how to use Github Actions to deploy your sites, all that’s left to do is to make nice websites 😉

If like me you are a WordPress developer but you host your sites at WPEngine, know that there is an action that allows you to deploy directly at WPEngine!

I’m new to this tool… if you have any questions… it’s not easy, but give it a try anyway by leaving a comment.

Add custom icons to your ACF with ACF projects: FontAwesome Field

If you’re used to creating your own web icons, you must be familiar with Icomoon. For a client project already using FontAwesome, I needed to add 2 customs icons from using the Icomoon service.

In this blog post, I’m using Icomoon, but you should be able to reproduce the same behaviour with another icon provider, maybe with some adaptation.

FontAwesome icons were used on menu items thanks to Advanced Custom Fields and Advanced Custom Fields: Font Awesome Field. The goal is to add to the list of FontAwesome icons, my 2 or 3 customs icomoon icons.

Show Icomoon icons in the ACF FontAwesome field

Advanced Custom Fields: Font Awesome Field that I will name ACFFA from now offers a filter to modify the list of icons:

ACFFA_get_icons
add_filter( 'ACFFA_get_icons', 'cot_test' );
function cot_test( $icons ) {
$icomoon = array(
'cierge' => array(
'name' => 'cierge',
'hex' => '\e90e',
'unicode' => ''
),
'calice' => array(
'name' => 'calice',
'hex' => '\e90f',
'unicode' => ''
),
);
foreach ( $icomoon as $ico ) {
$icons['list']['far'][ 'far ' . $ico['name'] ] = '<i class="icon-' . $ico['name'] . '"></i> ' . ucfirst(
$ico['name'] );
$icons['details']['far']['far ' . $ico['name']]['hex'] = $ico['hex'];
$icons['details']['far']['far ' . $ico['name']]['unicode'] = $ico['unicode'];
}

return $icons;
}

Here we filter the existing list of FA icons to add our custom icons. Unfortunately, since the Matt Keys extension is developed for FA, only the FontAwesome categories are available, namely Brand (fab), Regular (far), Style (fas). I decided to add my icons in the Regular or Far categories. This is purely arbitrary.

DéLet’s break down the code above

array(
'cierge' => array(
'name' => 'cierge',
'hex' => '\e90e',
'unicode' => '&#xe90e;'
),

Here I declare the Icomoon icons I’m going to want to use:

  • name: the name of the icon in the list
  • hex: the hexadecimal code provided by Icomoon
  • unicode: the unicode of the Icomoon icon

Then, with a foreach loop, I’m going to feed the $icons array containing the FontAwesome icons

Exemple d'icône personalisé dans le champs ACF

Display icons in the back office and front office

To display the icons in back-office, the Icomoon CSS sheet with all the icons will have to be loaded on the back-end side.

For that, we’re going to add this code:

add_action( 'admin_print_styles', 'cot_load_icomoon_css' );

If you are used to developing your themes you know the WordPress action hook wp_enqueue_scripts that allows you to load stylesheets and scripts in the frontend. There is also an equivalent for the back office called admin_print_styles.

add_action( 'admin_print_styles', 'cot_load_icomoon_css' );
add_action( 'wp_enqueue_scripts', 'cot_load_icomoon_css', 9, 1 );
function cot_load_icomoon_css(){
wp_enqueue_style( 'font-icomoon', get_template_directory_uri() . '/css/iconmoon.css');
}

So here I load the CSS sheet in front and back office.

I put a priority 9 at the front loading in order to pass just before the loading of the theme stylesheet and to be able to use them afterwards in my theme. By default an add_action() has a priority of 10.

Display the icons in front-office

In front-office, the ACF FontAwesome extension returns HTML markup formatted for FA (normal!). In some cases, we want to load another font (the one generated by Icomoon) in our custom icon.

Icomoon provided me with 4 font files (.eot, .svg, .ttf and .woff) that I placed in my projects. The icomoon.css file does a @font-face so that the fonts are loaded

The problem lies in the fact that the Matt Keys extension generates an HTML FA markup and it will be necessary to “trick” to hijack the operation only for Icomoon icons. And yes, we want to be able to continue to use FontAwesome normally.

<i class="far cierge" aria-hidden="true"></i>

We find the far class of FA which will load the FontAwesome font then the cierge class generated from the name given in the list of extensions to insert.

To make the site load my font instead of FontAwesome’s, I created a CSS sheet modifying the style of these classes:

.far.cierge:before {
font-family: "my-icomoon-font";
content: "\e90e";
}

For each custom icon I ask WordPress to load my-icomoon-font and use the content with the hex code corresponding to my icon.

 

Install WordPress Coding Standards on PHPStorm easily

WordPress, as all development frameworks, has its own rules to follow so that each developer can easily read and understand a colleague’s code.

This is what we call “Coding Standards”. For WordPress, we have the WordPress Coding Standards better known as WPCS!

These developments rules can be read here: Discover the WordPress Coding StandardsOK, we’ve got a thousand pages of docs… but I’m not gonna read all those docs while developing… couldn’t you make it easier?

Well you’re lucky because we can integrate these rules directly into our IDEs or text editors! So your code will be automatically checked during your development.

I use PHPStorm from Jetbrains every day but you should know that these rules can be used on almost any code editor.

Where do I find the WPCS?

WPCS are available on Github. I will detail how I do it but there are several methods. Take your pick!

Séb method for installing WPCS

So it’s not really my method… in the sense that it’s not me who developed it… but it’s the one I’m using and that I’m going to detail!

Before installing WPCS, you need to install PHPCS… and yes, WordPress is developed in PHP and to make it simple, WPCS adds rules or modify a little the rules of the PHP Codings Standards.

Keep in mind that, as in all the WordPress ecosystem, the goal is to make development accessible and easy.

Install PHPCS

Personally, I have a partition on my Ubuntu Linux dedicated to my projects and next to the projects, I have a folder with the WPCS and one with the PHPCS.

PHPCS are also available on Github

Clone PHPCS Github repo

Go to the place where you want to download the rules then:

or

if you haven’t entered an SSH key in your GitHub account.

You should now have a folder named phpcs.

Clone WPCS Github repo

We’ll do almost the same command to clone the WPCS repository this time:

or

I reassure you, the hardest part is over…

Setting PHPStorm

Go to “File | Settings | Languages & Frameworks | PHP | Quality Tools”.

Programme d'ordinateur
PHPStorm settings

By clicking on the 3 points to the right of Configuration you can specify the path to the PHPCS. It will be necessary to navigate and select the file “phpcs” present in the directory “phpcs/bin” obtained thanks to the preceding cloning.

Capture d'écran

If everything is OK, at the bottom of the window will show the version of Code Sniffer installed.

Let’s move on to WPCS, because after all… they’re the reason we’re here!

The path to the settings is File | Settings | Editor | Inspections

Logiciel
I love arrows!

the goal here is to get to the “PHP Code Sniffer Validation” menu in order to check “Installed Standard Paths” and to search after clicking on the folder on the right for the directory to be cloned with WPCS.

If all goes well… by clicking on the 2 circular arrows, you should have the choice of several WPCS… personally, I choose WordPress-Extra. then we validate by clicking OK…

Geolocalize an address without Google Maps

We’ve long been used to using Google’s tools and services to do a lot of things… only good things come to an end and some of these services close or become paid for.

As part of the development of my OpenAgenda plugin for WordPress, I need to find the longitude and latitude of an event address.

OpenStreetMap to the rescue

Fortunately, there are many services that can replace those of Google. Among them OpenStreetMap (OSM) for everything that revolves around maps and geolocation.

Nominatim for the conversion of address to location data.

So I discovered the Nominatim service.

That’s all very nice, but I want code!

If like me, you need to retrieve the latitude and longitude of a geographic point in PHP… for WordPress, here’s what I do:

A little explanation:

Line 1: a function is declared in which I pass by setting an address (street postal code city).

Line 2: I encode this address so that it can be integrated into a URL.

Line 3: I check if a transition with the coordinates exists. Nominatim limits access to its IAP… the transient allows you to store the information locally and not to make a request at each display… especially since the geographical coordinates of a point should not change too often….

Line 14: I create the transient that will be valid for 90 days. Basically, your WordPress will ask for geolocation only once every 90 days for the same address.

Between 2: if the address has never been geocoded then I request the Nominatim API.

Any ideas or improvements to make? Leave me a comment!

Develop your WordPress Project using Lando

I already told you about Lando a while ago… it was more of a cry from the heart to have found a simple solution to use Docker technology.

What’s Lando?

As I explained in my previous article, Lando is an overlay that makes it easy to use Docker Technology.

To put it simply, you have to see Docker as a modular tool where each brick is the component of your web server.
So you will need for example, to run a WordPress locally an Apache brick, a PHP brick and a Mariadb or Mysql brick.
Among these different bricks, there are different versions. For example PHP7.0 / PHP7.3 .

To develop soundly and avoid unpleasant surprises during production, it can be interesting to use during our developments the same software versions as those used by the server that will host the website in the end. In this way we reduce the risks of conflicts and errors.

Docker’s goal is to provide these different bricks and make them work together… except that…

Except that, as I said in my previous article, you have to set all this up… and have sysadmin knowledge that we don’t all have… fortunately Lando has arrived…

Install Lando

The hardest part of installing Lando is actually installing Docker which is quite simple if you are free like me, i.e. on a Linux distribution. It’s a bit more complicated if you’re trapped in a private system such as Windows or Mac OS. (For once we have something simpler under Ubuntu Linux… )

In short, Docker runs natively under Linux, you’ll have to add a virtual machine to install Docker on Windows or Mac OS.

Once Docker is installed, go to Lando’s website to install the overlay that will set all this to music.

Configuring Lando

I told you earlier that Lando was ignoring the complexity of Docker’s parameterization.

All the configuration of your development server will fit in a text file.

Let’s imagine that we want to create a WordPress site with the url https://mon-super-site.lndo.site.

Just create a folder on your machine where you have the write access and that you would name with the name you want: my-super-site.

In this folder we will add a .lando.yml file that will contain all the configuration of the server of our site:

That’s it, that’s all! Lando is ready to go! And you’re ready to develop!

Let’s decorate the .lando.yml:

  • Line 1: A name for your development environment. This name will be used to build the url of your project.
  • Line 2: “Recipe” means recipe in English. Lando has developed preconfigurations according to what you will do with it. Here is a WordPress recipe, it will install wp-cli. There are some for Pantheon, Drupal and more => Lando Recipe
  • Line 3: a config block where you can specify your software versions
  • Line 7: a Service block where we will list the services we want to see working. here PhpMyadmin and Mailhog
  • Line 16: Composing is still practical when working in the PHP world.
  • Line 19: This block is optional if you are working on a simple domain. It can be very useful when working on a Multisite WordPress subdomain for example.
  • Line 22: maybe the nicest part… do some tasks just after importing a database… here we do via wp-cli a search & replace to adapt the urls and we disable the excellent WordPress WP-Rocket extension because the cache in dev… how to say 😉

This .lando.yml is quite simple and there is a multitude of other configurations and possibilities to best fit your needs.

Note that Lando is oriented for development and is strongly discouraged by its developers for a production

To run Lando for the first time, you have to do a `lando rebuild –y’.

At this point, it will create and download what it needs. This step can be long depending on your web connection but it will only be done once…

Then for the next launches a `lando start` in your terminal in your working directory will be enough.

Attention the WordPress recipe prepares lando technically to receive WordPress but does not install WordPress. You will perform this installation step as you are used to or by using my wp-cli method.

I use Local By Flywheel… it’s the same thing!

Hummm each solution has its advantages and disadvantages. When I decided to test and finally adopt Lando, Local By Flywheel was not available under Linux. This is no longer the case since the summer of 2019.

I (too) quickly tested Local By Flywheel and I couldn’t get a WP to work :-/ (note it too much).

The only thing Flywheel brings for my case is a GUI (Graphical User Interface) whereas Lando… it’s command-line… but sincerely to make lando start or `lando rebuild –y`? install an interface that will necessarily use machine resources… I didn’t see the point.