Author: Abdullah Yahya

Cherry-picking individual commits to merge them from one git branch to another is a very common practice. For me, I usually cherry-pick one or more commits from a develop branch to a master / main branch. The problem, however, is when I need to cherry-pick multiple commits made over a long period of time, like one month. Most git clients don’t make it easy to filter commits by keyword, and if they do, then they don’t let you easily select all filtered commits. So, back to the command line it is.

This post will describe one way to easily find all commits matching one or more keywords and extract their commit IDs on one line so you can quickly run the git cherry-pick command.

Step 1: Find and review filtered commits

We usually prefix our commit messages with an identifier so we can filter them amongst many other commits. For example, all commits to related to one project may have a commit message prefix of “LP:8929” where, in this case, “LP” stands for Landing Page. Sometimes, we might accidentally use an equal sign (=) instead of a colon (:). So, we’ll have one or more keywords to search commit messages when we filter all commits.

Furthermore, we usually cherry-pick commits from a develop branch to the master branch while we’re on master. The following command does the following:

git log master..develop --oneline --grep='LP:8929' --grep='LP=8929' -i
48b1cbcef51 LP:8929 - Fixed images path by using imagekit variable
8efa19678b3 LP:8929 - Fixed text alignment
3ab0954c725 LP:8929 - Updated banner and VS page on the landing page
bba58a1c5eb LP:8929 - Updated main heading of features section
fb56da3b134 LP:8929 - Updated content and created carousel as per the requirement
c7c6d50fbdb LP:8929 - Updated styling of the page
bc8a7fc38a1 LP:8929 - Updated icons of the comparison section
9a90ccb91ad LP:8929 - Updated text styling
a3bc0baf65f LP:8929 - Updated headshots and styling of the page
13977d2eb12 LP:8929 - Updated content as per the doc
7e589debba8 LP:8929 - First Pass: Created Landing page of Tenable v Qualys competitive page

Now that we have our list of commits, we need to review them to make sure they are correct. In our case, we primarily want to make sure the keywords are at the beginning of the commit message since that’s our commit message format.

Step 2: Extract commit IDs

If our list of commits is correct, then we need to extract the commit IDs. However, the git cherry-pick command requires the list of commit IDs to be in chronological order. By default, the git log output is in reverse chronological order. Here’s the updated command with some modifications.

git log master..develop --reverse --oneline --grep='LP:8929' --grep='LP=8929' -i  | awk '{print $1}' | tr '\n' ' '
48b1cbcef51 8efa19678b3 3ab0954c725 bba58a1c5eb fb56da3b134 c7c6d50fbdb bc8a7fc38a1 9a90ccb91ad a3bc0baf65f 13977d2eb12 7e589debba8

Step 3: Run git cherry-pick command

Now that we have our list of commit IDs in the correct order, we can copy it, type “git cherry-pick” and paste the list to create a command like the following.

git cherry-pick 48b1cbcef51 8efa19678b3 3ab0954c725 bba58a1c5eb fb56da3b134 c7c6d50fbdb bc8a7fc38a1 9a90ccb91ad a3bc0baf65f 13977d2eb12 7e589debba8

Hit enter to run your git cherry-pick command.

Svelte is a JavaScript framework similar to React, Vue, etc, but with some fundamental differences that I think make it better. It’s also much more intuitive to understand as the web component structure of the code is dead simple. When building an app using Svelte, you’re encouraged to use Sveltekit, the official Svelte application framework. But, if you just want a simple frontend-only site/app, then you can just use Svelte (without Kit). In doing so, you can then easily take the built CSS and JS files and include them in any existing web page. Then, you just add a simple div

<div id="app"></div>

in your existing HTML file where you want the app go to.

Here’s an example.

Create a Vite Project

Vite is a front-end build tool and dev server. It includes Rollup to bundle and optimize assets. Run npm init vite and follow the prompts. As you can see in the screenshot below, I chose the default project name of “vite-project”.

This will create a folder named after your project and create some files and folders within it.

Then, as the instructions say, run the commands

cd vite-project
npm install
npm run dev

npm install will download the node packages specified in the package.json file into the node_modules folder.

Start Dev Server

npm run dev will launch a dev server. Running the dev script starts a program called Vite. Vite’s job is to take your application’s source files, pass them to other programs (including Svelte, in our case) and convert them into the code that will actually run when you open the application in a browser.

If you go to the local URL http://127.0.0.1:5173/, you will see the starter Svelte app that was created automatically.

The app includes a simple interactive counter. When you click the button, the counter increments.

src Folder

If you look at the src folder, you’ll see a few files:

• The assets folder contains static assets, like images
• The lib folder contains Svelte components. In this case, there’s the counter component.
• app.css is a global CSS file for the entire app
• App.svelte is your main app code. In this case, it imports the svelte.svg logo and the Counter component.

<script>
  import svelteLogo from './assets/svelte.svg'
  import Counter from './lib/Counter.svelte'
</script>

• main.js is the main JavaScript file. It loads your app’s CSS (app.css) and your app’s main code (App.svelte). It will also load the app in the HTML element with ID app.

import './app.css'
import App from './App.svelte'

const app = new App({
  target: document.getElementById('app'),
})

export default app

Build the App for Production

When you’re done developing your app (creating components, etc), you can run a build. This will bundle all CSS and JS files into a single CSS and JS file, optimize the bundle (e.g. minify it), and output the production files in the dist (distribution) folder. Run a build with the following command.

npm run build

You’ll see a dist folder created. In this case, with the following files:

If you open index.html, you’ll see it references the two bundled CSS and JS files along with a div element with ID app where the app will be injected.

<!DOCTYPE html>
<html lang="en">
  <head>
    <meta charset="UTF-8" />
    <link rel="icon" type="image/svg+xml" href="/vite.svg" />
    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
    <title>Vite + Svelte</title>
    <script type="module" crossorigin src="/assets/index-e83122bb.js"></script>
    <link rel="stylesheet" href="/assets/index-9ea02431.css">
  </head>
  <body>
    <div id="app"></div>
    
  </body>
</html>

Preview Built App

To preview the app from the built files, run

npm run preview

This will run a local server loading the built files.

There are many different types of website content management systems. This post will explain the different types and include some examples.

No-Code CMS

A no-code CMS is one where the entire website is managed by the CMS. For example, with webflow, all changes to your website must be done in the CMS using a WYSIWYG. You can only edit the HTML of the code widget. All other changes must be done using the UI. If you want to edit the source code of the entire site, you must export the entire site. But, you can’t import any changes you made from the export.

webflow

Database-Driven CMS

A database-driven CMS is one that stores some or all data in a database. For example, WordPress is one such CMS. Web pages and blog post content are stored in a database (usually mySQL). Unlike webflow, however, you can edit the source code of the web page templates. WordPress has a built-in versioning system, so you can see a history of changes made to a page.

• WordPress (Gutenberg Editor)
• Oxygen (Site builder, WordPress plugin)
• Elementor (Page builder, WordPress plugin)

Headless CMS

A headless CMS is one that stores data remotely. Data is integrated with your website using an API. For example, Contentful is one such CMS. You can create custom content models in Contentful. Your website will need to access the data at a particular URL that returns JSON data. Your website will then need to consume that JSON data to process it to show it on a web page, e.g. by using a template engine. Data in Contentful is likely stored in a database.

• Contentful
• Sanity
• Storyblok

Git-Based CMS

A git-based CMS stores data in git version control. Unlike many of the other CMSs, this CMS connects to your git repository as a web developer does. It then shows an interface that allows content editors to make edits to files in the repo. CloudCannon is one such CMS. With CloudCannon, web developers can add the class “editable” to HTML tags with a page’s source code. Then, the content of those tags will be the only elements that content editors can edit.

CloudCannon

FrontMatter CMS

Websites that are built using a static site generator (SSG) often include front matter. Frontmatter can come in different formats. One popular format is YAML. You can have YAML at the beginning of a page’s source code. Or, you can have a separate, standalone .yaml file. YAML contains contain name-value pairs, so they basically represent data. Netlify CMS is one type of CMS that supports editing YAML. It also supports editing pages like blog posts. However, unlike WordPress, where editors can add blocks of content to a page’s body, Netlify CMS requires the blocks of content to be predetermined upfront. So, for a blog post using Netlify CMS, editors can only use one widget to entire the entire body of the page. That widget is a rich text or markdown editor.

Netlify CMS

Structured vs Unstructured Data

The CMSs above can be categorized as CMSs that support structured and unstructured data. CloudCannon supports unstructured data because you can let editors edit specific elements of a complex page by marking those elements as “editable”. YAML and database-backed CMSs support structured data because there is a clear name/value mapping.

My Preference

Since WordPress is the most popular CMS and it allows editors to add blocks of different types of content to the body of a page, I think it’s worth using. However, I don’t like that WordPress keeps data in a database and relies on PHP to build pages before serving them. Also, not all pages need to be edited by non-technical people. In fact, most pages will still be edited by developers who need full access to the source code. To accommodate the needs of both developers and non-developers, I would use GitHub to version all code changes and I would use a static site generator like Eleventy that is integrated with WordPress using the WordPress API. When a WordPress editor saves or publishes a page, WordPress can trigger a save_post() action with a callback function that calls a URL, e.g. a PHP script on Heroku at myapi.com/?postid=123.

function call_commit_script( $post_id ) {
    $remote_url = 'https://www.myapi.com/?postid='.$post_id;
    $access_token = '3&^&2lhl3@#lsjk756'; //some secret password
    $args = array(
    'headers' => array(
       'Authorization' => 'Bearer ' . $access_token,
       ),
     );
    $response = wp_remote_get( $remote_url, $args );
    if ( is_array( $response ) && ! is_wp_error( $response ) ) {
        $headers = $response['headers']; // array of http header lines
        $body    = $response['body']; // use the content
    }
    $post_title = get_the_title( $post_id );
    $post_url = get_permalink( $post_id );
    $subject = 'A post has been updated';
    $message = "A post/page has been updated:\n\n";
    $message .= $post_title . ": " . $post_url\n\n$body;
    // Send email to admin.
    wp_mail( 'webmaster@qualys.com.com', $subject, $message );
}
add_action( 'save_post', 'call_commit_script' );

The PHP script on Heroku could then

1. verify the Authorization Bearer header contains the predetermined access token value, e.g. 3&^&2lhl3@#lsjk756. If it doesn’t, then the request didn’t come from an authorized origin and should be rejected.
2. pull any latest changes from GitHub
3. call the WordPress API to get the page content in JSON format, e.g.
   https://www.mywordpresssite.com/wp-json/wp/v2/pages/123
4. commit the JSON data to GitHub
5. use the static site generator to build the page using the JSON data as the data source

To perform git commands in PHP, this PHP library can be used.

Note: the PHP script on Heroku, e.g. https://www.myapi.com/?postid=’.$post_id, must be secured by SSL/TLS to encrypt the URL and headers so that attackers can’t see the authorization header value.

For allowing non-technical people to

• edit simple YAML files or frontmatter, I’d use Netlify CMS.
• edit complex, structured data, I’d use Contentful.
• edit specific, unstructured content, I’d use CloudCannon.