PHP 8.3 is out! Yay!

This is first time for last years when PHP version upgrade did not brake anything in our projects. All issues were just about some dependencies did not update php version requirements. Now everything is solved so we have PHP 8.3 on our production servers and everything works good. Much faster, much less code problems to solve comparing to 8.0 -> 8.1 or 8.1 -> 8.2 upgrades.

Debian 12 bookworm was released on June 10th, 2023. And next day we tried to use it (in development environment). And it went unexpectedly easy, there were no problems at all. So we started to plan migration for all servers in three steps: development servers, testing (staging) servers, production servers. And today we finished migration of all production servers to Debian 12 bookworm.

From our 20 years of experience of using Debian Linux transition from bullseye to bookworm was smoothest and easiest ever.

Drupal 10 is out for more than two months already! And this is extremely exciting.

Usually it takes from a year to switch to new core version: a lot of modules to upgrade, significant code changes. But Drupal 10 has almost the same core code as previous generation Drupal 9.5. There is a lot of deprecated code removed, there are some new features, but most of 9.5-compatible code runs without any problems with 10.0-core.

As usual long story short: https://packagist.org/packages/mitoteam/jpgraph.

We added patches to avoid PHP 8.2 warnings and some deprecation notices. Our internal tests are passing. Plus we have several confirmations from other users about no problems with PHP 8.2. So version 10.2.1 marked as release candidate and seems to be ready for production usage.

We'll mark it stable as soon as PHP 8.2 released (expected on November 24, 2022).

Long Story Short: PR accepted by PHPOffice/PhpSpreadsheet maintainers. So it will be included in next release.

One day our team has discussion about how to estimate amount of source code for one of our projects. First idea was just calculate total files size, but... A whole bunch of questions arised immediately: there are a lot of vendor code in project, there are binary files, there are images, generated code, multi-byte character encodings and so on.

Our goal was to understand just how many pure code was written by our team. And we want to know characters, lines and printed-pages estimate counts for projects. So this pet-project was born as an answer to this questions.