I’ve not posted in a few weeks, and rather than itemize specifics I’m just going to give a few high level summaries of recent activities.
The PHP 7.3 transition was my main focus the past month or so, involving rebuilds, re-tests and misc. fixes to a number of packages. Last week saw the completion of this endeavor with the removal of php7.2 from Eoan. PHP 7.3 brings some refinements to the language, but is an incremental release and not expected to cause too much disruption to developers and users; but even so it may be worth reviewing the 7.2 to 7.3 migration guide.
PHP version 7.4 is scheduled for release later this year, if you think it should (or should not) be included in Ubuntu 20.04, I’d love to hear rationale one way or another.
The server team has a merge report to indicate merges worth doing, however it’s suffered from both having too much information and not enough. As mentioned in past reports, over the past few months I’ve implemented the internal logic to calculate merge opportunities, and I’ve been manually sending text reports to the ubuntu-server@ mailing list.
The next phase of development for this is to turn this into a web page. The server team has another report implemented with JQuery, that allows sorting by columns and presents the information in a tabular format that feels more useful than the current merges page. Bileto also has an interesting UI, implemented via Angular.js, that I started looking at last week as well.
Apart from tooling, at the Paris sprint I also accepted the task of planning out our merges strategy for 20.04. This involves studying the package listing generated by the above tooling, and helping the team plan out the timing for various merges from Debian and upstream.
I’ve been digging into git-ubuntu and learning its internals the past few months, with the intention of lending a hand on development going forward. At the Paris sprint Robie and I took the opportunity to do some knowledge transfer and bat around some ideas. A goal of mine is to help improve the maintainability of the project via testsuite improvements, making the importer service more robust and reliable, and assist with retiring some of the accumulated tech debts. I am in particular interested in finding ways to make the developer tools more efficient and productive for heavily used workflows, such as for Debian package merges and bug SRUs.