People of the Internet,

While working on keystroke events, I realized my improvements to the event.change property were still inconsistent for certain Unicode characters. This led me to delve into code units, code points, graphemes, and other cool Unicode concepts. I found this blog post to be very enlightening.

Here’s an update on my progress over the past two weeks:

MRs merged:

• event.change : The change property of the event object now correctly handles Unicode, with adjustments to selStart and selEnd calculations. !MR998
• cursor position and undo/redo fix : Fixed cursor position calculations to account for rejected input text and resolved merging issues with undo/redo commands in text fields. !MR1011
• DocOpen Event implementation : Enabled document-level scripts to access the event object by implementing the DocOpen event. !MR1003
• Executing validation events correctly : Fixed a bug where validation scripts wouldn’t run after KeystrokeCommit scripts in Okular. !MR999
• Widget refresh functions for RadioButton, ListEdit and ComboEdit : Added refresh functions as slots for RadioButton, ListEdit, and ComboEdit widgets, aiding in reset functionality and script updates. !MR1012
• Additional document actions in Poppler : Implemented reading additional document actions (CloseDocument, SaveDocumentStart, SaveDocumentFinish, PrintDocumentStart, PrintDocumentFinish) in the qt5 and qt6 frontends for Poppler. !MR1561 (in Poppler)

MRs currently under review:

• Reset form implementation in qt6 frontend for Okular : Working on the reset form functionality in Okular, currently focusing on qt6 frontend details. !MR1564 (in Poppler)
• Reset form in Okular : Using the Poppler API to reset forms within Okular. !MR1007
• Fixing order of execution of events for text form fields : Addressing the incorrect execution order of certain events (e.g., calculation scripts) and ensuring keystroke commit, validation, and format scripts are evaluated correctly when setting fields via JavaScript. !MR1002

For the coming weeks, my focus will be on implementing reset forms, enhancing keystroking and formatting scripts, and possibly starting on submit forms. Let’s see how it goes.

See you next time. Cheers!

Like last year, the KDEPIM team convened in Toulouse to hold it’s traditional spring sprint. Unlike last year this time it was late spring, almost made it into summer. This time we’ve been hosted by Étincelle Coworking in one of their spaces. It was fairly nice. Lots of space, comfortable, well situated in the center. I definitely recommend.

The Warm-up

Some of the team (namely Carl and Volker) arrived early before the official start. I’m pretty sure Carl made it early to have lunch in the infamous Cake Place™ (at last). Of course a stupid amount of cake slices was involved. And after a short stroll we got to the meeting space we would use the rest of the time.

We started preparing the work board and finished collecting ideas as Dan arrived.

Day 1

Once again, we had two types of tasks: discussions and actual hacking. During that first day we mostly focused on the conversations but did some hacking on the side. Since the attendance was a bit low this year, we didn’t go for break out groups and did them all as a single group.

Said discussions covered mostly the following topics:

• Retiring some old software or now unused resources
• Trying to devise plans to remove some duplication of efforts (did you know we have several copies of the time tree parser?)
• Some minor stability issues were discussed to have a plan to solve them
• How to improve the usage of Akonadi on mobile which revolved mostly around some better partial syncing and some tight coupling with widgets at places
• Devising a strategy to have some end to end testing of the supported resources
• How to have better tag support and indexing in Akonadi (which opens the door to really cool features so stay tuned)
• And finally, we spent quite some time discussing how to improve our little PR efforts and how to recruit new contributors

Of course, like last year we’re contrained by the size of the team and so need to use smartly everyone’s time. Again no grand plans, but some of the items we plan to deliver during the year could have very nice consequences. We’ll focus on a better quality of life both for users and contributors.

Day 2

We already started some of the technical tasks the previous day, but Sunday was mostly devoted to them. We had a break for a nice brunch of course but still people worked on their branches around it.

As far as technical tasks, we got the following things done already:

• Some obscure and broken feature around notes in KMail has been removed
• KJots notes can now be imported into MarkNotes giving an upgrade path
• Akonadi tag discovery in caldav resources is a thing
• The Akonadi brand should finally not appear anywhere in the applications anymore
• A way to clean up the data coming from retired resources is now in place

But there’s more in the works, namely:

• the completion of the infrastructure for QtQuick based config dialogs
• the switch to KConfigDialogManager where is makes sense (instead of using an old PIM specific legacy system)
• the rewrite of the indexing system (which should improve performance and have instant indexing of payloads)
• the migration agent to apply tag discovery also to old and already existing items
• the move of KMime to KDE Frameworks is getting prepared
• the IMAP resource configuration system is in the works to be able to jump on the QtQuick infrastructure (it’s one of the more complex ones with lots of QtWidgets dependencies)
• the KDAV test suite is being completed for better coverage

The mentioned tasks and all the ones created during the conversations have been added to our Technical Roadmap on Gitlab. The work is still on going for some of them.

What now?

The most noticeable is probably one thing we acted on immediately after our discussions. We added milestones in the PIM group. Indeed, it’s probably not always easy for people wanting to join to know which are the bigger on going goals in the KDEPIM team and what to work on to help.

This is now encoded through those milestones. It’s the early days for them so they need to be improved and more tasks added, but at least this gives an idea of what has higher priority for now.

We plan to have a BoF at Akademy 2024 to be announced later when the schedule for the BoFs actually opens. There we’ll plan to focus at discussing the experience of contributors and see how it can be improved. We’ll also present our milestones. We might sign you up on tasks there, don’t hesitate to show up, we want to hear from you.

Last but not least

If you want to know more or engage with us, please join the KDEPIM matrix channel! Let’s chat further.

Also, I’d like to thank again Étincelle Coworking and KDE e.V. to make this event possible. This wouldn’t be possible without a venue and without at least partial support in the travel expenses.

Finally, if you like such meetings to happen in the future so that we can push forward your favorite software, please consider making a tax-deductible donation to the KDE e.V. foundation.

This year again I participated to the KDE PIM Sprint in Toulouse. As always it was really great to meet other KDE contributors and to work together for one weekend. And as you might have seen on my Mastodon account, a lot of food was also involved.

Day 1 (Friday Afternoon)

We started our sprint on Thursday with a lunch at the legendary cake place, which I missed last year due to my late arrival.

We then went to the coworking space where we would spend the remaining of this sprint and started working on defining tasks to work on and putting them on real kanban board.

To get a good summary of the specific topics we discussed, I invite you to consult the blog of Kevin.

That day, aside from the high level discussion, I proceeded to port away the IMAP authentification mechanism for Outlook accounts away from the KWallet API to use the more generic QtKeychain API. I also removed a large dependency from libkleo (the KDE library to interact with GPG).

Day 2 (Saturday)

On the second day, we were greated by a wonderful breakfast (thanks Kevin).

I worked on moving EventViews (the library that renders the calendar in KOrganizer) and IncidenceEditor (the library that provides the event/todo editor in KOrganizer) to KOrganizer. This will allow to reduce the number of libraries in PIM.

• Remove KOrganizer plugins from kdepim-addons and move them instead to KOrganizer
• Do not use a complex plugin system for handling calendar invitation and instead provide the incidence editor as a binary and call it instead. (MR 1, MR 2 and MR 3)

For lunch, we ended up eating at the excellent Mexican restaurant next to the location of the previous sprint.

I also worked on removing the “Add note” functionality in KMail. This feature allow to store notes to emails following RFC5257. Unfortunatelty this RFC never left the EXPERIMENTAL state and so these notes were only stored in Akonadi and not synchronized with any services.

This allow to remove the relevant widget from the pimcommon library and the Akonadi attribute.

I also started removing another type of notes: the KNotes app which provided sticky notes. This application was not maintained anymore, didn’t work so well with Wayland. If you were using KNotes, to make sure you don’t loose your notes, I added support in Marknote to import notes from KNotes.

Finally I worked on removing visible Akonadi branding from some KDE PIM applications. The branding was usually only visible when an issue occurred, which didn’t help with Akonadi reputation.

We ended up working quite late and ordering Pizzas. I personally got one with a lot of cheese (but no photo this time).

Day 3 (Sunday)

The final day, we didn’t had any breakfast :( but instead a wonderful brunch.

Aside from eating, I started writing a plugin system for the MimeTreeParser which powers the email viewer in Merkuro and in Kleopatra. In the short term, I want to be able to add Itinerary integration in Merkuro but in the longer term the goal is to bring this email viewer to feature parity with the email viewer from KMail and then replace the KMail email viewer with the one from Merkuro. Aside from removing duplicate code, this will improve the security since the individual email parts are isolated from each other and this will makes it easier for the email view to follow KDE styling as this is just normal QML instead of fake HTML components.

I also merged and rebased some WIP merge requests in Marknote in preparation of a new release soon and reviewed merge requests from the others.

Last but not least

If you want to know more or engage with us, please join the KDE PIM and the Merkuro matrix channels! Let’s chat further.

Also, I’d like to thank again Étincelle Coworking and KDE e.V. to make this event possible. This wouldn’t be possible without a venue and without at least partial support in the travel expenses.

Finally, if you like such meetings to happen in the future so that we can push forward your favorite software, please consider making a tax-deductible donation to the KDE e.V. foundation.

This is the first blog post of my GSOC journey. I will be sharing my works and experiences here. Stay tuned for more updates. In this blog, I’ll be sharing my experiences of the first two weeks of GSOC, what are the works I did, what are challenges I faced and how did I overcome them ( Did I really overcome them :P ).

On my first week I tried to understand the codebase of discover first, via doing small changes. So, the first thing I added was a new way of verification of snap publishers which is officially supported by snapcraft. Snapcraft has two tiers of verification:

This is the release schedule the release team agreed on

  https://community.kde.org/Schedules/KDE_Gear_24.08_Schedule

Dependency freeze is in around 4 weeks (July 18) and feature freeze one
after that. Get your stuff ready!
 

Let’s go for my web review for the week 2024-24.

Microsoft Will Switch Off Recall by Default After Security Backlash

Tags: tech, microsoft, privacy

Unsurprisingly they had to adjust under the pressure. The most blatant issues might be gone, it is still a bad idea at its core.

https://www.wired.com/story/microsoft-recall-off-default-security-concerns/

AI chatbots are intruding into online communities where people are trying to connect with other humans

Tags: tech, ai, machine-learning, gpt, criticism, ethics

Chatbots can be useful in some cases… but definitely not when people expect to connect with other humans.

https://theconversation.com/ai-chatbots-are-intruding-into-online-communities-where-people-are-trying-to-connect-with-other-humans-229473

Malicious VSCode extensions with millions of installs discovered

Tags: tech, vscode, security, ide

How trustworthy are the extensions you get in your editor or IDE? I’d expect most marketplaces to not be well harmed against such attacks.

https://www.bleepingcomputer.com/news/security/malicious-vscode-extensions-with-millions-of-installs-discovered/

HTTP/3 needs us (and other people) to make firewall changes

Tags: tech, http, quic, firewall

Good reminder that firewalls need to be adjusted for proper HTTP/3 support.

https://utcc.utoronto.ca/~cks/space/blog/sysadmin/HTTP3AndOurFirewalls

HTTP/3 in curl mid 2024 | daniel.haxx.se

Tags: tech, http, quic

Interesting status report about HTTP/3 support in curl. Shows quite well the various alternatives and how special HTTP/3 can be.

https://daniel.haxx.se/blog/2024/06/10/http-3-in-curl-mid-2024/

What is PID 0? · blog.dave.tf

Tags: tech, unix, linux, kernel, system, processes

Interesting deep dive in where the PIDs seen in user space come from. And also yes, there is something matching PID 0 which can be traced back to early UNIX systems.

https://blog.dave.tf/post/linux-pid0/

Scan HTML faster with SIMD instructions: Chrome edition – Daniel Lemire’s blog

Tags: tech, cpu, performance, SIMD

SIMD keeps providing interesting performance boosts for parsing work loads.

https://lemire.me/blog/2024/06/08/scan-html-faster-with-simd-instructions-chrome-edition/

Rolling your own fast matrix multiplication: loop order and vectorization – Daniel Lemire’s blog

Tags: tech, c++, compiler, performance, matrix

The ordering used for matrix multiplications definitely matters.

https://lemire.me/blog/2024/06/13/rolling-your-own-fast-matrix-multiplication-loop-order-and-vectorization/

You’ll regret using natural keys

Tags: tech, databases, design

Good advice on designing your database tables. The comments are good too, they allow to complete the picture.

https://blog.ploeh.dk/2024/06/03/youll-regret-using-natural-keys/

Brain dump – Pagination for database objects

Tags: tech, backend, databases

The right and wrong approaches for paginating results coming from a database.

https://www.n16f.net/blog/pagination-for-database-objects/

Optimal SQLite settings for Django

Tags: tech, django, databases, sqlite

Little and to the point reference on safer SQLite use. I should check if some of this would apply or is used by Akonadi as well.

https://gcollazo.com/optimal-sqlite-settings-for-django/

the Gilbert–Johnson–Keerthi algorithm explained as simply as possible

Tags: tech, geometry, mathematics, algorithm

Need to know if two shapes overlap? Good explanation of an elegant algorithm to do it.

https://computerwebsite.net/writing/gjk

Feynman’s Razor - by Defender of the Basic

Tags: tech, documentation, communication, gui

Nice reminder that even though we try to make things simpler to understand to people, there is a point where we can go too far.

https://defenderofthebasic.substack.com/p/feynmans-razor

Foreword for Fuzz Testing Book

Tags: tech, fuzzing, tests, history

Ever wondered where fuzz testing is coming from? This is an important bit of history.

https://pages.cs.wisc.edu/~bart/fuzz/Foreword1.html

Post-Architecture: An Open Approach to Software Engineering

Tags: tech, software, architecture

Indeed this is not for any environment and projects. So take it with a grain of salt. That said, I think this piece has a core truth to it which is more general. Software architectures shouldn’t be considered as something fixed as soon as they are planned, they need to be validated through use and to be prepared to evolve over time as needed.

https://arendjr.nl/blog/2024/06/post-architecture/

Bye for now!

After three months, KDE’s Kirigami tutorial has been ported to Qt6.

In case you are unaware of what Kirigami is:

• Qt provides two GUI technologies to create desktop apps: QtWidgets and QtQuick
• QtWidgets uses only C++ while QtQuick uses QML (plus optional C++ and JavaScript)
• Kirigami is a library made by KDE that extends QtQuick and provides a lot of niceties and quality-of-life components

Strictly speaking there weren’t that many API changes to Kirigami. The most notable changes were:

Hi! It has been over two weeks since the coding period began. In this blog post, I will provide a brief summary of my work during the first two weeks.

After spending some time reviewing the code, I decided to start by refactoring the existing code related to ASS format subtitles. This has two main goals: first, to enable Kdenlive to read as much information as possible from ASS subtitles (specifically, the features supported by libass) and load it into the memory; and second, to ensure that Kdenlive can save all this information back to the file. Since SubtitleModel already contains a significant amount of code for ASS format subtitles, my work mainly involved refining and expanding upon this existing code while maintaining compatibility.

So far, I have accomplished the following:

• Added initial support for reading and saving embedded fonts in ASS subtitles
• Optimized the storage method for subtitle styles
• Migrated from V4Style to V4+Style

Tasks still to be completed:

• Modify m_subtitleList to accommodate more information.
• Write unit tests for each feature

Once these steps are completed, we will have more comprehensive support for ASS format subtitles, marking the end of this phase of the ASS code refactoring. The next focus will be on refactoring the functionality for modifying subtitle styles. These two parts will be my primary focus for the next two weeks. Stay tuned!

A strange fact recently came into my purview – many users and developers don’t know what Open Source is.

Some time ago, there was a controversy regarding the Floorp web browser (a fork of Firefox) going closed source.

The cause for this was that some company forked Floorp and made a browser based on it. I’ll not comment on the irony that the author of a fork of a project complained that someone forked his project.

Obviously, this triggered a storm of negative reactions on quite a few platforms where fans of Free Software / Open Source software hang out.

The developer responded that this is just temporary, and that the browser will soon be Open Source again. After a while, the repositories became public again and all was fine.

The developer said that Floorp is again Open Source, the angry mob said good, the Floorp browser is again Open Source. And every discussion about Floorp got a plethora of comments that people should stop complaining, that Floorp is Open Source, and that the previous situation was just a misunderstanding.

Open Source

The term Open Source is well defined at opensource.org.

Making the source code of a program publicly available is not enough for something to be Open Source. Having an army of people saying that something is Open Source, is not enough for something to be Open Source.

If a license that the code is published under doesn’t conform to the criteria published on opensource.org, a program is not Open Source. A program whose license contains the following sentence, is, by definition, not Open Source:

You may not use or distribute this Software or any derivative works in any form for commercial purposes. Examples of commercial purposes would be running business operations, licensing, leasing, or selling the Software, or distributing the Software for use with commercial products.

– Floorp private components/LICENSE

This is strangely worded as it looks like you can not use the Floorp web browser to access, for example, a work e-mail account as that would be using it for commercial purposes. This is likely not what the license author intended – the intent of the license is to disallow creating commercial products by forking Floorp.

While it is a valid desire of the author not to have somebody else profit from his work, it is the thing that makes the Floorp web browser not Open Source.

You can call it ‘source available’, you can call it ‘fair code’ but you can not call it Open Source.

Update: A few days after this blog was published, Floorp moved the code from the private submodule to the main repository. So, it should be again Open Source. Let’s hope it remains like that.

Redefining Open Source

This blog post should have been written when the Floorp thing happened, but I thought this is just a random incident not worth the extra attention.

It seems I was wrong. It is something that people should start paying attention to.

A lot of people – developers and users alike, intentionally or not – misuse the term Open Source, and some of them like FUTO even go that far to redefine it and create their own incompatible The Open Source Definition that will suit their own purpose.

Open Source Confusion Cases

Now, the main purpose of this post isn’t for me to let off some steam, but to share a great project started by Dan Brown of attempting to find and list all projects which claim to be Open Source while their licenses say otherwise.

It can be found on his Github account.

Share your views to FSFE

Albert pointed out that FSFE is also interested in this topic:

The FSFE is looking for examples and thoughts about openwashing if you feel like it I’m sure they’ll welcome your input mastodon.social/@johas/112524760073638652

We’re already on Week 3 of the coding period of Google Summer of Code! As a reminder, I’m adding Python support to a few KDE Frameworks. During the first two weeks I added support for KWidgetsAddons, and it’s now almost finished except for two widgets whose bindings aren’t generated properly and don’t compile.

I also wrote (with the help of my mentor Carl) the necessary CMake code to build the library. That part is probably going to end (hopefully) in extra-cmake-modules so it can be used by anyone easily. The plan is to eventually submit each of the bindings to their repository so it’s easier to keep the C++ libraries in sync with their Python bindings.

This week I will be adding some examples for KWidgetAddons now that it’s on a usable status. If you want to see how it’s going, you can take a look at the code.