Wiki Reorganisation

A group of us at the CERN-based cross-team sprint are attempting to tame the wilderness of the KDE wikis – at least, TechBase and Community.

A little bit of history is needed here. Originally, TechBase was the only wiki, and it quickly became a dumping ground for pretty much everything. At some point, the other two wikis (Community and UserBase) were created so we could separate things out a bit, and people wanting tutorials for how to create Plasmoids, for example, wouldn’t be overwhelmed with meeting notes from the Plasma team.

While UserBase has a clear separate mission to the other wikis, the line between TechBase and Community has always been a bit fuzzy. The pages that were still on TechBase since before Community existed and were never removed didn’t help with this.

The first thing we did at the sprint, therefore, was clarify this line. The distinction we came up with was about target audience:

• TechBase is for “developer end-users” – people who want to build software on top of our libraries, or extend our applications with plugins, or write Plasmoids, or use the DBus interfaces of our software. These may be KDE developers or people not connected to KDE at all (at least, not yet!).
• Community is for the KDE community to talk to itself (and to people who want to join). This not only includes things like TODO lists and meeting notes, but also things that are specific to hosting a project on the KDE infrastructure (like how to integrate with our translation setup) and community organisation things like policies.

We have also spent quite some time discussing the details of how to organise both TechBase and Community with this distinction in mind, and are now working on moving the information around.

The work we’re doing at this sprint is possible thanks to our hosts, CERN, and people like you donating to the KDE e.V., who are helping to cover the costs of attendees to travel and stay here. Please consider joining in.

Advertisement

Library versioning

Do you maintain or contribute to a library, especially a KDE library (other than the Frameworks)? If so, I have a question for you.

First, some background. These days, the norm for libraries is to follow some variation of semantic versioning, where the version number follows the scheme x.y.z (eg: 1.4.3) and it should always be safe to upgrade the library to version with a higher y (minor) component as long as the x (major) component is the same. Qt follows this scheme, for example.

This scheme doesn’t really allow for alpha versions of software, though – the alpha releases for the (x+1) major version can be incompatible with both version x and with the final releases of version (x+1). The Semantic Versioning scheme prescribes using versions like x.y.z-alpha1, but this breaks the nicely numeric scheme we had, and not all tools can cope with that.

KDE Frameworks (and, in the past, kdelibs) approached this by defining some arbitrarily high minor version (typically 90+) to indicate pre-releases for the next major release. So the pre-release Frameworks were numbered like 4.90.1.

So where is all of this going? Well, CMake provides a helpful function to write package version information files that allow find_package() calls to only find compatible versions. So if you use the SameMajorVersion scheme, find_package(Foo 4.3) will find Foo 4.3.0 and Foo 4.5.2, but not Foo 4.2.5 or Foo 5.1.1. However, if project Foo uses the “high minor = next version prerelease” scheme, it will also find Foo 4.90.1, which is not compatible with Foo 4.3.

I wrote a SameMajorVersionWithPreleases scheme that would deal with this for Extra CMake Modules, but never committed it because I wasn’t sure if it would get used (KDE Frameworks doesn’t need it because we put KF5 in the package names). My question is: would your project find this useful? In other words, do you contribute to (or know of) a CMake-based project that uses the “high minor = next version prerelease” scheme I described above? If so, please comment!

CMake Library Tutorial

Interested in writing or maintaining a KDE Framework? Getting the build system right for a Framework, or indeed any other library, is not trivial. I’ve just finished writing a tutorial to help you do just that, though. It can be found at the CMake tutorials page on TechBase, and guides you through (an idealised version of) the KArchive buildsystem.

Comments, feedback and questions on the tutorial are welcome – here, by email or on the discussion page of the tutorial.

Akademy 2015

I’m just finishing up a week spend at Akademy, the annual conference of the KDE community.

We started last Friday with the AGM of the KDE e.V., the legal entity that deals with the financial and other legal aspects of KDE.

After that, we had two days of talks, and then from Monday onwards we’ve had “birds of a feather”, or “BoF”, sessions where people working on or interested in similar things get together to discuss things face-to-face.

It’s been great fun, and I’ve really enjoyed seeing people I’ve met before again and meeting people I’ve only known through IRC and blog posts or have never even come across.

The organising team have done a fantastic job: we’ve had free busses running from our accommodation to the venue, video recording of talks (which I’m sure someone will post about soon), easy to access food, two parties and people always on-hand to provide information.

There have been announcements of new technologies, talks about community things like how to write a vision statement or what artists are doing with Krita, technical talks about things like how to optimise your program, talks about development methodologies and more.

Personally, I’ve taken the opportunity to start improving our CMake documentation, including writing and planning new tutorials. I’ve also been working on some Extra CMake Modules things at the request of David Faure, looking at accessibility things with Frederik Gladhorn and trying out the upcoming release of Plasma.

I’ve had great fun taking lots of photographs this Akademy, and have been uploading them all to Flickr.

I’m not the only one, either – you can find other photo sets linked from the Akademy wiki.

I’m about to catch a flight home, but I very much hope I can make next year’s Akademy as well.

Buildsystem BoF at Akademy

I’ve just booked a room for Tuesday afternoon at Akademy for a couple of CMake / extra-cmake-modules BoF sessions.

The first part is for people who want support hacking the buildsystem of their own projects. Not sure of the best way to migrate to extra-cmake-modules as part of your shift to KF5? Getting “deprecated” or other CMake warnings you want to resolve? Can you think of something really useful the buildsystem could do in your project, but you don’t know how to do it? Come along, and I (and hopefully others) will be there to give you a hand. Likewise, if you feel comfortable writing stuff in CMake, and you want to help other developers, come along to lend a hand – you don’t need to be an expert!

The second part is for people who want to contribute to extra-cmake-modules. If you have some CMake code in your project you think other projects will find useful, or you have ideas for improvements in the existing modules, come along and pitch in.

Help wanted: Photoshop

As part of KDE Frameworks, we have a collection of QImage plugins that allow Qt applications to read various types of image files not natively supported by Qt. I’ve recently overhauled the one that reads Photoshop images (PSD files) but, without access to Photoshop, I have no way to create images that test the code. The one test image I have, I created using the Gimp, which is less than ideal when I want to test compatibility with Photoshop itself.

So, if you have access to Photoshop (Photoshop Elements might be sufficient, I’m not sure) and would be willing to make some test images (in pairs – one PSD, one PNG) of various specific configurations and in various save formats, please do get in touch (alex dot merry at kde dot org, alexmerry on IRC, or just post a comment here).

To the extent that the simple images I require would be copyrightable, you would need to be willing to license them under the LGPLv2+ or a permissive license (CC-BY-SA would be fine, for example).

New installation variables

I noticed a while ago that the CMake installation variable names used by KDE projects (defined either in kdelibs or KDEInstallDirs in extra-cmake-modules) are inconsistent with what GNUInstallDirs, for example, uses. The traditional variable names are things like BIN_INSTALL_DIR and CMAKECONFIG_INSTALL_PREFIX, while GNUInstallDirs defines things like CMAKE_INSTALL_BINDIR and CMAKE_INSTALL_DOCDIR.

In an effort to make the installation variables more consistent and compatible with GNUInstallDirs (for the ease of packagers and developers alike), I introduced GNUInstallDirs-style variables to KDEInstallDirs – this included the variables defined by GNUInstallDirs and other ones beginning CMAKE_INSTALL_. These were kept in sync with the old-style variable names.

Both of these styles of variables have namespacing problems: the ones starting CMAKE_ tread on CMake’s namespace (so CMake could potentially define one of those variables that doesn’t appear in GNUInstallDirs to have a different meaning), and the old-style variables are very generic and could conflict with another CMake module. Neither of these are likely to be an issue in reality, but it is good practice to respect namespaces in third-party modules.

KDEInstallDirs now primarily defines variables starting with KDE_INSTALL_ (which corresponds with the module name). The old-style names and all the CMAKE_INSTALL_ names that do not also exist in GNUInstallDirs are deprecated, and can be disabled entirely (see the documentation). It is possible to disable the variables that appear in GNUInstallDirs as well.

This appeared in extra-cmake-modules 1.6.0, but that version had some flaws (for example, it would break if you used both GNUInstallDirs and KDEInstallDirs in the same project, which was a major oversight), so you should use 1.6.1 instead.

There is a script in the kf5 directory of the kde-dev-scripts.git repository to port your CMakeLists.txt files to the new variables: cmakelists_install_vars.pl. Just run it in a directory with a CMakeLists.txt file.

return EBUSY;

I just want to apologise to those of you who are waiting on me for Review Requests, etc. Last month was quite busy for me, and this month is even more so – all good things, such as friends getting married, but all time-consuming. On top of that, Akonadi and the server I use for KDE-related email have decided they aren’t talking to one another, which makes it that bit harder to deal with (due to having to log into, and then deal with the interface of, webmail).

I’m expecting life to calm down considerably in October, so hopefully normal server should resume then.

Qt5 Pretty Printers for GDB

If you’ve ever debugged a Qt program in GDB and been frustrated when attempting to print a QString gives you an unhelpful result like

$1 = {d = 0x6e7c40}

then kde-dev-scripts has some useful tools for you. For some time now, there has been the kde-devel-gdb script, which you can include in your ~/.gdbinit. This gives you commands like “printq4string” and “printq5string” which will dump the actual string contents of a QString.

However, GDB 7 has the ability to add pretty-printers written in Python. This means that you can control how a value is output when you use “print” and when backtraces are printed. So “print someQString” will actually write out the string for you.

The gdb directory of kde-dev-scripts now has some of these pretty-printers for Qt5. Currently, it covers most of the main value classes and collection classes in QtCore (QByteArray, QString, QStringList, QList, QMap, etc). QHash and QSet are the most notable ones that are missing.

To use these pretty-printers, you either need to source the “load-qt5printers.py” file in your GDB session, or use the “create-qt5-autoloaders.py” script to generate files that will allow GDB to auto-load the printers when it loads the Qt5 libraries.

I plan to eventually submit these to Qt, so that pretty-printing of Qt value classes will work out of the box. For now, do let me know if you run into any issues with them (don’t forget to compile your application with debugging symbols!), and feel free to extend what’s there, or use it as a template for some Qt4 pretty-printers.

CMake Conditionals

[edited 17/04/14]

CMake has many wonderful features. However, CMake’s if() command handles variables in a way that can often trip people up. In particular, comparison operators like STREQUAL try to interpret their arguments as variable names, and only treat them as actual strings if there is no matching variable. Combined with the fact that basically any string you can write in a CMake file is a valid variable name, this can lead to some subtle bugs.

This script highlights all the potential pitfalls I could think of. The “FATAL_ERROR” branches are the ones that will not be taken, but many of them are ones you might expect to be taken if you either didn’t have a solid grasp of how if() parses its arguments or if you weren’t aware of what variables had been defined.