RFC: How to fix a tricky leak in QFormLayout?

We have a nasty leak in QFormLayout::setWidget(). In setWidget(), we create the QWidgetItemV2 corresponding to the passed QWidget, and pass that on to Private::setItem, which has a bunch of error returns (guard clauses) that do not delete the item, among them negative row index and cell already occupied.

We could easily fix that missing delete, but this function is also used by setLayout(), say, where the item is the layout.

Conceptually deleting that item (= the nested layout) is completely OK, because the user should be able to rely on the form layout to take ownership of the nested layout, without ifs and buts.

But then we have code in tst_qformlayout that breaks. Simplified, it reads:

   QFormLayout layout;
   QHBoxLayout l4;
   layout.addLayout(-1, QFormLayout::FieldRole, &l4);

I guess you spot the problem? If l4 had been added, everything would’ve been peachy: The QHBoxLayout destructor unregistered itself from layout, which does not attempt to delete l4 when it itself is deleted.

But if l4 is not added for some reason, like in the test code above, the fixed code will attempt to delete l4, which is undefined behaviour, of course (double-delete).

I think such broken code deserves to be broken, for the greater good of fixing a resource leak. Esp. since a double-delete should complain much louder than a leak, and the API user can do something about the double-delete while she can’t do anything about the leak (the pointer is not reachable from outside QFormLayout).

I feel uneasy about adding this to 5.6 LTS, though, so I’ll make a minimal fix there, just for setWidget().

What do you think about the larger fix?

Advertisement

Effective Qt @ Meeting C++ 2015: talk is now online

Effective Qt continues as a conference-driven series!

After the guys at Qt World Summit last year managed to mess up the recording of exactly one of the many, many talks there (my Effective Qt one, of course), this is to let you know that Jens Weller of Meeting C++ has uploaded the recording of my Effective Qt talk to his conference channel on YouTube.

I talk about Q_FOREACH vs. C++11 range-for, QLatin1String vs. QStringLiteral, as well as, of course, how bad QList is.

Enjoy!

Wild Beasts

In “Private Practice: Taming Templates”, I had (deliberately, of course :)) hidden a bug in the implementation of KDGenericFactory. I’ve now added a discussion about the bug and its fix.

Private Practice: What’s in a Proxy-Style?

A new session in the private practice, “What’s in a Proxy Style?”, examines QProxyStyle and derives some important caveat when applying a classical Design Pattern.

This is the original-language version of my recent Mythos Proxy-Style article.

Private Practice: Taming Templates

I’ve started a new column, Private Practice, that looks at existing implementations and examines them.

The first Private Practice article, Taming Templates, examines techniques to control template code bloat. There’s nothing particularly earth-shattering in there, but I noticed that a lot of people don’t seem to be aware of the fact that templates can cause a lot of duplicated code, so I thought I’d write about it.

Even if they haven’t been released under the Private Practice banner, the following articles also fit that category well:

• Mythos Proxy-Style [[Update 2011-11-24: EN version now available as “What’s in a Proxy-Style?”]]
• Dem Elfenbeinturm entreißen

Heise Developer: Mythos Proxy-Style

For you German speakers out there, my fifth Heise Developer article, “Mythos Proxy-Style: Wie eine verspätete Klasse neues Licht auf ein bekanntes Entwurfsmuster wirft” just went live on Heise’s site.

It examines the implementation of QProxyStyle in Qt 4.6 and, from that analysis, derives an important caveat regarding the applicability of a certain well-known design pattern.

This is the German translation of an unpublished English-language original, which I will publish here after a grace period.

From the article:

In der Qt-Gemeinde existiert ein Mythos: “QProxyStyle”. Auf der einen Seite stehen Entwickler wie David Faure, seines Zeichens KDE-Urgestein und geschätzter Kollege des Autors, die behaupten, ein QProxyStyle sei unmöglich. Auf der anderen Seite stehen Entwickler von Qt-Erweiterungen, die einen QProxyStyle entwickelt haben wollen. Eine Spurensuche im Märchenwald deckt interessante Komplikationen bei der Anwendung eines bekannten Entwurfsmusters auf.

Heise iX: QtQuick article, KDAB whitepaper; Qt Containers update

Together with my fellow KDABians Stephen Kelly and Volker Krause, I wrote a whitepaper on QtQuick usage in the KOMO project. From the whitepaper:

Developing a serious and large application from scratch, or from an existing desktop application, presents a different set of challenges compared to your normal smart-phone ‘app’. Portability to different platforms, and long-term maintainability in the face of ever-changing mobile target platforms suddenly come into play. This paper attempts to summarize the experience we gathered using the new declarative GUI framework, QtQuick, in this project.

Unfortunately, recent events (read: Tizen) have already invalidated much of the content in the supplemental box on the future of Qt, but that’s the bane of everyone writing about Qt’s future these days…

Get the English version (with slightly more details) at http://www.kdab.com/whitepapers, or read the German version (with better editing) in the current print issue of German computer magazine iX: http://www.heise.de/ix/inhalt/2011/11/114/.

In other news: I’ve updated “Understand the Qt Containers” with a discussion of the differences between STL and Qt containers regarding size types, associative container insertions, and error handling. I also mention QT_STRICT_ITERATORS now.

Early-Bird Registration open for “In-depth Multi-Threading with Qt” @ DevDays

I’ll be giving a two-day “In-depth multi-threading with Qt” training just after the DevDays in Munich, ie. Oct 27th-28th. The training language will be English. The training material is the one from the usual Qt training, but “seriously expanded” to fill the two-day schedule.

From the announcement:

This two-day training teaches how to use multithreading with the primitives provided by Qt. Participants will gain insight into multithreading problems in general, and how they pertain to Qt programs in particular.

Examples include how to offload work from the GUI thread, how to increase throughput to the maximum in your application, and optimal strategies for communication between threads.

The training covers topics such as cross-thread signal/slot connections, QThreadPool, QObjects and multithreading, QtConcurrent, QFuture, and Qt atomic operations.

Participants are expected to have a working knowledge of C++ and Qt. Prior experience with multithreaded programming is recommended, but not necessary. The course will start with a one hour multithreading refresher.

To book, visit http://www.kdab.com/schedule.

Until October 1st, you’ll save 200€ on the early-bird offer. Just use d94ec6fd in the Voucher(s) field.

Hope to see you there!

Understand the Qt Containers: The Data

This is just a short heads-up to those that have asked for quantitative measurements. I’ve begun updating “Understand the Qt Containers” with a section on data.

As of now, the memory and append performance is in. I’ll re-run the iteration test over the night once more (I’ve detected an anomaly that I’d like the verify first), and re-add them again, when done.

Effective Qt: Understand the Qt Containers

The third of my Effective Qt columns, “Understand the Qt Containers“, just went live on my site.

If you know me, you will know what awaits you there 🙂 I hope I have not disappointed you. That said, this is just a first part. I plan to eventually extend it to cover more containers (currently, just QVector and QList), and give more data and assembler code for the containers I’ve already described.

These are the guidelines this item covers:

• Remember that QList has nothing to do with std::list, and QSet has nothing to do with std::set.
• Be aware of the two Qt container APIs, Qt-ish and STL-compatible, and avoid mixing their use in the same project.
• Familiarise yourself with the STL containers and the additional features they offer.
• Prefer member-swap over assignment wherever possible to express move semantics. Use C++11 rvalue move semantics if the class and the compiler support them.
• [[EDIT 2011-08-25: new]] Always use constBegin()/constEnd() (cbegin()/cend()) when assigning the result to a const_iterator.
• Avoid modifying a container (remove/insert elements) while iterating. If you have to, use an STL algorithm, such as one of the std::remove() family.
• Prefer the STL-compatible iterators over the Java-style ones.
• If you do use the Java-style iterators, avoid using the mutating ones.
• Prefer to use const references as the first argument of Q_FOREACH, unless the element type is customarily passed by value.
• Familiarise yourself with BOOST_FOREACH, and the additional features it offers.
• Declare your enums and QFlags as Q_PRIMITIVE_TYPE if there’s a chance they will be held in Qt containers.
• Declare your value-types Q_MOVABLE_TYPE if there’s a chance they will be held in Qt containers.
• Don’t change the classification of a type if you need to maintain binary compatibility.
• Prefer vector (std or Q) over QList.
• Avoid QList<T> where T is not declared as either Q_MOVABLE_TYPE or Q_PRIMITIVE_TYPE or where sizeof(T) != sizeof(void*) (remember to check both 32 and 64-bit platforms).
• [[EDIT 2011-08-16: new]] Avoid using vectors of types for which Qt APIs customarily use QList, and for which QList is not inefficient.

Hope you enjoy!