A development team, working on a static analyzer, has already checked a game engine under Windows. In this check we used a Linux version and the result showed that there were a large number of serious errors. The article covers only the general analysis warnings and only “High” level of severity (there are also Medium and Low levels).
The following warnings are described in detail: typos in the Active() function, memory handling errors, strange code, null pointer dereference, problems with a comma, suspicious conditions, undefined behavior. CryEngine code has also several quite amusing comments of the developers. The article turned out to be quite informative, so it’s worth the time to read it!
Often people ask questions - which programming language is easier, which is the most popular, which one to start learning and so on. In this article we will compare two languages Python and Ruby; their reference implementations CPython and MRI, to be exact.
We took the latest versions of the source code from the repositories (Ruby, Python) for the analysis. There weren’t many glaring errors in these projects. Most of them are related to the usage of macros, although this code is quite innocent from the point of view of the developer. But at the same time, such suspicious fragments that occurred because of copy paste, comparing SOCKET type with null, undefined behavior, storing values to the variables that are already used or null pointer dereferencing are really worth reviewing.
Having analyzed all the warnings of general analysis diagnostics and removed all the false positives, we have come to the following conclusion concerning the error density:
More details about the code fragments where these suspicious code fragments were found:
It’s worth saying that despite these flaws, the code is still of high quality. We should also take such factors into account as the size of the codebase , or the fact that some fragments are erroneous only from the point of view of C++ language and they don’t affect the program in any way. That’s why this analysis may be rather subjective, because previously we haven’t evaluated the error density of these projects. We’ll try to do that in the future, so that we can later compare the result of the checks.
Nowadays a lot of projects are opening their source code and letting those who are interested in the development of it edit the code. OpenJDK is no exception, programmers PVS-Studio have found a lot of interesting errors that are worth paying attention to.
OpenJDK (Open Java Development Kit) - a project for the creation and implementation of Java (Java SE) platform, which is now free and open source. The project was started in 2006, by the Sun company. The project uses multiple languages- C, C++, and Java. We are interested in the source code written in C and C++. Let's take the 9th version of OpenJDK. The code of this implementation of Java platform is available at the Mercurial repository.
During verification, the analyzer found different errors in the project including: copy-paste, bugs in the operation precedence, errors in logical expressions and in pointer handling and other bugs, which are described in detail in this article.
It's always amusing to check a project which is used and maintained by a large number of people. The better and more accurate the code is, the more safely and effectively the program will work. Those bugs we found, are another proof of the usefulness of an analyzer, as it allows the detection of such errors which would otherwise be hard to detect doing simple code review.
Quite often the clash of views happens about the "copyright", but obviously the debaters do not understand where it came from, how it was developed, and what would happen to it later.
This post does not claim to be exhaustive, and it aims simply to describe the basic system regularities.
We will not describe here the distant times of the 13th century and a history of the first rights in the republic of Venice. We begin with what are closer and clearer to us.