Quality processes ensure functional, reliable and optimum development of Scilab software.

Our quality assurance is based in particular on:

Implementation of a new functionality follows a define process

• Specifications drafting: SEP (Scilab enhancement proposal),
• Respect of Scilab coding style for development,
• Writing of unitary tests are written,
• Writing of corresponding help page.

The functionality is integrated only after being reviewed and validated by another developer.

Code quality

• Coding style is defined for all languages used in Scilab (C, C++, Java, XML, Scilab, Makefiles, ...)
• Checkstyle is available for Java under Eclipse
• Code must be documented (comments using Javadoc, Doygen, ...)

As in the case of new functionality implementation, each modification in Scilab code is reviewed and accepted by another developer.

Tests

• Unitary tests (~1300): each new functionality is only integrated if it has a corresponding test and this test execution runs properly under all Scilab supported platforms,
• Non-regression tests (~1900): for each bug fix, a test is written to ensure its proper correction and prevent any future regression,
• Code coverage tools are used to check the tests quality.

Binary version generation

• Each commit into Scilab GIT repository launches a continuous integration process based on Jenkins,
• Daily build of Scilab binary through a nightly build process executes all unitary tests and non-regression tests under all supported platforms,
• Various compilers are supported (Visual Studio, Intel, GCC, Clang, ...).