Skip to content

Latest commit

 

History

History
130 lines (83 loc) · 3.08 KB

HACKING-QA.md

File metadata and controls

130 lines (83 loc) · 3.08 KB
Raw

Snoopy Hacking - QA

TLDR version:

  • Most of the code quality checks will run automatically (via GitHub Actions) on every PR and on every push to relevant branches. Find the automated runs' outputs on GitHub (in PR, in a commit or directly in the Actions tab.
  • If you want to run these checks locally, find the steps below.

Test suite

Runs automaticaly on GitHub, for:

  • A matrix of common OSes
  • A matrix of argument combinations to ./configure

This is a suite of test cases for verifying sanity of Snoopy's basic functionality.

Test suite - running manually

Start the in-repository test suite like this:

./configure --enable-everything &&
make &&
make tests

If there are errors, each test case has a corresponding .log file right next to it, where additional information about the error is (usually) available.

Autoscan

Runs automatically on GitHub.

The autoscan tool scans the local sources and generates configure.scan file containing all detected dependelcies.

Autoscan - running manually

AS1: Clean out the repository, or else autoscan picks up unrelated dependencies from build/aux/ltmain.sh, namely AC_PROG_CPP, AC_PROG_CXX and AC_PROG_RANLIB. The bootstrap+configure steps are present here to bring the repository into a known-good state, or else make gitclean might simply not work:

./bootstrap.sh &&
./configure --enable-everything &&
make gitclean

AS2: Run autoscan now:

autoscan

AS3: The configure.scan file is already committed to the repository from the previous autoscan run. This should make spotting newly-detected entries simple:

git diff

AS4: If git diff points out any changes, you will most likely want to add them to the configure.ac file.

Autoreconf

Runs automatically on GitHub.

The autoreconf tool (re)generates the config.h.in file from the latest configure.ac content (potentially updated by the Autoscan step above).

Autoreconf - running manually

AR1: Once the above Autoscan-related additions are done, update the config.h.in file with:

./bootstrap.sh

AR2: Commit.

Valgrind

Runs automatically on GitHub.

Valgrind is a local diagnostic tool that checks the binary for any memory and/or file descriptor leaks.

Valgrind - running manually

Run the valgrind check as follows:

./configure --enable-everything &&
make valgrind

Coverity

Exists as a GitHub workflow, but does not run automatically on GitHub. Maintainers can manually trigger this workflow when required (i.e. before releasing a new stable version).

Coverity is a static code analysis tool.

Coverity - running manually

To manually upload the build to Coverity, the run the following:

./dev-tools/submit-to-coverity.sh

The Coverity scan results will be available here.