This is a repository of test data sampled from various kinds of hardware, as well as a shared library which can be used with LD_PRELOAD to trick lshw into thinking it is scanning the sample hardware.

This makes it possible to work on lshw patches while checking for any unexpected changes or regressions in the output across a wide range of hardware.

Running the tests

To run the tests, first build the wrapper library:

make

You can then run all test cases, comparing your lshw XML output to the expected XML which is checked into this repository:

./run-tests.sh

If you have built lshw from source, pass the binary's path in LSHW:

LSHW=../lshw/src/lshw ./run-tests.sh

Status of the tests

These tests are expected to pass with the latest published release of lshw, B.02.17.

However, note that if you run them all on your x86_64 workstation some failures are expected due to compile-time differences in lshw across platforms.

Collecting new hardware samples

On the system you wish to collect a sample from, run the collect-sample.sh script. This script needs root access because many of the kernel special files it reads from are readable only by root.

sudo ./collect-sample.sh testdata/<new-sample-name>

Note that the script will emit many errors when it runs, due to attempting to read from various /sys special files that are not actually readable.

On pre-systemd distros it may be necessary to manually mount debugfs (TODO check if we can just grab /proc/bus/usb/devices instead):

sudo mount -t debugfs none /sys/kernel/debug