Checking simple commands:

When writing test scripts, you mainly want to have an overview of what test was successful and which test was a failure. In order to do that you would use the script ldstest_check like this:

ldstest_check "Testing: ls /" 5 ls /
ldstest_check "Testing: touch /" 5 touch /
ldstest_check "Testing: sleep 10" 5 sleep 10

Where arguments to the ldstest_check are:

• The description of the test
• The timeout of the test
• The command to run

In case of failure the output of the command is displayed. The previous example would show:

Testing: ls /               OK
Testing: touch /            FAILED

touch /
touch: setting times of `/': Permission denied
Testing: sleep 10           TIMEOUT

sleep 10
/home/lds/dev/ldstest/ldstest_check: line 52: 15596 Killed                  ldstest_log_to_file "$@"

Checking C code:

To test C code ldstest provides a ldstest_check_c command to transparently run the test under valgrind. The ldstest_check_c command exports a $STARTER environment variable that provides a helper script that will run the test under Valgrind. To use it:

ldstest_check_c "1 2 3" "Testing echo output" ./test_echo_output.sh

Where arguments to the ldstest_check are:

• The timeouts when running the test normally, the under valgrind, then under helgrind
• The description of the test
• The command to run

The test_echo_output.sh would contain :

#!/bin/bash

# Check the output of "echo blah" is really blah
if [ $("$STARTER" echo blah) != "blah" ]
then
    echo "Doo!"
    exit 1
fi

At runtime the $STARTER variable is replaced by the correct, issuing such tests:

Testing echo output         OK
Valgrind check: Testing echo output         OK
Helgrind check: Testing echo output         OK

Other options

Some options can be exported as global variables to modify the behavior of tests: