robotstatuschecker 3.0.1

Introduction

StatusChecker is a tool for validating that executed Robot Framework test cases have expected statuses and log messages. It is mainly useful for Robot Framework test library developers who want to use Robot Framework to also test their libraries. StatusChecker 1.3 and newer are compatible both with Python 2 and Python 3.

StatusChecker project is hosted at GitHub and downloads are at PyPI.

Installation instructions

The easiest way to install StatusChecker is by using pip:

pip install robotstatuschecker

Alternatively you can get the code by cloning the project from GitHub or downloading the source distribution from PyPI and extracting it. After that you can install the tool with:

python setup.py install

Usage

From the command line:

python -m robotstatuschecker infile [outfile]

Programmatically:

from robotstatuschecker import process_output

process_output('infile.xml', 'outfile.xml')

If an output file is not given, the input file is edited in place.

Defining expected test status

By default, all test cases are expected to PASS and have no message. Changing the expected status to FAIL is done by having the word FAIL (in uppercase) somewhere in the test case documentation. The expected error message must then follow the FAIL marker.

For robotframework version 4 you can also change the expected status to SKIP by adding the word SKIP in the test case documentation. Like Fail, the expected skip message must follow the word SKIP. If a test documentation contains the words FAIL and SKIP, SKIP will be ignored and the expected status will be FAIL.

If a test is expected to PASS with a certain message, the word PASS must be added to its documentation explicitly and the expected message given after that.

If a message check should happen in test setup or teardown, that check must be prefixed with SETUP or TEARDOWN word.

The expected message can also be specified as a regular expression by prefixing it with REGEXP:. The specified regular expression must match the error message fully. Having spaces between the status, the message and the possible regular expression prefix is optional.

An alternative to using regular expressions is using glob patterns where * matches anything (including newline) and ? matches any single character. This is can be accomplished by starting the expected message with GLOB:.

Finally, it is possible to test that the message starts with something by prefixing the expected message with STARTS:.

The following examples illustrate different ways to define test statuses and messages:

*** Test Cases ***
Simple failure
    [Documentation]    FAIL Expected error message
    Steps

Check in test setup is done by SETUP marker
    [Documentation]    LOG SETUP    This first log message in test setup
    [Setup]    Test specific setup
    Steps

Exclude documentation before marker
    [Documentation]    This text is ignored FAIL Expected error message
    Steps

Regexp example
    [Documentation]    FAIL REGEXP: (IOError|OSError): .*
    Steps

Glob example
    [Documentation]    FAIL GLOB: ??Error: *
    Steps

Start example
    [Documentation]    FAIL STARTS: IOError:
    Steps

Passing without message
    Steps

Passing with message
    [Documentation]    PASS Expected message
    Steps

Defining expected log messages

The expected keyword log messages can also be defined in the test case documentation using a syntax such as:

LOG x.y:z LEVEL Actual message

The part before the colon specifies the keyword to check. For example, 1 means first keyword, 1.2 is the second child keyword of the first keyword, and so on.

The part after the colon species the message. For example, 1:2 means the second message of the first keyword and 1.2:3 is the third message of the second child keyword of the first keyword. The message index is optional and defaults to 1. The message index also supports wildcard *. For example 1:* matches any message of the first keyword.

Message level is specified before the actual message, and it can be any of the valid log levels in capital letters. If the level is not given it defaults to INFO. Starting from 1.4 release also ERROR level is supported. The message level also supports wildcard ANY which will match all log levels.

Possible leading and trailing whitespace is ignored both in the expected and in the actual log message.

This syntax can be used multiple times to test multiple messages. It also works together with specifying the expected error message with FAIL, but it that case FAIL and the expected error must be first.

It is also possible to give the message as a regular expression or glob pattern or to give just the start of the message. This is accomplished by prefixing the message with REGEXP:, GLOB: or STARTS:, respectively, exactly like when defining expected test status.

Finally, to check that a keyword does not have a certain message, it is possible to use NONE in the place of the message.

*** Test cases ***
Simple example
    [Documentation]    LOG 1        Hello, world!
    Steps

Nested keywords
    [Documentation]    LOG 2.1      1st child of 2nd kw
    Steps

Message index
    [Documentation]    LOG 2:2      2nd msg of 2nd kw
    Steps

Nested and index
    [Documentation]    LOG 3.1:2    2nd msg of 3rd kw's 1st child
    Steps

Log levels
    [Documentation]    LOG 2        DEBUG Debug-level message
    ...                LOG 1.2:3    WARN Warning
    Steps

Multiple messages
    [Documentation]    LOG 1        First tested message
    ...                LOG 1.2      Second tested message
    ...                LOG 2.2.1    DEBUG Third tested message
    Steps

Status and log
    [Documentation]    FAIL         Expected error message
    ...                LOG 1.2      Expected log message
    Steps

Regexp message
    [Documentation]    LOG 1        REGEXP: (Hello|Hi) world!
    Steps

Glob message
    [Documentation]    LOG 1        GLOB: * world!
    Steps

Start of the message
    [Documentation]    LOG 1        STARTS: Hello w
    Steps

No message
    [Documentation]    LOG 1:1      Test that we have only 1 msg
    ...                LOG 1:2      NONE
    Steps

Count Messages
    [Documentation]    LOG 4 COUNT: 2    # Fourth keyword should have excatly 2 messages.
    Steps