These options let you control which information is collected in a run-log, if and when a run-log is shown and how to locate errors.

		Figure 29.25:  General run-log options

When to show run-logs (User)

A run-log is created for every execution of a test. A number of recent run-logs are available from the »Run« menu, the most recent run-log can also be opened by pressing [Control-L]. Additionally the run-log can be shown during execution or after an error as follows:

At start
This choice causes QF-Test to open the run-log when it begins executing a test sequence. The nodes of the log will be added as execution proceeds.

After finish
With this choice the run-log is shown after replay is finished.

On exception
The run-log is shown only if an uncaught exception is thrown.

Don't show
The run-log is not displayed automatically. You have to open it via the »Run« menu or by pressing [Control-L].

Number of run-logs in menu (User)

A limit for the number of menu items for recent run-logs kept in the »Run« menu.

3.0+Automatically save run-logs (User)

To prevent excessive memory use through run-logs and also to make the most recent run-logs persistent between QF-Test sessions, the recent run-logs kept in the »Run« menu are automatically saved in the directory .qftest in the current user's home directory. The filename for the run-log is based on a timestamp. QF-Test uses file locks to prevent collisions and accidental removal in case of parallel sessions and automatically keeps the .qftest directory clean by removing unreferenced logs, so there should be no reason to disable this feature. Still, you can do so by disabling this option.

3.1+Show expanded variable values in tree nodes (User)

The nodes in the tree view of a run-log can either be displayed with variables expanded to the value they had at run-time or with the original variables. Both views have their use, so you can toggle between them via this option or more quickly via the menu item »View«-»Show nodes expanded«.

Find next error function locates (User)

The search for an error in a run-log, accessible through the »Edit«-»Find next error« menu item or by pressing [Control-N], is a very useful utility. This option determines whether the search stops at exceptions only, errors and exceptions or even warnings.

Skip suppressed errors (User)

Like the previous one, this option controls the search for errors in a run-log. If activated, warnings, errors or exceptions that have not propagated to the top, are not found. Thus exceptions caught by a 'Try'/'Catch' clause or messages suppressed through the 'Maximum error level' attribute are skipped.

This option is also accessible through the »Edit«-»Skip suppressed errors« menu item.

3.0+ Cleanup tree when jumping to next or previous error (User)

When repeatedly jumping to errors in a run-log the tree can easily get cluttered with many expanded nodes. If this option is activated, QF-Test will automatically clean up the tree each time you navigate to an error so that only the parents of the error node are expanded.

Note When viewing split run-logs, partial run-logs containing an error will remain in memory as long as their nodes are expanded. Keeping this option activated will ensure that partial run-logs will be released as soon as possible, keeping memory use manageable even viewing the errors of a very large run-logs.

Save compressed run-logs (*.qrz) (System)

Server script name: OPT_LOG_SAVE_COMPRESSED

Run-logs can either be saved as plain or as compressed XML files. For large run-logs without screenshots the compression factor can be as high as 10, so it's advisable to use compressed logs where possible. The only reason not to use compression is if you want to transform the XML run-log afterwards. But even then compressed run-logs are an option because the compression method used is standard gzip format, so converting to and from compressed run-logs can easily be done using gzip.

When saving a log-file interactively you can always switch between compressed or non-compressed format by choosing the appropriate filter or by giving the file the extension .qrz or .qrl.

In batch mode, the default run-log format is compressed. To create an uncompressed run-log, simply specify the extension .qrl in the parameter for the -runlog [<file>] command line argument.

3.0+Create split run-logs (User)

Server script name: OPT_LOG_SAVE_SPLIT

A run-log can be split into several parts by setting the 'Name for separate run-log' attribute of a 'Data driver' or any of the various test nodes. By turning this option off you can temporarily disable support for split run-logs in order to get a normal, single run-log without having to modify any 'Name for separate run-log' attributes.

See subsection 9.1.4 for further information about split run-logs.

3.0+ Save split run-logs as ZIP files (*.qzp) (User)

Server script name: OPT_LOG_SAVE_SPLIT_ZIP

Split run-logs can either be saved as a single ZIP file with the extension .qzp, containing the main run-log and all partial logs together, or as a normal .qrl or .qrz run-log that is accompanied by a directory with the same base name and the suffix _logs, e.g. the file runlog.qrz plus the directory runlog_logs. This option determines the format in which split run-logs are created in interactive mode. It has no effect if the option Automatically save run-logs is turned off.

See subsection 9.1.4 for further information about split run-logs.

3.4+ Minimum size for automatic splitting (kB) (System)

Server script name: OPT_LOG_AUTO_SPLIT_SIZE

Split run-logs are the only reliable way to prevent running out of memory during very long running tests or when the run-log grows quickly due to screenshots or output from the SUT. They are also more efficient when transforming run-logs into reports. However, explicit setting of 'Name for separate run-log' attributes requires an understanding of the issues involved and either making decisions about where best to split a run-log or tedious typing when trying to split into small pieces.

As a compromise, QF-Test makes a very rough calculation about the size of a run-log during executing, taking screenshots and program output into account. When execution of a 'Test-case' or 'Test-set' has finished and the approximate size of the run-log pertaining to that node is larger than the threshold specified in this option, the run-log is split off and saved automatically. A value of 0 prevents automatic splitting.

See subsection 9.1.4 for further information about split run-logs.

3.4+ Name for automatically split 'Test-case' run-log (System)

Server script name: OPT_LOG_AUTO_SPLIT_TESTCASE_NAME

This option specifies the name to use for an external log when it is split off automatically after execution of a 'Test-case' has finished as described in the previous option. Variables can be used as well as the '%...' placeholders documented for the attribute 'Name for separate run-log'.

The special variable ${qftest:testcase.splitlogname} is a good base. It expands to a path name created from the name of the 'Test-case' with possible parent 'Test-set' nodes as directories.

See subsection 9.1.4 for further information about split run-logs.

3.4+ Name for automatically split 'Test-set' run-log (System)

Server script name: OPT_LOG_AUTO_SPLIT_TESTSET_NAME

This option specifies the name to use for an external log when it is split off automatically after execution of a 'Test-set' has finished as described in the option Minimum size for automatic splitting (kB). Variables can be used as well as the '%...' placeholders documented for the attribute 'Name for separate run-log'.

The special variable ${qftest:testset.splitlogname} is a good base. It expands to a path name created from the name of the 'Test-set' with possible parent 'Test-set' nodes as directories.

See subsection 9.1.4 for further information about split run-logs.

		Figure 29.26:  Options determining run-log content

Log variable expansion (System)

Server script name: OPT_LOG_CONTENT_VARIABLE_EXPANSION

If this option is activated, every variable expansion is logged.

Log parent nodes of components (System)

Server script name: OPT_LOG_CONTENT_PARENT_COMPONENTS

Setting this option will cause all direct and indirect parent nodes to be logged in addition to the target component node for every event, check, etc.

Log level for the SUT (System)

Server script name: OPT_LOG_CONTENT_SUT_LEVEL
Possible Values: VAL_LOG_SUT_LEVEL_MESSAGE, VAL_LOG_SUT_LEVEL_WARNING, VAL_LOG_SUT_LEVEL_ERROR

The level for automatically generated messages in the SUT during replay, e.g. details for component recognition. Only messages with the respective level, i.e. plain messages, warnings or errors will be logged. This option has no effect on messages created explicitly via rc.logMessage or qf.logMessage.

Number of events to log for error diagnosis (System)

SUT script name: OPT_LOG_CONTENT_DIAGNOSIS

During replay of a test QF-Test logs various events and other things going on behind the scenes. This information is quickly discarded except when an error happens. In that case the most recent events are written to a special run-log node. The information may also be useful to developers but is mostly required for error diagnosis when requesting support from Quality First Software GmbH.

This option determines the number of recent internal events to keep. Setting it to 0 disables the feature altogether. You should not set this value to less than about 400 without a good reason. Because the information is logged only for errors, the cost for gathering it is minimal.

Maximum number of errors with screenshots per run-log (System)

Server script name: OPT_LOG_CONTENT_SCREENSHOT_MAX

The maximum number of screenshots that QF-Test takes and stores in the run-log during a test-run on situations of exception or errors. Setting this value to 0 disables taking screenshots entirely, a negative value means unlimited screenshots.

Count screenshots individually for each split log (System)

Server script name: OPT_LOG_CONTENT_SCREENSHOT_PER_SPLIT_LOG

If this option is set, each partial log of a split run-log may contain the maximum number of screenshots defined above without affecting the count for the main run-log. Otherwise the limit applies for the sum of all parts belonging to the same main run-log.

See subsection 9.1.4 for further information about split run-logs.

Create screenshots of the whole screen (System)

Server script name: OPT_LOG_CONTENT_SCREENSHOT_FULLSCREEN

Activating this option causes QF-Test to take an image of the whole screen and save it in the run-log when a screenshot is triggered by an exception or error.

Create screenshots of individual windows (System)

Server script name: OPT_LOG_CONTENT_SCREENSHOT_WINDOW

Activating this option causes QF-Test to record images of all windows and dialogs of the SUT and store them in the run-log when screenshots are triggered due to exceptions or errors. Most of the time this will work even for windows that are covered by other windows or in cases where a full screenshot is not possible, for example when a screen is locked.

3.4+ Log successful advanced image checks (System)

Server script name: OPT_LOG_CONTENT_SCREENSHOT_ADVANCED_ALWAYS

If this option is activated, QF-Test will store the expected and actual images as well as the transformed images for successful advanced image checks in the run-log. Otherwise these details are kept for failed image checks only.

Activating this option can raise the size of the run-log drastically so be sure to use it in combination with compact run-logs and split run-logs.

Create compact run-log (System)

Server script name: OPT_LOG_CONTENT_COMPACT

Activating this option causes QF-Test to discard every node from a run-log that is neither relevant for error diagnosis, nor for the XML/HTML report. After an error or exception, as well as at the end of a test-run, the 100 most recent nodes are not discarded, so all the relevant information should remain available.

Large tests should almost always be run with this option enabled to reduce the risk of running out of memory due to excessively large run-logs.

This option is used only when QF-Test is run in interactive mode. It is ignored in batch mode (see section 1.5) to avoid accidental loss of information. To create a compact run-log in batch mode, use the -compact command line argument.

Don't create run-log (System)

Server script name: OPT_LOG_CONTENT_SUPPRESS

For very-long-running tests or demos that are run in an endless loop, memory consumption of the run-log is an issue, but split run-logs are an ideal solution. Before split run-logs were available, turning run-logs off completely via this option was sometimes the only way to get long-running tests to work. Now this option is only retained for backwards compatibility.

Note In batch mode this option is ignored. To suppress the run-log, use the argument -nolog.

3.0+ Log SUT output individually (System)

SUT script name: OPT_LOG_CONTENT_SPLIT_IO

If set, any text that an SUT client prints to its stdout or stderr stream is also logged in the run-log. For each interaction with the SUT QF-Test collects text printed before the event and after the event during synchronisation. This makes it possible to associate output like an excetption stacktrace that is triggered by an event with the event itself, something that is impossible if all output is kept in a single piece.

3.0+ Compactify individually logged SUT output (System)

Server script name: OPT_LOG_CONTENT_COMPACTIFY_SPLIT_IO

Output from an SUT client tends to accumulate and can consume a lot of memory. If this option is activated, individually logged SUT output for events that are no longer of interest can be removed along with the events in compact run-logs. Please see the option Create compact run-log for further information about compact run-logs.

Last update: 04/23/2012 Copyright © 1999-2012 Quality First Software GmbH