# Writing Test Suites [](){: #intro } ## Support for Test Suite Authors The `m:ct` module provides the main interface for writing test cases. This includes for example, the following: - Functions for printing and logging - Functions for reading configuration data - Function for terminating a test case with error reason - Function for adding comments to the HTML overview page For details about these functions, see module `m:ct`. The `Common Test` application also includes other modules named `ct_`, which provide various support, mainly simplified use of communication protocols such as RPC, SNMP, FTP, Telnet, and others. ## Test Suites A test suite is an ordinary Erlang module that contains test cases. It is recommended that the module has a name of the form `*_SUITE.erl`. Otherwise, the directory and auto compilation function in `Common Test` cannot locate it (at least not by default). It is also recommended that the `ct.hrl` header file is included in all test suite modules. Each test suite module must export function [`all/0`](`c:ct_suite:all/0`), which returns the list of all test case groups and test cases to be executed in that module. The callback functions to be implemented by the test suite are all listed in module [ct_suite ](`m:ct_suite`). They are also described in more detail later in this User's Guide. ## Init and End per Suite Each test suite module can contain the optional configuration functions [`init_per_suite/1`](`c:ct_suite:init_per_suite/1`) and [`end_per_suite/1`](`c:ct_suite:end_per_suite/1`). If the init function is defined, so must the end function be. If `init_per_suite` exists, it is called initially before the test cases are executed. It typically contains initializations common for all test cases in the suite, which are only to be performed once. `init_per_suite` is recommended for setting up and verifying state and environment on the System Under Test (SUT) or the `Common Test` host node, or both, so that the test cases in the suite executes correctly. The following are examples of initial configuration operations: - Opening a connection to the SUT - Initializing a database - Running an installation script `end_per_suite` is called as the final stage of the test suite execution (after the last test case has finished). The function is meant to be used for cleaning up after `init_per_suite`. `init_per_suite` and `end_per_suite` execute on dedicated Erlang processes, just like the test cases do. The result of these functions is however not included in the test run statistics of successful, failed, and skipped cases. The argument to `init_per_suite` is `Config`, that is, the same key-value list of runtime configuration data that each test case takes as input argument. `init_per_suite` can modify this parameter with information that the test cases need. The possibly modified `Config` list is the return value of the function. If `init_per_suite` fails, all test cases in the test suite are skipped automatically (so called _auto skipped_), including `end_per_suite`. Notice that if `init_per_suite` and `end_per_suite` do not exist in the suite, `Common Test` calls dummy functions (with the same names) instead, so that output generated by hook functions can be saved to the log files for these dummies. For details, see [Common Test Hooks](ct_hooks_chapter.md#manipulating). [](){: #per_testcase } ## Init and End per Test Case Each test suite module can contain the optional configuration functions [`init_per_testcase/2`](`c:ct_suite:init_per_testcase/2`) and [`end_per_testcase/2`](`c:ct_suite:end_per_testcase/2`). If the init function is defined, so must the end function be. If `init_per_testcase` exists, it is called before each test case in the suite. It typically contains initialization that must be done for each test case (analog to `init_per_suite` for the suite). `end_per_testcase/2` is called after each test case has finished, enabling cleanup after `init_per_testcase`. > #### Note {: .info } > > If `end_per_testcase` crashes, however, test results are unaffected. At the > same time, this occurrence is reported in the test execution logs. The first argument to these functions is the name of the test case. This value can be used with pattern matching in function clauses or conditional expressions to choose different initialization and cleanup routines for different test cases, or perform the same routine for many, or all, test cases. The second argument is the `Config` key-value list of runtime configuration data, which has the same value as the list returned by `init_per_suite`. `init_per_testcase/2` can modify this parameter or return it "as is". The return value of `init_per_testcase/2` is passed as parameter `Config` to the test case itself. The return value of `end_per_testcase/2` is ignored by the test server, with exception of the [`save_config`](dependencies_chapter.md#save_config) and `fail` tuple. `end_per_testcase` can check if the test case was successful. (which in turn can determine how cleanup is to be performed). This is done by reading the value tagged with `tc_status` from `Config`. The value is one of the following: - `ok` - `{failed,Reason}` where `Reason` is `timetrap_timeout`, information from [`exit/1`](`exit/1`), or details of a runtime error - `{skipped,Reason}` where `Reason` is a user-specific term Function `end_per_testcase/2` is even called if a test case terminates because of a call to `ct:abort_current_testcase/1`, or after a timetrap time-out. However, `end_per_testcase` then executes on a different process than the test case function. In this situation, `end_per_testcase` cannot change the reason for test case termination by returning `{fail,Reason}` or save data with `{save_config,Data}`. The test case is skipped in the following two cases: - If `init_per_testcase` crashes (called _auto skipped_). - If `init_per_testcase` returns a tuple `{skip,Reason}` (called _user skipped_). The test case can also be marked as failed without executing it by returning a tuple `{fail,Reason}` from `init_per_testcase`. > #### Note {: .info } > > If `init_per_testcase` crashes, or returns `{skip,Reason}` or `{fail,Reason}`, > function `end_per_testcase` is not called. If it is determined during execution of `end_per_testcase` that the status of a successful test case is to be changed to failed, `end_per_testcase` can return the tuple `{fail,Reason}` (where `Reason` describes why the test case fails). As `init_per_testcase` and `end_per_testcase` execute on the same Erlang process as the test case, printouts from these configuration functions are included in the test case log file. [](){: #test_cases } ## Test Cases The smallest unit that the test server is concerned with is a test case. Each test case can test many things, for example, make several calls to the same interface function with different parameters. The author can choose to put many or few tests into each test case. Some things to keep in mind follows: - Many small test cases tend to result in extra, and possibly duplicated code, as well as slow test execution because of large overhead for initializations and cleanups. Avoid duplicated code, for example, by using common help functions. Otherwise, the resulting suite becomes difficult to read and understand, and expensive to maintain. - Larger test cases make it harder to tell what went wrong if it fails. Also, large portions of test code risk being skipped when errors occur. - Readability and maintainability suffer when test cases become too large and extensive. It is not certain that the resulting log files reflect very well the number of tests performed. The test case function takes one argument, `Config`, which contains configuration information such as `data_dir` and `priv_dir`. (For details about these, see section [Data and Private Directories](write_test_chapter.md#data_priv_dir). The value of `Config` at the time of the call, is the same as the return value from `init_per_testcase`, mentioned earlier. > #### Note {: .info } > > The test case function argument `Config` is not to be confused with the > information that can be retrieved from the configuration files (using > [`ct:get_config/1/2`](`ct:get_config/1`)). The test case argument `Config` is > to be used for runtime configuration of the test suite and the test cases, > while configuration files are to contain data related to the SUT. These two > types of configuration data are handled differently. As parameter `Config` is a list of key-value tuples, that is, a data type called a property list, it can be handled by the `m:proplists` module. A value can, for example, be searched for and returned with function `proplists:get_value/2`. Also, or alternatively, the general `m:lists` module contains useful functions. Normally, the only operations performed on `Config` are insertion (adding a tuple to the head of the list) and lookup. To look up a value in the config, `proplists:get_value` can be used. For example: `PrivDir = proplists:get_value(priv_dir, Config)`. The test case result can be customized in several ways. See the manual for [`Module:Testcase/1`](`c:ct_suite:'Testcase'/1`) in the `m:ct_suite` module for details. [](){: #info_function } ## Test Case Information Function For each test case function there can be an extra function with the same name but without arguments. This is the test case information function. It is expected to return a list of tagged tuples that specifies various properties regarding the test case. The following tags have special meaning: - **`timetrap`** - Sets the maximum time the test case is allowed to execute. If this time is exceeded, the test case fails with reason `timetrap_timeout`. Notice that `init_per_testcase` and `end_per_testcase` are included in the timetrap time. For details, see section [Timetrap Time-Outs](write_test_chapter.md#timetraps). - **`userdata`** - Specifies any data related to the test case. This data can be retrieved at any time using the `ct:userdata/3` utility function. - **`silent_connections`** - For details, see section [Silent Connections](run_test_chapter.md#silent_connections). - **`require`** - Specifies configuration variables required by the test case. If the required configuration variables are not found in any of the test system configuration files, the test case is skipped. A required variable can also be given a default value to be used if the variable is not found in any configuration file. To specify a default value, add a tuple of the form `{default_config,ConfigVariableName,Value}` to the test case information list (the position in the list is irrelevant). _Examples:_ ```erlang testcase1() -> [{require, ftp}, {default_config, ftp, [{ftp, "my_ftp_host"}, {username, "aladdin"}, {password, "sesame"}]}}]. ``` ```erlang testcase2() -> [{require, unix_telnet, unix}, {require, {unix, [telnet, username, password]}}, {default_config, unix, [{telnet, "my_telnet_host"}, {username, "aladdin"}, {password, "sesame"}]}}]. ``` For more information about `require`, see section [Requiring and Reading Configuration Data](config_file_chapter.md#require_config_data) in section External Configuration Data and function [`ct:require/1/2`](`ct:require/1`). > #### Note {: .info } > > Specifying a default value for a required variable can result in a test case > always getting executed. This might not be a desired behavior. If `timetrap` or `require`, or both, is not set specifically for a particular test case, default values specified by function [`suite/0`](`c:ct_suite:suite/0`) are used. Tags other than the earlier mentioned are ignored by the test server. An example of a test case information function follows: ```erlang reboot_node() -> [ {timetrap,{seconds,60}}, {require,interfaces}, {userdata, [{description,"System Upgrade: RpuAddition Normal RebootNode"}, {fts,"http://someserver.ericsson.se/test_doc4711.pdf"}]} ]. ``` [](){: #suite } ## Test Suite Information Function Function [`suite/0`](`c:ct_suite:suite/0`) can, for example, be used in a test suite module to set a default `timetrap` value and to `require` external configuration data. If a test case, or a group information function also specifies any of the information tags, it overrides the default values set by `suite/0`. For details, see [Test Case Information Function](write_test_chapter.md#info_function) and [Test Case Groups](write_test_chapter.md#test_case_groups). The following options can also be specified with the suite information list: - `stylesheet`, see [HTML Style Sheets](run_test_chapter.md#html_stylesheet) - `userdata`, see [Test Case Information Function](write_test_chapter.md#info_function) - `silent_connections`, see [Silent Connections](run_test_chapter.md#silent_connections) An example of the suite information function follows: ```erlang suite() -> [ {timetrap,{minutes,10}}, {require,global_names}, {userdata,[{info,"This suite tests database transactions."}]}, {silent_connections,[telnet]}, {stylesheet,"db_testing.css"} ]. ``` [](){: #test_case_groups } ## Test Case Groups A test case group is a set of test cases sharing configuration functions and execution properties. Test case groups are defined by function [`groups/0`](`c:ct_suite:groups/0`) that should return a term having the following syntax: ```text groups() -> GroupDefs Types: GroupDefs = [GroupDef] GroupDef = {GroupName,Properties,GroupsAndTestCases} GroupName = atom() GroupsAndTestCases = [GroupDef | {group,GroupName} | TestCase | {testcase,TestCase,TCRepeatProps}] TestCase = atom() TCRepeatProps = [{repeat,N} | {repeat_until_ok,N} | {repeat_until_fail,N}] ``` `GroupName` is the name of the group and must be unique within the test suite module. Groups can be nested, by including a group definition within the `GroupsAndTestCases` list of another group. `Properties` is the list of execution properties for the group. The possible values are as follows: ```erlang Properties = [parallel | sequence | Shuffle | {GroupRepeatType,N}] Shuffle = shuffle | {shuffle,Seed} Seed = {integer(),integer(),integer()} GroupRepeatType = repeat | repeat_until_all_ok | repeat_until_all_fail | repeat_until_any_ok | repeat_until_any_fail N = integer() | forever ``` _Explanations:_ - **`parallel`** - `Common Test` executes all test cases in the group in parallel. - **`sequence`** - The cases are executed in a sequence as described in section [Sequences](dependencies_chapter.md#sequences) in section Dependencies Between Test Cases and Suites. - **`shuffle`** - The cases in the group are executed in random order. - **`repeat, repeat_until_*`** - Orders `Common Test` to repeat execution of all the cases in the group a given number of times, or until any, or all, cases fail or succeed. _Example:_ ```erlang groups() -> [{group1, [parallel], [test1a,test1b]}, {group2, [shuffle,sequence], [test2a,test2b,test2c]}]. ``` To specify in which order groups are to be executed (also with respect to test cases that are not part of any group), add tuples on the form `{group,GroupName}` to the `all/0` list. _Example:_ ```erlang all() -> [testcase1, {group,group1}, {testcase,testcase2,[{repeat,10}]}, {group,group2}]. ``` Execution properties with a group tuple in `all/0`: `{group,GroupName,Properties}` can also be specified. These properties override those specified in the group definition (see `groups/0` earlier). This way, the same set of tests can be run, but with different properties, without having to make copies of the group definition in question. If a group contains subgroups, the execution properties for these can also be specified in the group tuple: `{group,GroupName,Properties,SubGroups}` Where, `SubGroups` is a list of tuples, `{GroupName,Properties}` or `{GroupName,Properties,SubGroups}` representing the subgroups. Any subgroups defined in `groups/0` for a group, that are not specified in the `SubGroups` list, executes with their predefined properties. _Example:_ ```erlang groups() -> [{tests1, [], [{tests2, [], [t2a,t2b]}, {tests3, [], [t31,t3b]}]}]. ``` To execute group `tests1` twice with different properties for `tests2` each time: ```erlang all() -> [{group, tests1, default, [{tests2, [parallel]}]}, {group, tests1, default, [{tests2, [shuffle,{repeat,10}]}]}]. ``` This is equivalent to the following specification: ```erlang all() -> [{group, tests1, default, [{tests2, [parallel]}, {tests3, default}]}, {group, tests1, default, [{tests2, [shuffle,{repeat,10}]}, {tests3, default}]}]. ``` Value `default` states that the predefined properties are to be used. The following example shows how to override properties in a scenario with deeply nested groups: ```erlang groups() -> [{tests1, [], [{group, tests2}]}, {tests2, [], [{group, tests3}]}, {tests3, [{repeat,2}], [t3a,t3b,t3c]}]. all() -> [{group, tests1, default, [{tests2, default, [{tests3, [parallel,{repeat,100}]}]}]}]. ``` For ease of readability, all syntax definitions can be replaced by a function call whose return value should match the expected syntax case. _Example:_ ```erlang all() -> [{group, tests1, default, test_cases()}, {group, tests1, default, [shuffle_test(), {tests3, default}]}]. test_cases() -> [{tests2, [parallel]}, {tests3, default}]. shuffle_test() -> {tests2, [shuffle,{repeat,10}]}. ``` The described syntax can also be used in test specifications to change group properties at the time of execution, without having to edit the test suite. For more information, see section [Test Specifications](run_test_chapter.md#test_specifications) in section Running Tests and Analyzing Results. As illustrated, properties can be combined. If, for example, `shuffle`, `repeat_until_any_fail`, and `sequence` are all specified, the test cases in the group are executed repeatedly, and in random order, until a test case fails. Then execution is immediately stopped and the remaining cases are skipped. Before execution of a group begins, the configuration function [`init_per_group(GroupName, Config)`](`c:ct_suite:init_per_group/2`) is called. The list of tuples returned from this function is passed to the test cases in the usual manner by argument `Config`. `init_per_group/2` is meant to be used for initializations common for the test cases in the group. After execution of the group is finished, function [`end_per_group(GroupName, Config)`](`c:ct_suite:end_per_group/2`) is called. This function is meant to be used for cleaning up after `init_per_group/2`. If the init function is defined, so must the end function be. Whenever a group is executed, if `init_per_group` and `end_per_group` do not exist in the suite, `Common Test` calls dummy functions (with the same names) instead. Output generated by hook functions are saved to the log files for these dummies. For more information, see section [Manipulating Tests](ct_hooks_chapter.md#manipulating) in section Common Test Hooks. > #### Note {: .info } > > `init_per_testcase/2` and `end_per_testcase/2` are always called for each > individual test case, no matter if the case belongs to a group or not. The properties for a group are always printed in the top of the HTML log for `init_per_group/2`. The total execution time for a group is included at the bottom of the log for `end_per_group/2`. Test case groups can be nested so sets of groups can be configured with the same `init_per_group/2` and `end_per_group/2` functions. Nested groups can be defined by including a group definition, or a group name reference, in the test case list of another group. _Example:_ ```erlang groups() -> [{group1, [shuffle], [test1a, {group2, [], [test2a,test2b]}, test1b]}, {group3, [], [{group,group4}, {group,group5}]}, {group4, [parallel], [test4a,test4b]}, {group5, [sequence], [test5a,test5b,test5c]}]. ``` In the previous example, if `all/0` returns group name references in the order `[{group,group1},{group,group3}]`, the order of the configuration functions and test cases becomes the following (notice that `init_per_testcase/2` and `end_per_testcase/2:` are also always called, but not included in this example for simplification): ```text init_per_group(group1, Config) -> Config1 (*) test1a(Config1) init_per_group(group2, Config1) -> Config2 test2a(Config2), test2b(Config2) end_per_group(group2, Config2) test1b(Config1) end_per_group(group1, Config1) init_per_group(group3, Config) -> Config3 init_per_group(group4, Config3) -> Config4 test4a(Config4), test4b(Config4) (**) end_per_group(group4, Config4) init_per_group(group5, Config3) -> Config5 test5a(Config5), test5b(Config5), test5c(Config5) end_per_group(group5, Config5) end_per_group(group3, Config3) ``` (\*) The order of test case `test1a`, `test1b`, and `group2` is undefined, as `group1` has a shuffle property. (\*\*) These cases are not executed in order, but in parallel. Properties are not inherited from top-level groups to nested subgroups. For instance, in the previous example, the test cases in `group2` are not executed in random order (which is the property of `group1`). ## Parallel Property and Nested Groups If a group has a parallel property, its test cases are spawned simultaneously and get executed in parallel. However, a test case is not allowed to execute in parallel with `end_per_group/2`, which means that the time to execute a parallel group is equal to the execution time of the slowest test case in the group. A negative side effect of running test cases in parallel is that the HTML summary pages are not updated with links to the individual test case logs until function `end_per_group/2` for the group has finished. A group nested under a parallel group starts executing in parallel with previous (parallel) test cases (no matter what properties the nested group has). However, as test cases are never executed in parallel with `init_per_group/2` or `end_per_group/2` of the same group, it is only after a nested group has finished that remaining parallel cases in the previous group become spawned. ## Parallel Test Cases and I/O A parallel test case has a private I/O server as its group leader. (For a description of the group leader concept, see [ERTS](`e:erts:index.html`)). The central I/O server process, which handles the output from regular test cases and configuration functions, does not respond to I/O messages during execution of parallel groups. This is important to understand to avoid certain traps, like the following: If a process, `P`, is spawned during execution of, for example, `init_per_suite/1`, it inherits the group leader of the `init_per_suite` process. This group leader is the central I/O server process mentioned earlier. If, at a later time, _during parallel test case execution_, some event triggers process `P` to call [`io:format/1/2`](`io:format/1`), that call never returns (as the group leader is in a non-responsive state) and causes `P` to hang. ## Repeated Groups [](){: #repeated_groups } A test case group can be repeated a certain number of times (specified by an integer) or indefinitely (specified by `forever`). The repetition can also be stopped too early if any or all cases fail or succeed, that is, if any of the properties `repeat_until_any_fail`, `repeat_until_any_ok`, `repeat_until_all_fail`, or `repeat_until_all_ok` is used. If the basic `repeat` property is used, status of test cases is irrelevant for the repeat operation. The status of a subgroup can be returned (`ok` or `failed`), to affect the execution of the group on the level above. This is accomplished by, in `end_per_group/2`, looking up the value of `tc_group_properties` in the `Config` list and checking the result of the test cases in the group. If status `failed` is to be returned from the group as a result, `end_per_group/2` is to return the value `{return_group_result,failed}`. The status of a subgroup is taken into account by `Common Test` when evaluating if execution of a group is to be repeated or not (unless the basic `repeat` property is used). The value of `tc_group_properties` is a list of status tuples, each with the key `ok`, `skipped`, and `failed`. The value of a status tuple is a list with names of test cases that have been executed with the corresponding status as result. The following is an example of how to return the status from a group: ```erlang end_per_group(_Group, Config) -> Status = proplists:get_value(tc_group_result, Config), case proplists:get_value(failed, Status) of [] -> % no failed cases {return_group_result,ok}; _Failed -> % one or more failed {return_group_result,failed} end. ``` It is also possible, in `end_per_group/2`, to check the status of a subgroup (maybe to determine what status the current group is to return). This is as simple as illustrated in the previous example, only the group name is stored in a tuple `{group_result,GroupName}`, which can be searched for in the status lists. _Example:_ ```erlang end_per_group(group1, Config) -> Status = proplists:get_value(tc_group_result, Config), Failed = proplists:get_value(failed, Status), case lists:member({group_result,group2}, Failed) of true -> {return_group_result,failed}; false -> {return_group_result,ok} end; ... ``` > #### Note {: .info } > > When a test case group is repeated, the configuration functions > `init_per_group/2` and `end_per_group/2` are also always called with each > repetition. ## Shuffled Test Case Order The order in which test cases in a group are executed is under normal circumstances the same as the order specified in the test case list in the group definition. With property `shuffle` set, however, `Common Test` instead executes the test cases in random order. You can provide a seed value (a tuple of three integers) with the shuffle property `{shuffle,Seed}`. This way, the same shuffling order can be created every time the group is executed. If no seed value is specified, `Common Test` creates a "random" seed for the shuffling operation (using the return value of `erlang:timestamp/0`). The seed value is always printed to the `init_per_group/2` log file so that it can be used to recreate the same execution order in a subsequent test run. > #### Note {: .info } > > If a shuffled test case group is repeated, the seed is not reset between > turns. If a subgroup is specified in a group with a `shuffle` property, the execution order of this subgroup in relation to the test cases (and other subgroups) in the group, is random. The order of the test cases in the subgroup is however not random (unless the subgroup has a `shuffle` property). [](){: #group_info } ## Group Information Function The test case group information function, `group(GroupName)`, serves the same purpose as the suite- and test case information functions previously described. However, the scope for the group information function, is all test cases and subgroups in the group in question (`GroupName`). _Example:_ ```erlang group(connection_tests) -> [{require,login_data}, {timetrap,1000}]. ``` The group information properties override those set with the suite information function, and can in turn be overridden by test case information properties. For a list of valid information properties and more general information, see the [Test Case Information Function](write_test_chapter.md#info_function). ## Information Functions for Init- and End-Configuration Information functions can also be used for functions `init_per_suite`, `end_per_suite`, `init_per_group`, and `end_per_group`, and they work the same way as with the [Test Case Information Function](write_test_chapter.md#info_function). This is useful, for example, for setting timetraps and requiring external configuration data relevant only for the configuration function in question (without affecting properties set for groups and test cases in the suite). The information function `init/end_per_suite()` is called for `init/end_per_suite(Config)`, and information function `init/end_per_group(GroupName)` is called for `init/end_per_group(GroupName,Config)`. However, information functions cannot be used with `init/end_per_testcase(TestCase, Config)`, as these configuration functions execute on the test case process and use the same properties as the test case (that is, the properties set by the test case information function, `TestCase()`). For a list of valid information properties and more general information, see the [Test Case Information Function](write_test_chapter.md#info_function). [](){: #data_priv_dir } ## Data and Private Directories In the data directory, `data_dir`, the test module has its own files needed for the testing. The name of `data_dir` is the name of the test suite followed by `"_data"`. For example, `"some_path/foo_SUITE.beam"` has the data directory `"some_path/foo_SUITE_data/"`. Use this directory for portability, that is, to avoid hardcoding directory names in your suite. As the data directory is stored in the same directory as your test suite, you can rely on its existence at runtime, even if the path to your test suite directory has changed between test suite implementation and execution. `priv_dir` is the private directory for the test cases. This directory can be used whenever a test case (or configuration function) needs to write something to file. The name of the private directory is generated by `Common Test`, which also creates the directory. By default, `Common Test` creates one central private directory per test run, shared by all test cases. This is not always suitable. Especially if the same test cases are executed multiple times during a test run (that is, if they belong to a test case group with property `repeat`) and there is a risk that files in the private directory get overwritten. Under these circumstances, `Common Test` can be configured to create one dedicated private directory per test case and execution instead. This is accomplished with the flag/option `create_priv_dir` (to be used with the [`ct_run`](ct_run_cmd.md) program, the `ct:run_test/1` function, or as test specification term). There are three possible values for this option as follows: - `auto_per_run` - `auto_per_tc` - `manual_per_tc` The first value indicates the default `priv_dir` behavior, that is, one private directory created per test run. The two latter values tell `Common Test` to generate a unique test directory name per test case and execution. If the auto version is used, _all_ private directories are created automatically. This can become very inefficient for test runs with many test cases or repetitions, or both. Therefore, if the manual version is used instead, the test case must tell `Common Test` to create `priv_dir` when it needs it. It does this by calling the function `ct:make_priv_dir/0`. > #### Note {: .info } > > Do not depend on the current working directory for reading and writing data > files, as this is not portable. All scratch files are to be written in the > `priv_dir` and all data files are to be located in `data_dir`. Also, the > `Common Test` server sets the current working directory to the test case log > directory at the start of every case. ## Execution Environment Each test case is executed by a dedicated Erlang process. The process is spawned when the test case starts, and terminated when the test case is finished. The configuration functions `init_per_testcase` and `end_per_testcase` execute on the same process as the test case. The configuration functions `init_per_suite` and `end_per_suite` execute, like test cases, on dedicated Erlang processes. [](){: #timetraps } ## Timetrap Time-Outs The default time limit for a test case is 30 minutes, unless a `timetrap` is specified either by the suite-, group-, or test case information function. The timetrap time-out value defined by `suite/0` is the value that is used for each test case in the suite (and for the configuration functions `init_per_suite/1`, `end_per_suite/1`, `init_per_group/2`, and `end_per_group/2`). A timetrap value defined by `group(GroupName)` overrides one defined by `suite()` and is used for each test case in group `GroupName`, and any of its subgroups. If a timetrap value is defined by `group/1` for a subgroup, it overrides that of its higher level groups. Timetrap values set by individual test cases (by the test case information function) override both group- and suite- level timetraps. A timetrap can also be set or reset dynamically during the execution of a test case, or configuration function. This is done by calling `ct:timetrap/1`. This function cancels the current timetrap and starts a new one (that stays active until time-out, or end of the current function). Timetrap values can be extended with a multiplier value specified at startup with option `multiply_timetraps`. It is also possible to let the test server decide to scale up timetrap time-out values automatically. That is, if tools such as `cover` or `trace` are running during the test. This feature is disabled by default and can be enabled with start option `scale_timetraps`. If a test case needs to suspend itself for a time that also gets multiplied by `multiply_timetraps` (and possibly also scaled up if `scale_timetraps` is enabled), the function `ct:sleep/1` can be used (instead of, for example, `timer:sleep/1`). A function (`fun/0` or `{Mod,Func,Args}` (MFA) tuple) can be specified as timetrap value in the suite-, group- and test case information function, and as argument to function `ct:timetrap/1`. _Examples:_ `{timetrap,{my_test_utils,timetrap,[?MODULE,system_start]}}` `ct:timetrap(fun() -> my_timetrap(TestCaseName, Config) end)` The user timetrap function can be used for two things as follows: - To act as a timetrap. The time-out is triggered when the function returns. - To return a timetrap time value (other than a function). Before execution of the timetrap function (which is performed on a parallel, dedicated timetrap process), `Common Test` cancels any previously set timer for the test case or configuration function. When the timetrap function returns, the time-out is triggered, _unless_ the return value is a valid timetrap time, such as an integer, or a `{SecMinOrHourTag,Time}` tuple (for details, see module `m:ct_suite`). If a time value is returned, a new timetrap is started to generate a time-out after the specified time. The user timetrap function can return a time value after a delay. The effective timetrap time is then the delay time _plus_ the returned time. [](){: #logging } ## Logging - Categories and Verbosity Levels `Common Test` provides the following three main functions for printing strings: - `ct:log(Category, Importance, Format, FormatArgs, Opts)` - `ct:print(Category, Importance, Format, FormatArgs)` - `ct:pal(Category, Importance, Format, FormatArgs)` The [`log/1,2,3,4,5`](`ct:log/1`) function prints a string to the test case log file. The [`print/1,2,3,4`](`ct:print/1`) function prints the string to screen. The [`pal/1,2,3,4`](`ct:pal/1`) function prints the same string both to file and screen. The functions are described in module `m:ct`. The optional `Category` argument can be used to categorize the log printout. Categories can be used for two things as follows: - To compare the importance of the printout to a specific verbosity level. - To format the printout according to a user-specific HTML Style Sheet (CSS). Argument `Importance` specifies a level of importance that, compared to a verbosity level (general and/or set per category), determines if the printout is to be visible. `Importance` is any integer in the range 0..99. Predefined constants exist in the `ct.hrl` header file. The default importance level, `?STD_IMPORTANCE` (used if argument `Importance` is not provided), is 50. This is also the importance used for standard I/O, for example, from printouts made with `io:format/2`, `io:put_chars/1`, and so on. `Importance` is compared to a verbosity level set by the `verbosity` start flag/option. The level can be set per category or generally, or both. If `verbosity` is not set by the user, a level of 100 (`?MAX_VERBOSITY` = all printouts visible) is used as default value. `Common Test` performs the following test: ```text Importance >= (100-VerbosityLevel) ``` The constant `?STD_VERBOSITY` has value 50 (see `ct.hrl`). At this level, all standard I/O gets printed. If a lower verbosity level is set, standard I/O printouts are ignored. Verbosity level 0 effectively turns all logging off (except from printouts made by `Common Test` itself). The general verbosity level is not associated with any particular category. This level sets the threshold for the standard I/O printouts, uncategorized `ct:log/print/pal` printouts, and printouts for categories with undefined verbosity level. _Examples:_ Some printouts during test case execution: ```erlang io:format("1. Standard IO, importance = ~w~n", [?STD_IMPORTANCE]), ct:log("2. Uncategorized, importance = ~w", [?STD_IMPORTANCE]), ct:log(info, "3. Categorized info, importance = ~w", [?STD_IMPORTANCE]), ct:log(info, ?LOW_IMPORTANCE, "4. Categorized info, importance = ~w", [?LOW_IMPORTANCE]), ct:log(error, ?HI_IMPORTANCE, "5. Categorized error, importance = ~w", [?HI_IMPORTANCE]), ct:log(error, ?MAX_IMPORTANCE, "6. Categorized error, importance = ~w", [?MAX_IMPORTANCE]), ``` If starting the test with a general verbosity level of 50 (`?STD_VERBOSITY`): ```text $ ct_run -verbosity 50 ``` the following is printed: ```text 1. Standard IO, importance = 50 2. Uncategorized, importance = 50 3. Categorized info, importance = 50 5. Categorized error, importance = 75 6. Categorized error, importance = 99 ``` If starting the test with: ```text $ ct_run -verbosity 1 and info 75 ``` the following is printed: ```erlang 3. Categorized info, importance = 50 4. Categorized info, importance = 25 6. Categorized error, importance = 99 ``` Note that the category argument is not required in order to only specify the importance of a printout. Example: ```erlang ct:pal(?LOW_IMPORTANCE, "Info report: ~p", [Info]) ``` Or perhaps in combination with constants: ```erlang -define(INFO, ?LOW_IMPORTANCE). -define(ERROR, ?HI_IMPORTANCE). ct:log(?INFO, "Info report: ~p", [Info]) ct:pal(?ERROR, "Error report: ~p", [Error]) ``` The functions `ct:set_verbosity/2` and `ct:get_verbosity/1` may be used to modify and read verbosity levels during test execution. The arguments `Format` and `FormatArgs` in `ct:log/print/pal` are always passed on to the STDLIB function `io:format/3` (For details, see the `m:io` manual page). `ct:pal/4` and `ct:log/5` add headers to strings being printed to the log file. The strings are also wrapped in div tags with a CSS class attribute, so that stylesheet formatting can be applied. To disable this feature for a printout (i.e. to get a result similar to using `io:format/2`), call `ct:log/5` with the `no_css` option. How categories can be mapped to CSS tags is documented in section [HTML Style Sheets](run_test_chapter.md#html_stylesheet) in section Running Tests and Analyzing Results. Common Test will escape special HTML characters (<, > and &) in printouts to the log file made with `ct:pal/4` and `io:format/2`. In order to print strings with HTML tags to the log, use the `ct:log/3,4,5` function. The character escaping feature is per default disabled for `ct:log/3,4,5` but can be enabled with the `esc_chars` option in the `Opts` list, see [`ct:log/3,4,5`](`ct:log/5`). If the character escaping feature needs to be disabled (typically for backwards compatibility reasons), use the `ct_run` start flag `-no_esc_chars`, or the `ct:run_test/1` start option `{esc_chars,Bool}` (this start option is also supported in test specifications). For more information about log files, see section [Log Files](run_test_chapter.md#log_files) in section Running Tests and Analyzing Results. ## Illegal Dependencies Even though it is highly efficient to write test suites with the `Common Test` framework, mistakes can be made, mainly because of illegal dependencies. Some of the more frequent mistakes from our own experience with running the Erlang/OTP test suites follows: - Depending on current directory, and writing there: This is a common error in test suites. It is assumed that the current directory is the same as the author used as current directory when the test case was developed. Many test cases even try to write scratch files to this directory. Instead `data_dir` and `priv_dir` are to be used to locate data and for writing scratch files. - Depending on execution order: During development of test suites, make no assumptions on the execution order of the test cases or suites. For example, a test case must not assume that a server it depends on is already started by a previous test case. Reasons for this follows: - The user/operator can specify the order at will, and maybe a different execution order is sometimes more relevant or efficient. - If the user specifies a whole directory of test suites for the test, the execution order of the suites depends on how the files are listed by the operating system, which varies between systems. - If a user wants to run only a subset of a test suite, there is no way one test case could successfully depend on another. - Depending on Unix: Running Unix commands through `os:cmd` are likely not to work on non-Unix platforms. - Nested test cases: Starting a test case from another not only tests the same thing twice, but also makes it harder to follow what is being tested. Also, if the called test case fails for some reason, so do the caller. This way, one error gives cause to several error reports, which is to be avoided. Functionality common for many test case functions can be implemented in common help functions. If these functions are useful for test cases across suites, put the help functions into common help modules. - Failure to crash or exit when things go wrong: Making requests without checking that the return value indicates success can be OK if the test case fails later, but it is never acceptable just to print an error message (into the log file) and return successfully. Such test cases do harm, as they create a false sense of security when overviewing the test results. - Messing up for subsequent test cases: Test cases are to restore as much of the execution environment as possible, so that subsequent test cases do not crash because of their execution order. The function [`end_per_testcase`](`c:ct_suite:end_per_testcase/2`) is suitable for this.