Robot Framework 7.0 is a new major release with enhanced listener interface (#3296), native VAR
syntax for creating variables (#3761), support for mixing embedded and normal arguments with library keywords (#4710), JSON result format (#4847) and various other enhancements and bug fixes. Robot Framework 7.0 requires Python 3.8 or newer (#4294).
Robot Framework 7.0 rc 2 was released on Thursday January 4, 2024, with all features and fixes planned to be included in the final release. It is targeted for anyone interested to see how they can use the interesting new features and how backwards incompatible changes and deprecations possibly affect their tests, tasks, tools and libraries. The target date for the final release is Wednesday January 10, 2024.
Questions and comments related to the release can be sent to the #devel
channel on Robot Framework Slack and possible bugs submitted to the issue tracker.
If you have pip installed, just run
pip install --pre --upgrade robotframework
to install the latest available release or use
pip install robotframework==7.0rc2
to install exactly this version. Alternatively you can download the package from PyPI and install it manually. For more details and other installation approaches, see the installation instructions.
If you are interested to learn more about the new features in Robot Framework 7.0, join the RoboCon conference in February, 2024. Pekka Klärck, Robot Framework lead developer, will go through the key features briefly in the onsite conference in Helsinki and more thoroughly in the online edition.
The conference has also dozens of other great talks, workshops and a lot of possibilities to meet other community members as well as developers of various tools and libraries in the ecosystem. All profits from the conference will be used for future Robot Framework development.
Robot Framework's listener interface is a very powerful mechanism to get notifications about various events during execution and it also allows modifying data and results on the fly. It is not typically directly used by normal Robot Framework users, but they are likely to use tools that are based on it. The listener API has been significantly enhanced making it possible to create even more powerful and interesting tools in the future.
The major limitation with the listener API has been that the listener API version 2 only supports getting notifications and that the more powerful listener API version 3 has only supported suites and tests/tasks.
The biggest enhancement in the whole Robot Framework 7.0 is that the listener version 3 has been extended to support also keywords and control structures (#3296). For example, a listener having the following methods prints information about started keywords and ended WHILE loops:
python
from robot.running import Keyword as KeywordData, While as WhileData from robot.result import Keyword as KeywordResult, While as WhileResult
- def start_keyword(data: KeywordData, result: KeywordResult):
print(f"Keyword '{result.full_name}' used on line {data.lineno} started.")
- def end_while(data: WhileData, result: WhileResult):
- print(f"WHILE loop on line {data.lineno} ended with status {result.status} "
f"after {len(result.body)} iterations.")
With keywords it is possible to also get more information about the actually executed keyword. For example, the following listener prints some information about the executed keyword and the library it belongs to:
python
from robot.running import Keyword as KeywordData, LibraryKeyword from robot.result import Keyword as KeywordResult
- def start_library_keyword(data: KeywordData,
implementation: LibraryKeyword, result: KeywordResult):
- library = implementation.owner
- print(f"Keyword '{implementation.name}' is implemented in library "
f"'{library.name}' at '{implementation.source}' on line " f"{implementation.lineno}. The library has {library.scope.name} " f"scope and the current instance is {library.instance}.")
As the above example already illustrated, it is even possible to get an access to the actual library instance. This means that listeners can inspect the library state and also modify it. With user keywords it is even possible to modify the keyword itself or, via the owner
resource file, any other keyword in the resource file.
Listeners can also modify results if needed. Possible use cases include hiding sensitive information and adding more details to results based on external sources.
Notice that although listener can change status of any executed keyword or control structure, that does not directly affect the status of the executed test. In general listeners cannot directly fail keywords so that execution would stop or handle failures so that execution would continue. This kind of functionality may be added in the future if there are needs.
The new listener v3 methods are so powerful and versatile that going them through thoroughly in these release notes is not possible. For more examples, you can see the acceptance tests using the methods in various interesting and even crazy ways.
Earlier listeners needed to specify the API version they used with the ROBOT_LISTENER_API_VERSION
attribute. Now that the listener version 3 got the new methods, it is considered so much more powerful than the version 2 that it was made the default listener version (#4910).
The listener version 2 continues to work, but using it requires specifying the listener version as earlier. The are no plans to deprecate the listener version 2, but we nevertheless highly recommend using the version 3 whenever possible.
Listeners are typically enabled from the command line, but libraries can register listeners as well. Often libraries themselves want to act as listeners, and that has earlier required using self.ROBOT_LIBRARY_LISTENER = self
in the __init__
method. Robot Framework 7.0 makes it possible to use string SELF
(case-insensitive) for this purpose as well (#4910), which means that a listener can be specified as a class attribute and not only in __init__
. This is especially convenient when using the @library
decorator:
python
from robot.api.deco import keyword, library
@library(listener='SELF') class Example:
- def start_suite(self, data, result):
...
@keyword def example(self, arg): ...
Listeners have methods like output_file
and log_file
that are called when result files are ready so that they get the file path as an argument. Earlier paths were strings, but nowadays listener version 3 methods get them as more convenient pathlib.Path objects.
Modifying keyword call arguments programmatically has been made more convenient (#5000). This enhancement eases modifying arguments using the new listener version 3 start/end_keyword
methods.
The new VAR
syntax (#3761) makes it possible to create local variables as well as global, suite and test/task scoped variables dynamically during execution. The motivation is to have a more convenient syntax than using the Set Variable
keyword for creating local variables and to unify the syntax for creating variables in different scopes. Except for the mandatory VAR
marker, the syntax is also the same as when creating variables in the Variables section. The syntax is best explained with examples:
robotframework
* Test Cases* Example # Create a local variable ${local}
with a value value
. VAR ${local} value
# Create a variable that is available throughout the whole suite. # Supported scopes are GLOBAL, SUITE, TEST, TASK and LOCAL (default). VAR ${suite} value scope=SUITE
# Validate created variables. Should Be Equal ${local} value Should Be Equal ${suite} value
- Example continued
# Suite level variables are seen also by subsequent tests. Should Be Equal ${suite} value
When creating ${scalar}
variables having long values, it is possible to split the value to multiple lines. Lines are joined together with a space by default, but that can be changed with the separator
configuration option. Similarly as in the Variables section, it is possible to create also @{list}
and &{dict}
variables. Unlike in the Variables section, variables can be created conditionally using IF/ELSE structures:
robotframework
* Test Cases* Long value VAR ${long} ... This value is rather long. ... It has been split to multiple lines. ... Parts will be joined together with a space.
- Multiline
VAR ${multiline} ... First line. ... Second line. ... Last line. ... separator=n
- List
# Creates a list with three items. VAR @{list} a b c
- Dictionary
# Creates a dictionary with two items. VAR &{dict} key=value second=item
- Normal IF
- IF 1 > 0
VAR ${x} true value
- ELSE
VAR ${x} false value
END
- Inline IF
IF 1 > 0 VAR ${x} true value ELSE VAR ${x} false value
User keywords got support to use both embedded and normal arguments in Robot Framework 6.1 (#4234) and now that support has been added also to library keywords (#4710). The syntax works so, that if the function or method implementing the keyword accepts more arguments than there are embedded arguments, the remaining arguments can be passed in as normal arguments. This is illustrated by the following example keyword:
python
@keyword('Number of ${animals} should be') def example(animals, count): ...
The above keyword could be used like this:
robotframework
* Test Cases* Example Number of horses should be 2 Number of horses should be count=2 Number of dogs should be 3
Robot Framework 6.1 added support to convert test/task data to JSON and back and Robot Framework 7.0 extends the JSON serialization support to execution results (#4847). One of the core use cases for data serialization was making it easy to transfer data between process and machines, and now it is also easy to pass results back.
Also the built-in Rebot tool that is used for post-processing results supports JSON files both in output and in input. Creating JSON output files is done using the normal --output
option so that the specified file has a .json
extension:
rebot --output output.json output.xml
When reading output files, JSON files are automatically recognized by the extension:
rebot output.json
rebot output1.json output2.json
When combining or merging results, it is possible to mix JSON and XML files:
rebot output1.xml output2.json
rebot --merge original.xml rerun.json
The JSON output file structure is documented in the result.json
schema file.
The plan is to enhance the support for JSON output files in the future so that they could be created already during execution. For more details see issue #3423.
Automatic argument conversion is a very powerful feature that library developers can use to avoid converting arguments manually and to get more useful Libdoc documentation. There are two important new enhancements to it.
In Python, the Literal type makes it possible to type arguments so that type checkers accept only certain values. For example, this function only accepts strings x
, y
and `z`:
python
- def example(arg: Literal['x', 'y', 'z']):
...
Robot Framework has been enhanced so that it validates that an argument having a Literal
type can only be used with the specified values (#4633). For example, using a keyword with the above implementation with a value xxx
would fail.
In addition to validation, arguments are also converted. For example, if an argument accepts Literal[-1, 0, 1]
, used arguments are converted to integers and then validated. In addition to that, string matching is case, space, underscore and hyphen insensitive. In all cases exact matches have a precedence and the argument that is passed to the keyword is guaranteed to be in the exact format used with Literal
.
Literal
conversion is in many ways similar to Enum conversion that Robot Framework has supported for long time. Enum
conversion has benefits like being able to use a custom documentation and it is typically better when using the same type multiple times. In simple cases being able to just use arg: Literal[...]
without defining a new type is very convenient, though.
Python's type hinting syntax has evolved so that generic types can be parameterized like list[int]
(new in Python 3.9) and unions written as int | float
(new in Python 3.10). Using these constructs with older Python versions causes errors, but Python type checkers support also "stringified" type hints like 'list[int]'
and 'int | float'
that work regardless the Python version.
Support for stringified generics and unions has now been added also to Robot Framework's argument conversion (#4711). For example, the following typing now also works with Python 3.8:
python
- def example(a: 'list[int]', b: 'int | float'):
...
These stringified types are also compatible with the Remote library API and other scenarios where using actual types is not possible.
Individual tests and keywords can nowadays remove tags that have been set in the Settings section with Test Tags
or Keyword Tags
settings by using the -tag
syntax with their own [Tags]
setting (#4374). For example, tests T1
and T3
below get tags all
and most
, and test T2
gets tags all
and `one`:
robotframework
* Settings* Test Tags all most
* Test Cases* T1 No Operation T2 [Tags] one -most No Operation T3 No Operation
With tests it is possible to get the same effect by using the Default Tags
setting and overriding it where needed. That syntax is, however, considered deprecated (#4365) and using the new -tag
syntax is recommended. With keywords there was no similar functionality earlier.
Dynamic and hybrid libraries nowadays support asynchronous execution. In practice the special methods like get_keyword_names
and run_keyword
can be implemented as async methods (#4803).
Async support was added to the normal static library API in Robot Framework 6.1 (#4089). A bug related to handling asynchronous keywords if execution is stopped gracefully has also been fixed (#4808).
Timestamps used in the result model and stored to the output.xml file used custom format like 20231107 19:57:01.123
earlier. Non-standard formats are seldom a good idea, and in this case parsing the custom format turned out to be slow as well.
Nowadays the result model stores timestamps as standard datetime objects and elapsed times as timedelta (#4258). This makes creating timestamps and operating with them more convenient and considerably faster. The new objects can be accessed via start_time
, end_time
and elapsed_time
attributes that were added as forward compatibility already in Robot Framework 6.1 (#4765). Old information is still available via the old starttime
, endtime
and elapsedtime
attributes so this change is fully backwards compatible.
The timestamp format in output.xml has also been changed from the custom YYYYMMDD HH:MM:SS.mmm
format to ISO 8601 compatible YYYY-MM-DDTHH:MM:SS.mmmmmm
. Using a standard format makes it easier to process output.xml files, but this change also has big positive performance effect. Now that the result model stores timestamps as datetime objects, formatting and parsing them with the available isoformat() and fromisoformat() methods is very fast compared to custom formatting and parsing.
A related change is that instead of storing start and end times of each executed item in output.xml, we nowadays store their start and elapsed times. Elapsed times are represented as floats denoting seconds. Having elapsed times directly available is a lot more convenient than calculating them based on start and end times. Storing start and elapsed times also takes less space than storing start and end times.
As the result of these changes, times are available in the result model and in output.xml in higher precision than earlier. Earlier times were stored in millisecond granularity, but nowadays they use microseconds. Logs and reports still use milliseconds, but that can be changed in the future if there are needs.
Changes to output.xml are backwards incompatible and affect all external tools that process timestamps. This is discussed more in Changes to output.xml section below along with other output.xml changes.
Report and log got a new dark mode (#3725). It is enabled automatically based on browser and operating system preferences, but there is also a toggle to switch between the modes.
Robot Framework 7.0 requires Python 3.8 or newer (#4294). The last version that supports Python 3.6 and 3.7 is Robot Framework 6.1.1.
The output.xml file has changed in different ways making Robot Framework 7.0 incompatible with external tools processing output.xml files until these tools are updated. We try to avoid this kind of breaking changes, but in this case especially the changes to timestamps were considered so important that we eventually would have needed to do them anyway.
Due to the changes being relatively big, it can take some time before external tools are updated. To allow users to take Robot Framework 7.0 into use also if they depend on an incompatible tool, it is possible to use the new --legacy-output
option both as part of execution and with the Rebot tool to generate output.xml files that are compatible with older versions.
The biggest changes in output.xml are related to timestamps (#4258). With earlier versions start and end times of executed items, as well as timestamps of the logged messages, were stored using a custom YYYYMMDD HH:MM:SS.mmm
format, but nowadays the format is ISO 8601 compatible YYYY-MM-DDTHH:MM:SS.mmmmmm
. In addition to that, instead of saving start and end times to starttime
and endtime
attributes and message times to timestamp
, start and elapsed times are now stored to start
and elapsed
attributes and message times to time
.
Examples:
xml
<!-- Old format --> <msg timestamp="20231108 15:36:34.278" level="INFO">Hello world!</msg> <status status="PASS" starttime="20231108 15:37:35.046" endtime="20231108 15:37:35.046"/>
<!-- New format --> <msg time="2023-11-08T15:36:34.278343" level="INFO">Hello world!</msg> <status status="PASS" start="2023-11-08T15:37:35.046153" elapsed="0.000161"/>
The new format is standard compliant, contains more detailed times, makes the elapsed time directly available and makes the <status>
elements over 10% shorter. These are all great benefits, but we are still sorry for all the extra work this causes for those developing tools that process output.xml files.
How keyword names are stored in output.xml has changed slightly as well (#4884). With each executed keywords we store both the name of the keyword and the name of the library or resource file containing it. Earlier the latter was stored to attribute library
also with resource files, but nowadays the attribute is generic owner
. In addition to owner
being a better name in general, it also matches the new owner
attribute keywords in the result model have.
Another change is that the original name stored with keywords using embedded arguments is nowadays in source_name
attribute when it used to be in sourcename
. This change was done to make the attribute consistent with the attribute in the result model.
Examples:
xml
<!-- Old format --> <kw name="Log" library="BuiltIn">...</kw> <kw name="Number of horses should be" sourcename="Number of ${animals} should be" library="my_resource">...</kw>
<!-- New format --> <kw name="Log" owner="BuiltIn">...</kw> <kw name="Number of horses should be" source_name="Number of ${animals} should be" owner="my_resource">...</kw>
Nowadays keywords and control structures can have a message. Messages are represented as the text of the <status>
element, and they have been present already earlier with tests and suites. Related to this, control structured cannot anymore have <doc>
. (#4883)
These changes should not cause problems for tools processing output.xml files, but storing messages with each failed keyword and control structure may increase the output.xml size.
The output.xml schema has been updated and can be found via https://github.com/robotframework/robotframework/tree/master/doc/schema/.
There have been some changes to the result model that unfortunately affect external tools using it. The main motivation for these changes has been cleaning up the model before creating a JSON representation for it (#4847).
The biggest changes are related to keyword names (#4884). Earlier Keyword
objects had a name
attribute that contained the full keyword name like BuiltIn.Log
. The actual keyword name and the name of the library or resource file that the keyword belonged to were in kwname
and libname
attributes, respectively. In addition to these, keywords using embedded arguments also had a sourcename
attribute containing the original keyword name.
Due to reasons explained in #4884, the following changes have been made in Robot Framework 7.0:
- Old
kwname
is renamed toname
. This is consistent with the execution sideKeyword
. - Old
libname
is renamed to genericowner
. - New
full_name
is introduced to replace the oldname
. sourcename
is renamed tosource_name
.kwname
,libname
andsourcename
are preserved as properties. They are considered deprecated, but accessing them does not cause a deprecation warning yet.
The backwards incompatible part of this change is changing the meaning of the name
attribute. It used to be a read-only property yielding the full name like BuiltIn.Log
, but now it is a normal attribute that contains just the actual keyword name like Log
. All other old attributes have been preserved as properties and code using them does not need to be updated immediately.
The following attributes that were deprecated already in Robot Framework 4.0 have been removed (#4846):
TestSuite.keywords
. UseTestSuite.setup
andTestSuite.teardown
instead.TestCase.keywords
. UseTestCase.body
,TestCase.setup
andTestCase.teardown
instead.Keyword.keywords
. UseKeyword.body
andKeyword.teardown
instead.Keyword.children
. UseKeyword.body
andKeyword.teardown
instead.TestCase.critical
. The whole criticality concept has been removed.
Additionally, TestSuite.keywords
and TestCase.keywords
have been removed from the execution model.
There have been some changes also to the parsing model:
The node representing the deprecated
[Return]
setting has been renamed fromReturn
toReturnSetting
. At the same time, the node representing theRETURN
statement has been renamed fromReturnStatement
toReturn
(#4939).To ease transition,
ReturnSetting
has existed as an alias forReturn
starting from Robot Framework 6.1 (#4656) andReturnStatement
is preserved as an alias now. In addition to that, theModelVisitor
base class has special handling forvisit_ReturnSetting
andvisit_ReturnStatement
visitor methods so that they work correctly withReturnSetting
andReturnStatement
with Robot Framework 6.1 and newer. Issue #4939 explains this in more detail and has a concrete example how to support also older Robot Framework versions.- The node representing the
Test Tags
setting as well as the deprecatedForce Tags
setting has been renamed fromForceTags
toTestTags
(#4385).ModelVisitor
has special handling for thevisit_ForceTags
method so that it will continue to work also after the change. - The token type used with
AS
(orWITH NAME
) in library imports has been changed toToken.AS
(#4375).Token.WITH_NAME
still exists as an alias forToken.AS
. - Statement
type
andtokens
have been moved from_fields
to_attributes
(#4912). This may affect debugging the model.
The following deprecated constructs have been removed from Libdoc spec files (#4667):
datatypes
have been removed from XML or JSON spec files. They were deprecated in favor oftypedocs
already in Robot Framework 5.0 (#4160).- Type names are not anymore written to XML specs as content of the
<type>
elements. The name is available as thename
attribute of<type>
elements since Robot Framework 6.1 (#4538). types
andtypedocs
attributes have been removed from arguments in JSON specs. Thetype
attribute introduced in RF 6.1 (#4538) needs to be used instead.
Libdoc schema files have been updated and can be found via https://github.com/robotframework/robotframework/tree/master/doc/schema/.
There are two changes related to selecting tests:
- When using
--test
and--include
together, tests matching either of them are selected (#4721). Earlier tests need to match both options to be selected. - When selecting a suite using its parent suite as a prefix like
--suite parent.suite
, the given name must match the full suite name (#4720). Earlier it was enough if the prefix matched the closest parent or parents.
- The default value of the
stdin
argument used withProcess
library keyword has been changed fromsubprocess.PIPE
toNone
(#4103). This change ought to avoid processes hanging in some cases. Those who depend on the old behavior need to usestdin=PIPE
explicitly to enable that. - When type hints are specified as strings, they must use format
type
,type[param]
,type[p1, p2]
ort1 | t2
(#4711). Using other formats will cause errors taking keywords into use. In practice problems occur if the special characters[
,]
,,
and|
occur in unexpected places. For example,arg: "Hello, world!"
will cause an error due to the comma. datetime
,date
andtimedelta
objects are sent over the Remote interface differently than earlier (#4784). They all used to be converted to strings, but nowadaysdatetime
is sent as-is,date
is converted todatetime
and sent like that, andtimedelta
is converted to afloat
by usingtimedelta.total_seconds()
.- Argument conversion support with
collections.abc.ByteString
has been removed (#4983). The reason is thatByteString
is deprecated and will be removed in Python 3.14. It has not been too often needed, but if you happen to use it, you can changearg: ByteString
toarg: bytes | bytearray
and the functionality stays exactly the same. - Paths passed to listener version 3 methods like
output_file
andlog_file
have been changed from strings topathlib.Path
objects (#4988). Most of the time both kinds of paths work interchangeably, so this change is unlikely to cause issues. If you need to handle these paths as strings, they can be converted by usingstr(path)
. robot.utils.normalize
does not anymore support bytes (#4936).- Deprecated
accept_plain_values
argument has been removed from thetimestr_to_secs
utility function (#4861).
The [Return]
setting for specifying the return value from user keywords has been "loudly" deprecated (#4876). It has been "silently" deprecated since Robot Framework 5.0 when the much more versatile RETURN
setting was introduced (#4078), but now using it will cause a deprecation warning. The plan is to preserve the [Return]
setting at least until Robot Framework 8.0.
If you have lot of data that uses [Return]
, the easiest way to update it is using the Robotidy tool that can convert [Return]
to RETURN
automatically. If you have data that is executed also with Robot Framework versions that do not support RETURN
, you can use the Return From Keyword
keyword instead. That keyword will eventually be deprecated and removed as well, though.
Using singular section headers like *** Test Case ***
or *** Setting ***
nowadays causes a deprecation warning (#4432). They were silently deprecated in Robot Framework 6.0 for reasons explained in issue #4431.
- In the parsing model,
For.variables
,ForHeader.variables
,Try.variable
andExceptHeader.variable
attributes have been deprecated in favor of the newassign
attribute (#4708). - In running and result models,
For.variables
andTryBranch.variable
have been deprecated in favor of the newassign
attribute (#4708). - In the result model, control structures like
FOR
were earlier modeled so that they looked like keywords. Nowadays they are considered totally different objects and their keyword specific attributesname
,kwnane
,libname
,doc
,args
,assign
,tags
andtimeout
have been deprecated (#4846). starttime
,endtime
andelapsed
time attributes in the result model have been silently deprecated (#4258). Accessing them does not yet cause a deprecation warning, but users are recommended to usestart_time
,end_time
andelapsed_time
attributes that are available since Robot Framework 6.1.kwname
,libname
andsourcename
attributes used by theKeyword
object in the result model have been silently deprecated (#4884). New code should usename
,owner
andsource_name
instead.
- Using embedded arguments with a variable that has a value not matching custom embedded argument patterns nowadays causes a deprecation warning (#4524). Earlier variables used as embedded arguments were always accepted without validating values.
- Using
FOR IN ZIP
loops with lists having different lengths without explicitly usingmode=SHORTEST
has been deprecated (#4685). The strict mode where lengths must match will be the default mode in the future. - Various utility functions in the
robot.utils
package that are no longer used by Robot Framework itself, including the whole Python 2/3 compatibility layer, have been deprecated (#4501). If you need some of these utils, you can copy their code to your own tool or library. This change may affect existing libraries and tools in the ecosystem. case_insensitive
andwhitespace_insensitive
arguments used by some Collections and String library keywords have been deprecated in favor ofignore_case
andignore_whitespace
. The new arguments were added for consistency reasons (#4954) and the old arguments will continue to work for the time being.- Passing time as milliseconds to the
elapsed_time_to_string
utility function has been deprecated (#4862).
Robot Framework development is sponsored by the Robot Framework Foundation and its over 60 member organizations. If your organization is using Robot Framework and benefiting from it, consider joining the foundation to support its development as well.
Robot Framework 7.0 team funded by the foundation consists of Pekka Klärck and Janne Härkönen (part time). In addition to work done by them, the community has provided some great contributions:
- Ygor Pontelo added async support to the dynamic and hybrid library APIs (#4803) and fixed a bug with handling async keywords when execution is stopped gracefully (#4808).
- Topi 'top1' Tuulensuu fixed a performance regression when using
Run Keyword
so that the name of the executed keyword contains a variable (#4659). - Pasi Saikkonen added dark mode to reports and logs (#3725).
- René added return type information to Libdoc's HTML output (#3017), fixed
DotDict
equality comparisons (#4956) and helped finalizing the dark mode support (#3725). - Robin added type hints to modules that did not yet have them under the public
robot.api
package (#4841). - Mark Moberts added case-insensitive list and dictionary comparison support to the Collections library (#4343).
- Daniel Biehl enhanced performance of traversing the parsing model using
ModelVisitor
(#4934).
Big thanks to Robot Framework Foundation, to community members listed above, and to everyone else who has tested preview releases, submitted bug reports, proposed enhancements, debugged problems, or otherwise helped with Robot Framework 7.0 development.
Robot Framework lead developer
ID | Type | Priority | Summary | Added |
---|---|---|---|---|
#3296 | enhancement | critical | Support keywords and control structures with listener version 3 | beta 1 |
#3761 | enhancement | critical | Native VAR syntax to create variables inside tests and keywords |
alpha 1 |
#4294 | enhancement | critical | Drop Python 3.6 and 3.7 support | alpha 1 |
#4710 | enhancement | critical | Support library keywords with both embedded and normal arguments | alpha 1 |
#4847 | enhancement | critical | Support JSON serialization with result model | rc 1 |
#4659 | bug | high | Performance regression when using Run Keyword and keyword name contains a variable |
alpha 1 |
#4921 | bug | high | Log levels don't work correctly with robot:flatten
|
alpha 1 |
#3725 | enhancement | high | Support dark theme with report and log | rc 1 |
#4258 | enhancement | high | Change timestamps from custom strings to datetime in result model and to ISO 8601 format in output.xml |
alpha 1 |
#4374 | enhancement | high | Support removing tags set globally by using -tag syntax with [Tags] setting |
alpha 1 |
#4633 | enhancement | high | Automatic argument conversion and validation for Literal
|
beta 1 |
#4711 | enhancement | high | Support type aliases in formats 'list[int]' and 'int | float' in argument conversion |
alpha 1 |
#4803 | enhancement | high | Async support to dynamic and hybrid library APIs | alpha 2 |
#4808 | bug | medium | Async keywords are not stopped when execution is stopped gracefully | alpha 2 |
#4859 | bug | medium | Parsing errors in reStructuredText files have no source | alpha 1 |
#4880 | bug | medium | Initially empty test fails even if pre-run modifier adds content to it | alpha 1 |
#4886 | bug | medium |
Set Variable If is slow if it has several conditions |
alpha 1 |
#4898 | bug | medium | Resolving special variables can fail with confusing message | alpha 1 |
#4915 | bug | medium |
cached_property attributes are called when importing library |
alpha 1 |
#4924 | bug | medium | WHILE on_limit missing from listener v2 attributes |
alpha 1 |
#4926 | bug | medium | WHILE and TRY content are not removed with --removekeywords all
|
alpha 1 |
#4945 | bug | medium |
TypedDict with forward references do not work in argument conversion |
alpha 2 |
#4956 | bug | medium | DotDict behaves inconsistent on equality checks. x == y != not x != y and not x != y == not x == y
|
beta 1 |
#4964 | bug | medium | Variables set using Set Suite Variable with children=True cannot be properly overwritten |
rc 1 |
#4980 | bug | medium | DateTime library uses deprecated datetime.utcnow()
|
rc 1 |
#3017 | enhancement | medium | Add return type to Libdoc specs and HTML output | alpha 2 |
#4103 | enhancement | medium | Process: Change the default stdin behavior from subprocess.PIPE to None
|
alpha 1 |
#4302 | enhancement | medium | Remove Reserved library |
alpha 1 |
#4343 | enhancement | medium | Collections: Support case-insensitive list and dictionary comparisons | alpha 2 |
#4375 | enhancement | medium | Change token type of AS (or WITH NAME ) used with library imports to Token.AS
|
alpha 1 |
#4385 | enhancement | medium | Change the parsing model object produced by Test Tags (and Force Tags ) to TestTags
|
alpha 1 |
#4432 | enhancement | medium | Loudly deprecate singular section headers | alpha 1 |
#4501 | enhancement | medium | Loudly deprecate old Python 2/3 compatibility layer and other deprecated utils | alpha 1 |
#4524 | enhancement | medium | Loudly deprecate variables used as embedded arguments not matching custom patterns | alpha 1 |
#4545 | enhancement | medium | Support creating assigned variable name based on another variable like ${${var}} = Keyword
|
alpha 1 |
#4667 | enhancement | medium | Remove deprecated constructs from Libdoc spec files | alpha 1 |
#4685 | enhancement | medium | Deprecate SHORTEST mode being default with FOR IN ZIP loops |
alpha 1 |
#4708 | enhancement | medium | Use assing , not variable , with FOR and TRY/EXCEPT model objects when referring to assigned variables |
alpha 1 |
#4720 | enhancement | medium | Require --suite parent.suite to match the full suite name |
alpha 1 |
#4721 | enhancement | medium | Change behavior of --test and --include so that they are cumulative |
alpha 1 |
#4747 | enhancement | medium | Support [Setup] with user keywords |
alpha 1 |
#4784 | enhancement | medium | Remote: Enhance datetime , date and timedelta conversion |
alpha 1 |
#4841 | enhancement | medium | Add typing to all modules under robot.api
|
alpha 2 |
#4846 | enhancement | medium | Result model: Loudly deprecate not needed attributes and remove already deprecated ones | alpha 1 |
#4872 | enhancement | medium | Control continue-on-failure mode by using recursive and non-recursive tags together | rc 1 |
#4876 | enhancement | medium | Loudly deprecate [Return] setting |
alpha 1 |
#4877 | enhancement | medium | XML: Support ignoring element order with Elements Should Be Equal
|
beta 1 |
#4883 | enhancement | medium | Result model: Add message to keywords and control structures and remove doc from controls |
alpha 1 |
#4884 | enhancement | medium | Result model: Enhance storing keyword name | alpha 1 |
#4896 | enhancement | medium | Support separator=<value> configuration option with scalar variables in Variables section |
alpha 1 |
#4903 | enhancement | medium | Support argument conversion and named arguments with dynamic variable files | alpha 1 |
#4905 | enhancement | medium | Support creating variable name based on another variable like ${${VAR}} in Variables section |
alpha 1 |
#4910 | enhancement | medium | Make listener v3 the default listener API | beta 1 |
#4912 | enhancement | medium | Parsing model: Move type and tokens from _fields to _attributes
|
alpha 1 |
#4930 | enhancement | medium | BuiltIn: New Reset Log Level keyword for resetting the log level to the original value |
rc 1 |
#4939 | enhancement | medium | Parsing model: Rename Return to ReturnSetting and ReturnStatement to Return
|
alpha 2 |
#4942 | enhancement | medium | Add public argument conversion API for libraries and other tools | alpha 2 |
#4952 | enhancement | medium | Collections: Make ignore_order and ignore_keys recursive |
alpha 2 |
#4960 | enhancement | medium | Support integer conversion with strings representing whole number floats like '1.0' and '2e10'
|
beta 1 |
#4976 | enhancement | medium | Support string SELF (case-insenstive) when library registers itself as listener |
beta 1 |
#4979 | enhancement | medium | Add robot.result.TestSuite.to/from_xml methods |
rc 1 |
#4982 | enhancement | medium | DateTime: Support datetime.date as an input format with date related keywords |
rc 1 |
#4983 | enhancement | medium | Type conversion: Remove support for deprecated ByteString
|
rc 1 |
#5000 | enhancement | medium | Nicer API for setting keyword call arguments programmatically | rc 2 |
#4934 | --- | medium | Enhance performance of visiting parsing model | alpha 1 |
#4621 | bug | low | OperatingSystem library docs have broken link / title | rc 1 |
#4798 | bug | low |
--removekeywords passed doesn't remove test setup and teardown |
beta 1 |
#4867 | bug | low | Original order of dictionaries is not preserved when they are pretty printed in log messages | alpha 1 |
#4870 | bug | low | User keyword teardown missing from running model JSON schema | alpha 1 |
#4904 | bug | low | Importing static variable file with arguments does not fail | alpha 1 |
#4913 | bug | low | Trace log level logs arguments twice for embedded arguments | alpha 1 |
#4927 | bug | low | WARN level missing from the log level selector in log.html | alpha 1 |
#4967 | bug | low | Variables are not resolved in keyword name in WUKS error message | beta 1 |
#4861 | enhancement | low | Remove deprecated accept_plain_values from timestr_to_secs utility function |
alpha 1 |
#4862 | enhancement | low | Deprecate elapsed_time_to_string accepting time as milliseconds |
alpha 1 |
#4864 | enhancement | low | Process: Make warning about processes hanging if output buffers get full more visible | alpha 1 |
#4885 | enhancement | low | Add full_name to replace longname to suite and test objects |
alpha 1 |
#4900 | enhancement | low | Make keywords and control structures in log look more like original data | alpha 1 |
#4922 | enhancement | low | Change the log level of Set Log Level message from INFO to DEBUG |
alpha 1 |
#4933 | enhancement | low | Type conversion: Ignore hyphens when matching enum members | alpha 1 |
#4935 | enhancement | low | Use casefold , not lower , when comparing strings case-insensitively |
alpha 1 |
#4936 | enhancement | low | Remove bytes support from robot.utils.normalize function |
alpha 1 |
#4954 | enhancement | low | Collections and String: Add ignore_case as alias for case_insensitive
|
alpha 2 |
#4958 | enhancement | low | Document robot_running and dry_run_active properties of the BuiltIn library in the User Guide |
beta 1 |
#4975 | enhancement | low | Support times and x suffixes with WHILE limit to make it more compatible with Wait Until Keyword Succeeds
|
beta 1 |
#4988 | enhancement | low | Change paths passed to listener v3 methods to pathlib.Path instances |
rc 1 |
Altogether 86 issues. View on the issue tracker.