This document contains a specification of the API provided by the analysis server. The API in this document is currently under development. Changes to the API will be accompanied by an update to the protocol version number according to the principles of semantic versioning (semver.org).
The analysis server API is a bi-directional client-server API. The API is independent of the transport mechanism used, but is heavily influenced by a model in which sockets or character streams are used to transport JSON-RPC encoded information.
The characters passed to the server are expected to be encoded using UTF-8.
When character streams are used as the transport, messages are delineated by newlines. This means, in particular, that the JSON encoding process must not introduce newlines within a message. Note however that newlines are used in this document for readability.
It is the client's responsibility to read output from the server to avoid its blocking.
To ease interoperability with Lisp-based clients (which may not be able to easily distinguish between empty lists, empty maps, and null), client-to-server communication is allowed to replace any instance of "{}" or "[]" with null. The server will always properly represent empty lists as "[]" and empty maps as "{}".
Clients can make a request of the server and the server will provide a response for each request that it receives. While many of the requests that can be made by a client are informational in nature, we have chosen to always return a response so that clients can know whether the request was received and was correct.
There is no guarantee concerning the order in which responses will be returned, but there is a guarantee that the server will process requests in the order in which they are sent as long as the transport mechanism also makes this guarantee. Responses can be returned in an order that is different from the order in which the requests were received because some requests take longer to process than others.
Every request is required to have two fields and may have two additional optional fields. The first required field is the ‘id’ field, which is only used by the server to associate a response with the request that generated the response. The second required field is the ‘method’ field, which is used to determine what the server is being requested to do. One optional field is the ‘params’ field, whose structure is dependent on the method being requested. The structure of this field is described with each request for which it is required. The other optional field is the 'clientRequestTime' field, which is a number indicating the time at which the client made the request (milliseconds since epoch). Providing clientRequestTime helps us track how responsive analysis server is to client requests and better address any issues that occur.
Every response has up to three fields. The first field is the ‘id’ field, which is always present and whose value is the identifier that was passed to the request that generated the response. The second field is the ‘error’ field, which is only present if an error was encountered while processing the request. The third field is the ‘result’ field, whose structure is dependent on the method being responded to, and is described with each request that will produce it.
The server can also communicate to the clients by sending a notification. The purpose of these notifications is to provide information to clients as it becomes available rather than to require that clients poll for it. Unless explicitly stated, all notifications are designed to return the complete information available at the time the notification is sent; clients are not required to update previously communicated results. Consequently, the server can and should return partial results before all results are available. For example, the syntactic errors for a file can be returned as soon as the syntactic analysis is complete, and both syntactic and semantic errors can be returned together at a later time.
Each notification has two fields. The first field is the ‘event’ field, which identifies the kind of notification. The second field is the ‘params’ field, whose structure is dependent on the kind of notification being sent. The structure of this field is described with each notification.
In order to be backward compatible, clients should ignore fields that were not specified in the version of the API on which they were based. Clients should also use the server.getVersion request to test that the version of the server supports an API before using it.
The analysis server satisfies requests under the principle of eventual consistency. That is, in some cases it may return responses with the currently available results while it's catching up with unprocessed changes.
For convenience, the API is divided into domains. Each domain is specified in a separate section below. The specifications of the API’s refer to data structures beyond the standard JSON primitives. These data structures are documented in the section titled Types.
The command-line arguments that can be passed to the server.
- --client-id
Specifies an identifier associated with the client. Used when generating error reports.
Clients are strongly encouraged to provide this information in order to improve the quality of information that can be provided to them.
- --client-version
Specifies the version of the client that is communicating with the server. Used when generating error reports.
Clients are strongly encouraged to provide this information in order to improve the quality of information that can be provided to them.
- --no-error-notification
Deprecated: clients should no longer pass this option in
Disable notifications about errors (see analysis.error). If this flag is not specified then notifications will be sent for all errors produced for all files in the actual analysis roots.
- --no-index
Deprecated: clients should no longer pass this option in
This flag used to disable the server from generating an index, but now it has no effect.
- --file-read-mode
Deprecated: clients should no longer pass this option in
An enumeration of the ways files can be read from disk. Some clients normalize end of line characters which would make the file offset and range information incorrect. The default option is as-is, but can also be set to normalize-eol-always. The default option (as-is) reads files as they are on disk. The normalize-eol-always option does the following:
- '\r\n' is converted to '\n';
- '\r' by itself is converted to '\n';
- this happens regardless of the OS editor is running on.
The server domain contains API’s related to the execution of the server.
Return the version number of the analysis server.
The version number of the analysis server.
Cleanly shutdown the analysis server. Requests that are received after this request will not be processed. Requests that were received before this request, but for which a response has not yet been sent, will not be responded to. No further responses or notifications will be sent after the response to this request has been sent.
Subscribe for services. All previous subscriptions are replaced by the given set of services.
It is an error if any of the elements in the list are not valid services. If there is an error, then the current subscriptions will remain unchanged.
A list of the services being subscribed to.
Reports that the server is running. This notification is issued once after the server has started running but before any requests are processed to let the client know that it started correctly.
It is not possible to subscribe to or unsubscribe from this notification.
The version number of the analysis server.
The process id of the analysis server process.
The session id for this session.
Reports that an unexpected error has occurred while executing the server. This notification is not used for problems with specific requests (which are returned as part of the response) but is used for exceptions that occur while performing other tasks, such as analysis or preparing notifications.
It is not possible to subscribe to or unsubscribe from this notification.
True if the error is a fatal error, meaning that the server will shutdown automatically after sending this notification.
The error message indicating what kind of error was encountered.
The stack trace associated with the generation of the error, used for debugging the server.
Reports the current status of the server. Parameters are omitted if there has been no change in the status represented by that parameter.
This notification is not subscribed to by default. Clients can subscribe by including the value "STATUS" in the list of services passed in a server.setSubscriptions request.
The current status of analysis, including whether analysis is being performed and if so what is being analyzed.
The current status of pub execution, indicating whether we are currently running pub.
The analysis domain contains API’s related to the analysis of files.
Return the errors associated with the given file. If the errors for the given file have not yet been computed, or the most recently computed errors for the given file are out of date, then the response for this request will be delayed until they have been computed. If some or all of the errors for the file cannot be computed, then the subset of the errors that can be computed will be returned and the response will contain an error to indicate why the errors could not be computed. If the content of the file changes after this request was received but before a response could be sent, then an error of type CONTENT_MODIFIED will be generated.
This request is intended to be used by clients that cannot asynchronously apply updated error information. Clients that can apply error information as it becomes available should use the information provided by the 'analysis.errors' notification.
If a request is made for a file which does not exist, or which is not currently subject to analysis (e.g. because it is not associated with any analysis root specified to analysis.setAnalysisRoots), an error of type GET_ERRORS_INVALID_FILE will be generated.
The file for which errors are being requested.
The errors associated with the file.
Return the hover information associate with the given location. If some or all of the hover information is not available at the time this request is processed the information will be omitted from the response.
The file in which hover information is being requested.
The offset for which hover information is being requested.
The hover information associated with the location. The list will be empty if no information could be determined for the location. The list can contain multiple items if the file is being analyzed in multiple contexts in conflicting ways (such as a part that is included in multiple libraries).
Return the transitive closure of reachable sources for a given file.
If a request is made for a file which does not exist, or which is not currently subject to analysis (e.g. because it is not associated with any analysis root specified to analysis.setAnalysisRoots), an error of type GET_REACHABLE_SOURCES_INVALID_FILE will be generated.
The file for which reachable source information is being requested.
A mapping from source URIs to directly reachable source URIs. For example, a file "foo.dart" that imports "bar.dart" would have the corresponding mapping { "file:///foo.dart" : ["file:///bar.dart"] }. If "bar.dart" has further imports (or exports) there will be a mapping from the URI "file:///bar.dart" to them. To check if a specific URI is reachable from a given file, clients can check for its presence in the resulting key set.
Return library dependency information for use in client-side indexing and package URI resolution.
Clients that are only using the libraries field should consider using the analyzedFiles notification instead.
A list of the paths of library elements referenced by files in existing analysis roots.
A mapping from context source roots to package maps which map package names to source directories for use in client-side package URI resolution.
Return the navigation information associated with the given region of the given file. If the navigation information for the given file has not yet been computed, or the most recently computed navigation information for the given file is out of date, then the response for this request will be delayed until it has been computed. If the content of the file changes after this request was received but before a response could be sent, then an error of type CONTENT_MODIFIED will be generated.
If a navigation region overlaps (but extends either before or after) the given region of the file it will be included in the result. This means that it is theoretically possible to get the same navigation region in response to multiple requests. Clients can avoid this by always choosing a region that starts at the beginning of a line and ends at the end of a (possibly different) line in the file.
If a request is made for a file which does not exist, or which is not currently subject to analysis (e.g. because it is not associated with any analysis root specified to analysis.setAnalysisRoots), an error of type GET_NAVIGATION_INVALID_FILE will be generated.
The file in which navigation information is being requested.
The offset of the region for which navigation information is being requested.
The length of the region for which navigation information is being requested.
A list of the paths of files that are referenced by the navigation targets.
A list of the navigation targets that are referenced by the navigation regions.
A list of the navigation regions within the requested region of the file.
Force the re-analysis of everything contained in the specified analysis roots. This will cause all previously computed analysis results to be discarded and recomputed, and will cause all subscribed notifications to be re-sent.
If no analysis roots are provided, then all current analysis roots will be re-analyzed. If an empty list of analysis roots is provided, then nothing will be re-analyzed. If the list contains one or more paths that are not currently analysis roots, then an error of type INVALID_ANALYSIS_ROOT will be generated.
A list of the analysis roots that are to be re-analyzed.
Sets the root paths used to determine which files to analyze. The set of files to be analyzed are all of the files in one of the root paths that are not either explicitly or implicitly excluded. A file is explicitly excluded if it is in one of the excluded paths. A file is implicitly excluded if it is in a subdirectory of one of the root paths where the name of the subdirectory starts with a period (that is, a hidden directory).
Note that this request determines the set of requested analysis roots. The actual set of analysis roots at any given time is the intersection of this set with the set of files and directories actually present on the filesystem. When the filesystem changes, the actual set of analysis roots is automatically updated, but the set of requested analysis roots is unchanged. This means that if the client sets an analysis root before the root becomes visible to server in the filesystem, there is no error; once the server sees the root in the filesystem it will start analyzing it. Similarly, server will stop analyzing files that are removed from the file system but they will remain in the set of requested roots.
If an included path represents a file, then server will look in the directory containing the file for a pubspec.yaml file. If none is found, then the parents of the directory will be searched until such a file is found or the root of the file system is reached. If such a file is found, it will be used to resolve package: URI’s within the file.
A list of the files and directories that should be analyzed.
A list of the files and directories within the included directories that should not be analyzed.
A mapping from source directories to package roots that should override the normal package: URI resolution mechanism.
If a package root is a directory, then the analyzer will behave as though the associated source directory in the map contains a special pubspec.yaml file which resolves any package: URI to the corresponding path within that package root directory. The effect is the same as specifying the package root directory as a "--package_root" parameter to the Dart VM when executing any Dart file inside the source directory.
If a package root is a file, then the analyzer will behave as though that file is a ".packages" file in the source directory. The effect is the same as specifying the file as a "--packages" parameter to the Dart VM when executing any Dart file inside the source directory.
Files in any directories that are not overridden by this mapping have their package: URI's resolved using the normal pubspec.yaml mechanism. If this field is absent, or the empty map is specified, that indicates that the normal pubspec.yaml mechanism should always be used.
Subscribe for general services (that is, services that are not specific to individual files). All previous subscriptions are replaced by the given set of services.
It is an error if any of the elements in the list are not valid services. If there is an error, then the current subscriptions will remain unchanged.
A list of the services being subscribed to.
Set the priority files to the files in the given list. A priority file is a file that is given priority when scheduling which analysis work to do first. The list typically contains those files that are visible to the user and those for which analysis results will have the biggest impact on the user experience. The order of the files within the list is significant: the first file will be given higher priority than the second, the second higher priority than the third, and so on.
Note that this request determines the set of requested priority files. The actual set of priority files is the intersection of the requested set of priority files with the set of files currently subject to analysis. (See analysis.setSubscriptions for a description of files that are subject to analysis.)
If a requested priority file is a directory it is ignored, but remains in the set of requested priority files so that if it later becomes a file it can be included in the set of actual priority files.
The files that are to be a priority for analysis.
Subscribe for services that are specific to individual files. All previous subscriptions are replaced by the current set of subscriptions. If a given service is not included as a key in the map then no files will be subscribed to the service, exactly as if the service had been included in the map with an explicit empty list of files.
Note that this request determines the set of requested subscriptions. The actual set of subscriptions at any given time is the intersection of this set with the set of files currently subject to analysis. The files currently subject to analysis are the set of files contained within an actual analysis root but not excluded, plus all of the files transitively reachable from those files via import, export and part directives. (See analysis.setAnalysisRoots for an explanation of how the actual analysis roots are determined.) When the actual analysis roots change, the actual set of subscriptions is automatically updated, but the set of requested subscriptions is unchanged.
If a requested subscription is a directory it is ignored, but remains in the set of requested subscriptions so that if it later becomes a file it can be included in the set of actual subscriptions.
It is an error if any of the keys in the map are not valid services. If there is an error, then the existing subscriptions will remain unchanged.
A table mapping services to a list of the files being subscribed to the service.
Update the content of one or more files. Files that were previously updated but not included in this update remain unchanged. This effectively represents an overlay of the filesystem. The files whose content is overridden are therefore seen by server as being files with the given content, even if the files do not exist on the filesystem or if the file path represents the path to a directory on the filesystem.
A table mapping the files whose content has changed to a description of the content change.
Deprecated: all of the options can be set by users in an analysis options file.
Update the options controlling analysis based on the given set of options. Any options that are not included in the analysis options will not be changed. If there are options in the analysis options that are not valid, they will be silently ignored.
The options that are to be used to control analysis.
Reports the paths of the files that are being analyzed.
This notification is not subscribed to by default. Clients can subscribe by including the value "ANALYZED_FILES" in the list of services passed in an analysis.setGeneralSubscriptions request.
A list of the paths of the files that are being analyzed.
Reports the errors associated with a given file. The set of errors included in the notification is always a complete list that supersedes any previously reported errors.
The file containing the errors.
The errors contained in the file.
Reports that any analysis results that were previously associated with the given files should be considered to be invalid because those files are no longer being analyzed, either because the analysis root that contained it is no longer being analyzed or because the file no longer exists.
If a file is included in this notification and at some later time a notification with results for the file is received, clients should assume that the file is once again being analyzed and the information should be processed.
It is not possible to subscribe to or unsubscribe from this notification.
The files that are no longer being analyzed.
Reports the folding regions associated with a given file. Folding regions can be nested, but will not be overlapping. Nesting occurs when a foldable element, such as a method, is nested inside another foldable element such as a class.
This notification is not subscribed to by default. Clients can subscribe by including the value "FOLDING" in the list of services passed in an analysis.setSubscriptions request.
The file containing the folding regions.
The folding regions contained in the file.
Reports the highlight regions associated with a given file.
This notification is not subscribed to by default. Clients can subscribe by including the value "HIGHLIGHTS" in the list of services passed in an analysis.setSubscriptions request.
The file containing the highlight regions.
The highlight regions contained in the file. Each highlight region represents a particular syntactic or semantic meaning associated with some range. Note that the highlight regions that are returned can overlap other highlight regions if there is more than one meaning associated with a particular region.
Reports the classes that are implemented or extended and class members that are implemented or overridden in a file.
This notification is not subscribed to by default. Clients can subscribe by including the value "IMPLEMENTED" in the list of services passed in an analysis.setSubscriptions request.
The file with which the implementations are associated.
The classes defined in the file that are implemented or extended.
The member defined in the file that are implemented or overridden.
Reports that the navigation information associated with a region of a single file has become invalid and should be re-requested.
This notification is not subscribed to by default. Clients can subscribe by including the value "INVALIDATE" in the list of services passed in an analysis.setSubscriptions request.
The file whose information has been invalidated.
The offset of the invalidated region.
The length of the invalidated region.
The delta to be applied to the offsets in information that follows the invalidated region in order to update it so that it doesn't need to be re-requested.
Reports the navigation targets associated with a given file.
This notification is not subscribed to by default. Clients can subscribe by including the value "NAVIGATION" in the list of services passed in an analysis.setSubscriptions request.
The file containing the navigation regions.
The navigation regions contained in the file. The regions are sorted by their offsets. Each navigation region represents a list of targets associated with some range. The lists will usually contain a single target, but can contain more in the case of a part that is included in multiple libraries or in Dart code that is compiled against multiple versions of a package. Note that the navigation regions that are returned do not overlap other navigation regions.
The navigation targets referenced in the file. They are referenced by NavigationRegions by their index in this array.
The files containing navigation targets referenced in the file. They are referenced by NavigationTargets by their index in this array.
Reports the occurrences of references to elements within a single file.
This notification is not subscribed to by default. Clients can subscribe by including the value "OCCURRENCES" in the list of services passed in an analysis.setSubscriptions request.
The file in which the references occur.
The occurrences of references to elements within the file.
Reports the outline associated with a single file.
This notification is not subscribed to by default. Clients can subscribe by including the value "OUTLINE" in the list of services passed in an analysis.setSubscriptions request.
The file with which the outline is associated.
The kind of the file.
The name of the library defined by the file using a "library" directive, or referenced by a "part of" directive. If both "library" and "part of" directives are present, then the "library" directive takes precedence. This field will be omitted if the file has neither "library" nor "part of" directives.
The outline associated with the file.
Reports the overriding members in a file.
This notification is not subscribed to by default. Clients can subscribe by including the value "OVERRIDES" in the list of services passed in an analysis.setSubscriptions request.
The file with which the overrides are associated.
The overrides associated with the file.
The code completion domain contains commands related to getting code completion suggestions.
Request that completion suggestions for the given offset in the given file be returned.
The file containing the point at which suggestions are to be made.
The offset within the file at which suggestions are to be made.
The identifier used to associate results with this completion request.
Reports the completion suggestions that should be presented to the user. The set of suggestions included in the notification is always a complete list that supersedes any previously reported suggestions.
The id associated with the completion.
The offset of the start of the text to be replaced. This will be different than the offset used to request the completion suggestions if there was a portion of an identifier before the original offset. In particular, the replacementOffset will be the offset of the beginning of said identifier.
The length of the text to be replaced if the remainder of the identifier containing the cursor is to be replaced when the suggestion is applied (that is, the number of characters in the existing identifier).
The completion suggestions being reported. The notification contains all possible completions at the requested cursor position, even those that do not match the characters the user has already typed. This allows the client to respond to further keystrokes from the user without having to make additional requests.
True if this is that last set of results that will be returned for the indicated completion.
The search domain contains commands related to searches that can be performed against the code base.
Perform a search for references to the element defined or referenced at the given offset in the given file.
An identifier is returned immediately, and individual results will be returned via the search.results notification as they become available.
The file containing the declaration of or reference to the element used to define the search.
The offset within the file of the declaration of or reference to the element.
True if potential matches are to be included in the results.
The identifier used to associate results with this search request.
If no element was found at the given location, this field will be absent, and no results will be reported via the search.results notification.
The element referenced or defined at the given offset and whose references will be returned in the search results.
If no element was found at the given location, this field will be absent.
Perform a search for declarations of members whose name is equal to the given name.
An identifier is returned immediately, and individual results will be returned via the search.results notification as they become available.
The name of the declarations to be found.
The identifier used to associate results with this search request.
Perform a search for references to members whose name is equal to the given name. This search does not check to see that there is a member defined with the given name, so it is able to find references to undefined members as well.
An identifier is returned immediately, and individual results will be returned via the search.results notification as they become available.
The name of the references to be found.
The identifier used to associate results with this search request.
Perform a search for declarations of top-level elements (classes, typedefs, getters, setters, functions and fields) whose name matches the given pattern.
An identifier is returned immediately, and individual results will be returned via the search.results notification as they become available.
The regular expression used to match the names of the declarations to be found.
The identifier used to associate results with this search request.
Return the type hierarchy of the class declared or referenced at the given location.
The file containing the declaration or reference to the type for which a hierarchy is being requested.
The offset of the name of the type within the file.
True if the client is only requesting superclasses and interfaces hierarchy.
A list of the types in the requested hierarchy. The first element of the list is the item representing the type for which the hierarchy was requested. The index of other elements of the list is unspecified, but correspond to the integers used to reference supertype and subtype items within the items.
This field will be absent if the code at the given file and offset does not represent a type, or if the file has not been sufficiently analyzed to allow a type hierarchy to be produced.
Reports some or all of the results of performing a requested search. Unlike other notifications, this notification contains search results that should be added to any previously received search results associated with the same search id.
The id associated with the search.
The search results being reported.
True if this is that last set of results that will be returned for the indicated search.
The edit domain contains commands related to edits that can be applied to the code.
Format the contents of a single file. The currently selected region of text is passed in so that the selection can be preserved across the formatting operation. The updated selection will be as close to matching the original as possible, but whitespace at the beginning or end of the selected region will be ignored. If preserving selection information is not required, zero (0) can be specified for both the selection offset and selection length.
If a request is made for a file which does not exist, or which is not currently subject to analysis (e.g. because it is not associated with any analysis root specified to analysis.setAnalysisRoots), an error of type FORMAT_INVALID_FILE will be generated. If the source contains syntax errors, an error of type FORMAT_WITH_ERRORS will be generated.
The file containing the code to be formatted.
The offset of the current selection in the file.
The length of the current selection in the file.
The line length to be used by the formatter.
The edit(s) to be applied in order to format the code. The list will be empty if the code was already formatted (there are no changes).
The offset of the selection after formatting the code.
The length of the selection after formatting the code.
Return the set of assists that are available at the given location. An assist is distinguished from a refactoring primarily by the fact that it affects a single file and does not require user input in order to be performed.
The file containing the code for which assists are being requested.
The offset of the code for which assists are being requested.
The length of the code for which assists are being requested.
The assists that are available at the given location.
Get a list of the kinds of refactorings that are valid for the given selection in the given file.
The file containing the code on which the refactoring would be based.
The offset of the code on which the refactoring would be based.
The length of the code on which the refactoring would be based.
The kinds of refactorings that are valid for the given selection.
Return the set of fixes that are available for the errors at a given offset in a given file.
The file containing the errors for which fixes are being requested.
The offset used to select the errors for which fixes will be returned.
The fixes that are available for the errors at the given offset.
Get the changes required to perform a refactoring.
If another refactoring request is received during the processing of this one, an error of type REFACTORING_REQUEST_CANCELLED will be generated.
The kind of refactoring to be performed.
The file containing the code involved in the refactoring.
The offset of the region involved in the refactoring.
The length of the region involved in the refactoring.
True if the client is only requesting that the values of the options be validated and no change be generated.
Data used to provide values provided by the user. The structure of the data is dependent on the kind of refactoring being performed. The data that is expected is documented in the section titled Refactorings, labeled as "Options". This field can be omitted if the refactoring does not require any options or if the values of those options are not known.
The initial status of the refactoring, i.e. problems related to the context in which the refactoring is requested. The array will be empty if there are no known problems.
The options validation status, i.e. problems in the given options, such as light-weight validation of a new name, flags compatibility, etc. The array will be empty if there are no known problems.
The final status of the refactoring, i.e. problems identified in the result of a full, potentially expensive validation and / or change creation. The array will be empty if there are no known problems.
Data used to provide feedback to the user. The structure of the data is dependent on the kind of refactoring being created. The data that is returned is documented in the section titled Refactorings, labeled as "Feedback".
The changes that are to be applied to affect the refactoring. This field will be omitted if there are problems that prevent a set of changes from being computed, such as having no options specified for a refactoring that requires them, or if only validation was requested.
The ids of source edits that are not known to be valid. An edit is not known to be valid if there was insufficient type information for the server to be able to determine whether or not the code needs to be modified, such as when a member is being renamed and there is a reference to a member from an unknown type. This field will be omitted if the change field is omitted or if there are no potential edits for the refactoring.
Get the changes required to convert the partial statement at the given location into a syntactically valid statement. If the current statement is already valid the change will insert a newline plus appropriate indentation at the end of the line containing the offset. If a change that makes the statement valid cannot be determined (perhaps because it has not yet been implemented) the statement will be considered already valid and the appropriate change returned.
The file containing the statement to be completed.
The offset used to identify the statement to be completed.
The change to be applied in order to complete the statement.
Will be true if the change contains nothing but whitespace characters, or is empty.
Sort all of the directives, unit and class members of the given Dart file.
If a request is made for a file that does not exist, does not belong to an analysis root or is not a Dart file, SORT_MEMBERS_INVALID_FILE will be generated.
If the Dart file has scan or parse errors, SORT_MEMBERS_PARSE_ERRORS will be generated.
The Dart file to sort.
The file edit that is to be applied to the given file to effect the sorting.
Organizes all of the directives - removes unused imports and sorts directives of the given Dart file according to the Dart Style Guide.
If a request is made for a file that does not exist, does not belong to an analysis root or is not a Dart file, FILE_NOT_ANALYZED will be generated.
If directives of the Dart file cannot be organized, for example because it has scan or parse errors, or by other reasons, ORGANIZE_DIRECTIVES_ERROR will be generated. The message will provide details about the reason.
The Dart file to organize directives in.
The file edit that is to be applied to the given file to effect the organizing.
The execution domain contains commands related to providing an execution or debugging experience.
Create an execution context for the executable file with the given path. The context that is created will persist until execution.deleteContext is used to delete it. Clients, therefore, are responsible for managing the lifetime of execution contexts.
The path of the Dart or HTML file that will be launched, or the path of the directory containing the file.
The identifier used to refer to the execution context that was created.
Delete the execution context with the given identifier. The context id is no longer valid after this command. The server is allowed to re-use ids when they are no longer valid.
The identifier of the execution context that is to be deleted.
Map a URI from the execution context to the file that it corresponds to, or map a file to the URI that it corresponds to in the execution context.
Exactly one of the file and uri fields must be provided. If both fields are provided, then an error of type INVALID_PARAMETER will be generated. Similarly, if neither field is provided, then an error of type INVALID_PARAMETER will be generated.
If the file field is provided and the value is not the path of a file (either the file does not exist or the path references something other than a file), then an error of type INVALID_PARAMETER will be generated.
If the uri field is provided and the value is not a valid URI or if the URI references something that is not a file (either a file that does not exist or something other than a file), then an error of type INVALID_PARAMETER will be generated.
If the contextRoot used to create the execution context does not exist, then an error of type INVALID_EXECUTION_CONTEXT will be generated.
The identifier of the execution context in which the URI is to be mapped.
The path of the file to be mapped into a URI.
The URI to be mapped into a file path.
The file to which the URI was mapped. This field is omitted if the uri field was not given in the request.
The URI to which the file path was mapped. This field is omitted if the file field was not given in the request.
Deprecated: the analysis server no longer fires LAUNCH_DATA events.
Subscribe for services. All previous subscriptions are replaced by the given set of services.
It is an error if any of the elements in the list are not valid services. If there is an error, then the current subscriptions will remain unchanged.
A list of the services being subscribed to.
Reports information needed to allow a single file to be launched.
This notification is not subscribed to by default. Clients can subscribe by including the value "LAUNCH_DATA" in the list of services passed in an execution.setSubscriptions request.
The file for which launch data is being provided. This will either be a Dart library or an HTML file.
The kind of the executable file. This field is omitted if the file is not a Dart file.
A list of the Dart files that are referenced by the file. This field is omitted if the file is not an HTML file.
The diagnostic domain contains server diagnostics APIs.
Return server diagnostics.
The list of analysis contexts.
Return the port of the diagnostic web server. If the server is not running this call will start the server. If unable to start the diagnostic web server, this call will return an error of DEBUG_PORT_COULD_NOT_BE_OPENED.
The diagnostic server port.
This section contains descriptions of the data types referenced in the API’s of the various domains.
A list of fixes associated with a specific error.
Deprecated: the only reference to this type has been deprecated.
A set of options controlling what kind of analysis is to be performed. If the value of a field is omitted the value of the option will not be changed.
An enumeration of the services provided by the analysis domain that are related to a specific list of files.
FOLDINGHIGHLIGHTSIMPLEMENTEDINVALIDATE
This service is not currently implemented and will become a GeneralAnalysisService in a future release.
NAVIGATIONOCCURRENCESOUTLINEOVERRIDESAn indication of the current state of analysis.
An identifier used to associate completion results with a completion request.
Information about an analysis context.
A description of an executable file.
An enumeration of the kinds of executable files.
CLIENTEITHERNOT_EXECUTABLESERVERThe identifier for a execution context.
An enumeration of the services provided by the execution domain.
LAUNCH_DATAAn enumeration of the kinds of files.
LIBRARYPARTAn enumeration of the services provided by the analysis domain that are general in nature (that is, are not specific to some list of files).
ANALYZED_FILESThe hover information associated with a specific location.
A description of a class that is implemented or extended.
A description of a class member that is implemented or overridden.
A description of a member that overrides an inherited member.
A description of a member that is being overridden.
An indication of the current state of pub execution.
An abstract superclass of all refactoring feedbacks.
An abstract superclass of all refactoring options.
An indication of a problem with the execution of the server, typically in response to a request.
An enumeration of the types of errors that can occur in the execution of the server.
CONTENT_MODIFIED
An "analysis.getErrors" or "analysis.getNavigation" request could not be satisfied because the content of the file changed before the requested results could be computed.
DEBUG_PORT_COULD_NOT_BE_OPENED
The server was unable to open a port for the diagnostic server.
FILE_NOT_ANALYZED
A request specified a FilePath which does not match a file in an analysis root, or the requested operation is not available for the file.
FORMAT_INVALID_FILE
An "edit.format" request specified a FilePath which does not match a Dart file in an analysis root.
FORMAT_WITH_ERRORS
An "edit.format" request specified a file that contains syntax errors.
GET_ERRORS_INVALID_FILE
An "analysis.getErrors" request specified a FilePath which does not match a file currently subject to analysis.
GET_NAVIGATION_INVALID_FILE
An "analysis.getNavigation" request specified a FilePath which does not match a file currently subject to analysis.
GET_REACHABLE_SOURCES_INVALID_FILE
An "analysis.getReachableSources" request specified a FilePath which does not match a file currently subject to analysis.
INVALID_ANALYSIS_ROOT
A path passed as an argument to a request (such as analysis.reanalyze) is required to be an analysis root, but isn't.
INVALID_EXECUTION_CONTEXT
The context root used to create an execution context does not exist.
INVALID_FILE_PATH_FORMAT
The format of the given file path is invalid, e.g. is not absolute and normalized.
INVALID_OVERLAY_CHANGE
An "analysis.updateContent" request contained a ChangeContentOverlay object which can't be applied, due to an edit having an offset or length that is out of range.
INVALID_PARAMETER
One of the method parameters was invalid.
INVALID_REQUEST
A malformed request was received.
ORGANIZE_DIRECTIVES_ERROR
An "edit.organizeDirectives" request specified a Dart file that cannot be analyzed. The reason is described in the message.
REFACTORING_REQUEST_CANCELLED
Another refactoring request was received during processing of this one.
SERVER_ALREADY_STARTED
The analysis server has already been started (and hence won't accept new connections).
This error is included for future expansion; at present the analysis server can only speak to one client at a time so this error will never occur.
SERVER_ERROR
An internal error occurred in the analysis server. Also see the server.error notification.
SORT_MEMBERS_INVALID_FILE
An "edit.sortMembers" request specified a FilePath which does not match a Dart file in an analysis root.
SORT_MEMBERS_PARSE_ERRORS
An "edit.sortMembers" request specified a Dart file that has scan or parse errors.
UNANALYZED_PRIORITY_FILES
An "analysis.setPriorityFiles" request includes one or more files that are not being analyzed.
This is a legacy error; it will be removed before the API reaches version 1.0.
UNKNOWN_REQUEST
A request was received which the analysis server does not recognize, or cannot handle in its current configuration.
UNKNOWN_SOURCE
The analysis server was requested to perform an action on a source that does not exist.
UNSUPPORTED_FEATURE
The analysis server was requested to perform an action which is not supported.
This is a legacy error; it will be removed before the API reaches version 1.0.
An identifier used to associate search results with a search request.
A single result from a search request.
An enumeration of the kinds of search results returned by the search domain.
DECLARATION
The declaration of an element.
INVOCATION
The invocation of a function or method.
READ
A reference to a field, parameter or variable where it is being read.
READ_WRITE
A reference to a field, parameter or variable where it is being read and written.
REFERENCE
A reference to an element.
UNKNOWN
Some other kind of search result.
WRITE
A reference to a field, parameter or variable where it is being written.
An enumeration of the services provided by the server domain.
STATUSA representation of a class in a type hierarchy.
This section contains additional information for each kind of refactoring. In addition to a brief description of the refactoring, there is a specification of the feedback that is provided when a refactoring is requested using the edit.getRefactoring request (designed to improve the UX) and the options that may be provided to edit.getRefactoring.
Convert a getter into a method by removing the keyword get and adding an empty parameter list.
It is an error if the range contains anything other than all or part of the name of a single getter.
Convert a method into a getter by adding the keyword get and removing the parameter list.
It is an error if the range contains anything other than all or part of the name of a single method or if the method has a non-empty parameter list.
Create a local variable initialized by the expression that covers the specified selection.
It is an error if the selection range is not covered by a complete expression.
The offsets of the expressions that cover the specified selection, from the down most to the up most.
The lengths of the expressions that cover the specified selection, from the down most to the up most.
The proposed names for the local variable.
The offsets of the expressions that would be replaced by a reference to the variable.
The lengths of the expressions that would be replaced by a reference to the variable. The lengths correspond to the offsets. In other words, for a given expression, if the offset of that expression is offsets[i], then the length of that expression is lengths[i].
The name that the local variable should be given.
True if all occurrences of the expression within the scope in which the variable will be defined should be replaced by a reference to the local variable. The expression used to initiate the refactoring will always be replaced.
Create a method whose body is the specified expression or list of statements, possibly augmented with a return statement.
It is an error if the range contains anything other than a complete expression (no partial expressions are allowed) or a complete sequence of statements.
The offset to the beginning of the expression or statements that will be extracted.
The length of the expression or statements that will be extracted.
The proposed return type for the method. If the returned element does not have a declared return type, this field will contain an empty string.
The proposed names for the method.
True if a getter could be created rather than a method.
The proposed parameters for the method.
The offsets of the expressions or statements that would be replaced by an invocation of the method.
The lengths of the expressions or statements that would be replaced by an invocation of the method. The lengths correspond to the offsets. In other words, for a given expression (or block of statements), if the offset of that expression is offsets[i], then the length of that expression is lengths[i].
The return type that should be defined for the method.
True if a getter should be created rather than a method. It is an error if this field is true and the list of parameters is non-empty.
The name that the method should be given.
The parameters that should be defined for the method.
It is an error if a REQUIRED or NAMED parameter follows a POSITIONAL parameter. It is an error if a REQUIRED or POSITIONAL parameter follows a NAMED parameter.
True if all occurrences of the expression or statements should be replaced by an invocation of the method. The expression or statements used to initiate the refactoring will always be replaced.
Inline the initializer expression of a local variable in place of any references to that variable.
It is an error if the range contains anything other than all or part of the name of a single local variable.
The name of the variable being inlined.
The number of times the variable occurs.
Inline a method in place of one or all references to that method.
It is an error if the range contains anything other than all or part of the name of a single method.
The name of the class enclosing the method being inlined. If not a class member is being inlined, this field will be absent.
The name of the method (or function) being inlined.
True if the declaration of the method is selected. So all references should be inlined.
True if the method being inlined should be removed. It is an error if this field is true and inlineAll is false.
True if all invocations of the method should be inlined, or false if only the invocation site used to create this refactoring should be inlined.
Move the given file and update all of the references to that file and from it. The move operation is supported in general case - for renaming a file in the same folder, moving it to a different folder or both.
The refactoring must be activated before an actual file moving operation is performed.
The "offset" and "length" fields from the request are ignored, but the file specified in the request specifies the file to be moved.
The new file path to which the given file is being moved.
Rename a given element and all of the references to that element.
It is an error if the range contains anything other than all or part of the name of a single function (including methods, getters and setters), variable (including fields, parameters and local variables), class or function type.
The offset to the beginning of the name selected to be renamed.
The length of the name selected to be renamed.
The human-readable description of the kind of element being renamed (such as "class" or "function type alias").
The old name of the element before the refactoring.
The name that the element should have after the refactoring.
This section contains a list of all of the errors that are produced by the server and the data that is returned with each.
TODO: TBD