* CodeView displays comments as atomic regions. They are either displayed as a whole or not at all. Do nothing halfway!
* added CommentHandler that handles comments detected by clang.
* stored CommentLocations in database.
* saving errors in separate sqlite table
* use MessageShowErrors to show the errors in the UI
* show error count permanently in the right side of the status bar with click to show
* fixed clicking of error locations
* fixed expanding of files with errors
* the codeview now displays a hatched background for a filename when the file's content has changed or the file does not exist anymore.
* added TimePoint class to encapsulate boost.
The version string of the application is stored on every parse. The utility class Version is used to parse the version
string and compare application and stored string. The third number in the version string is now the storage refresh
number, the project will be refreshed when it changes. (0.2.1 -> 0.2.2)
* added status messages that are displayed when sending commands to the IDE.
* added status messages that are displayed when receiving commands from the IDE.
* handled communication from VS to Coati and vice versa.
* one bug still remains: Coati regards a tab as a single character while VS regards it as multiple characters. This causes offsets in communicated positions.
* fixed rpath in mac bundle
* added Qt5Network and Qt5DBus frameworks to bundle
* fixed project settings not saved
* fixed preferences not shown on first start
* added coati_manual.pdf
* When content that is shown in a snippet is located in the file scope, the snippet's title bar displays the file's name. Clicking this name has the same effect as maximizing the file (the file's entire content is shown but the file symbol is not activated).
* extracted code for translating input messages to token activations from the FeatureController to the helper class ActivationTranslator.
* used ActivationTranslator in UndoRedoController to determine if the symbols that have been selected before refresh are still available after parsing again.
* Added a new edge that connects a method of an implicitly specialized template class to its definition in the unspecialized template class.
* Added a test case for this.
* Made methods that have an invalid source location stored as undefined.
* expand toggle also for undefined classes
* changed direction of some template edges
* save same access component to specialized template functions and classes
* added new logos to setup ui
* changed wording in status messages and bundle nodes
* removed ununsed Vector.h
* save NameHierarchy to SearchMatch
* fixed NameHierarchyElement retrieval for name elements without parent
* fixed match weight to score last letters higher
This change passes all file node ids for clearing to the sqlite storage at once and makes use of ON DELETE CASCADE for
everything else, instead of manally deleting all source locations.
* fixed refreshing to send another ui refresh message after parsing
* keepContent field for MessageBase so that views won't update
* fixed searchmatching to only weigh uppercase letters next to lowercase ones
* fixed file not jumping back to snippets when activating edge
* clicking into empty space deselects active edge
* clear file manager files on every fetch
* added db transaction for file clearing
* use zooming instead of changing font size
This change makes MessageActivateTokens being used consistently throughout all Controllers. The FeatureController
accepts all incoming activation messages and turns them into MessageActivateTokens, which are then accepted by all
view controllers.
* showing project name in main window title
* keep reference count visible when maximizing code file
* show first node in db when no main function is available
* log Storage stats after parse
* fixed macro locations saved multiply
* using NameHierarchy when retrieving ids
* don't save macros from external files
* fixed macro and include saving duplicates
* added NODE_UNDEFINED_MACRO
TIL = You need to cut a 1kg cake about 85 times in half until you have a piece with only a single atom.
This change shows the files in the code view minimized, if more than 5. This greatly improves performance, because code
and annotations won't get loaded until needed. Files including declarations or definitions of the active symbol are
still expanded. The number of references in each file is now shown in it's title bar. Clicking the title bar allows for
expanding and collapsing as well now.
Namespace splitting is now handled in active token id retrieval from the Storage, so that namespace nodes are not added
to the Graph at all and complexity in the GraphController is reduced.
This change introduces the HierarchyCache that caches the member relationships of nodes in memory for faster graph
retrieval. Storing these relationships allows for retrieving nodes and edges as lists from the SqliteStorage at once
which results in a 10 fold speed improvement.
This change introduces bucket grid layouting, which sorts the nodes into buckets based on how they are connected. Each
bucket is then layouted individually before they are arranged next to each other. Edges go from left to right, except
for inheritance edges that go from bottom to top.
* readded Storage logs
* turned type name resolver error logs to info logs
* fixed inheritance not parsed in forward declarations
* check file node id before saving source location
* show errors in unparsed files again
* getting all incoming and outgogin edges to children at once in aggregation edge computation
* getting all edges connected to active node at once
* getting all access components for all member edges at once.
* removed header search in subdirectories of header search paths
* cache file node ids in Storage for faster lookup
* retrieve node names from SearchIndex
* fixed addNodeAndAllChildren also called for member edges
* added check if node already present in addEdgeAndAllChildren
* database stores modification time of files.
* FileManager updates internal state based on database info.
* reimplemented FileManager's check if a file is in project.
* implemented clearing data of updated or removed files before reparsing the updated and added files.
templateSpecializationArgsAsWritten returning null caused a crash
VisitMemberExprInDeclBody for vardecl was missing
subdirectories for headersearchpaths, that the user does not need to add a hugh amount of folders
* fixed first match in autocompletion not preselected
* fixed maximize file button extending snippets
* changed names of views and setting name of composite view
* fixed number of error locations
* opens on first start autmatically
* open from the menu under Preferences
* show version number on splash screen
* added recent projects markup to app settings template
* moved all qt window classes to qt/window
* put common form stuff from project setup and app settings into QtSettingsWindow
* realigned buttons and elements
* fixed retina styles for mac
* removed MessageLoadSource and subsequent functionality
* edit existing project through menu
* browsing paths from the line item in QtDirectoryListBox
* grow QtDirectoryListBox instead of scroll
* take any file or directory as valid path
* fixed new project setup screen background
* fixed screen handling on cancelation
This change removes the query syntax from the QtSmartSearchbox and all its classes. Now it only autocompletes to one
single Token, but the UI functionality for composing a query of multiple elements is still left for later use. The struct
SearchMatch is now used in the whole pipeline for passing search results back and forth.
* removed self assignment of id from Token
* removed FilterableGraph interface from Graph
* creating Nodes and Edges on Graph with addNode and addGraph methods
* removed obsolete Aggregation edge filtering
* fixed clang warnings
* fixed memory leak for Node creation
* replaced storage by a new version that utilizes sqlite for persistence.
* added SqliteStorage to manage direct access to the database.
* added testsuit for sqlite.
* slimmed down the protected interface of the storage.
* changed tests in StorageTestSuit to avoid using protected methods of the storage.
* added functions to begin and commit a transaction to the storage.
* removed GraphFilterTestSuite.
* removed GraphFilterConductorTestSuite.
* added Node and Edge constructor that takes id as parameter
* added constructor for TokenComponentNameCached that accepts NameHierarchy as parameter
* added TokenLocation constructor that takes locationId as parameter.
* created respective construction functions in TokenLocatonLine, -File and -Collection.
fortune cookie message = Recognition will come from unexpected sources.
added missing virtual destructor
fixed a warning in token.h: expression with side effects will be evaluated despite being used as an operand to 'typeid'
CMP00043 Cmake warning
added missing override keyword
* aggregation edges only connect toplevel nodes
* fixed aggregation click for undo and redo
* don't show edge name in search field, because they are not searchable
* clear search field when clicking aggregation
* leave active nodes active when clicking edges
This change checks all paths in the ProjectSettings file for relativity and makes them relative to the location of the
ProjectSettings file if necessary. The paths are also correctly saved relative to the file.
This change lets the user click the snippet title in the CodeView to insert the full scope. The existing snippets are
merged with the new lines accordingly.
Put all colors into single xml file. Accessing colors is handled via the singleton class ColorScheme. The default scheme
"bright.xml" holds the original colors. The scheme "dark.xml" is a start, but still a poor alternative. Color schemes
can be switched via the View menu.
This change adds font size and font family to the ApplicationSettings to have a single point for manipulating them. The
font size can also be changed via the menu or the zoomIn and zoomOut shortcuts on all platforms.
This change introdoces a classification of commands by order in the undo stack. Undoing the last action will resend the
message of the last active token change and all messages changing the ui after that.
* saving search query to undo stack with MessageSearch
* saving node and edge with name to undo stack with MessageActivateNode and MessageActivateEdge
* added FeatureController that handels distribution of MessageActivateTokens
* clearing views when refreshing or changing project
* clearing undo stack when changing project
* added StorageCache derived from StorageAccessProxy
This change bundles similar nodes in the graph view together. When the bundled node or it's aggregation edge are clicked,
then the bundle is split. This change also adds an arrow to the aggregation edge based on the direction of it's edges.
* fixed Storage saving specialized template functions as undefined function node
* fixed usages of same type in functions in return or parameter not clickable in graph, because the edges are on top
* moved graph postprocessing to GraphController to avoid reprocessing on edge activation
* animat sceneRect of the GraphView to avoid jump at end of animation
* enabled highDpi icons in QApplication for Searchbar Icons
* realigned StartScreen contents
* showing version number 0.1
* "Search" as placeholder text in SearchBar
* same content margins for each dockwidget view
* brighter variable color in Graph
* press ESC to close StartScreen
* renamed admin directory to php because adblocker blocked ajax
* don't refresh page when pressing enter
* improved english
* added db_config for live page
* save forward declarations of classes and structs as NODE_UNDEFINED_TYPE
* put whole files on top in snippet sorting
* uses aggregation count as weight in graph layouting instead of adding all edges to graph
* refreshing all views of components on MessageRefresh in ComponentManager
* reloading ApplikatinoSettings on MessageRefresh
* smoothing for edge lines
* fixed activation of token not working when multiple tokens active
* changed inheritance arrow
* thicker width for edges on focus
* common base class QtGraphNode
* derived class QtGraphNodeData for nodes representing the data model
* removed GraphNode and GraphEdge interface classes
* moved and renamed files for DummyNode and DummyEdge
* refactored includes throughout Graph UI related files
This change switches the code view UI to use minimize and maximize buttons within the QtCodeView instead of creating
separate windows for each maximized code file. Real icons for the buttons are still missing.
* use DEL key to delete next element
* cursor not jumping to end of word anymore when autocompletions are updated
* holding down arrows for moving to start or end of word won't turn word into element
fortune cookie message = You deserve to have a good time after a hard days work.
* added title for snippets
* title displays name of parent scope.
* when new TokenLocationFiles are created in the storage they are now returned as shared_ptr to ensure correct linking in the contained data.
* Storage::getTokenLocationsForLinesInFile() now always returns pairs of start and end locations.
* Adjusted code snippet view to regard the change described above.
This change ist just an interim solution and does not cover all cases.
The TokenLocation saved for the template argument is the location of the whole expression.
* fixed crash when hovering GraphNode while expanding/collapsing it
* fixed clang warnings
* fixed GraphNodeExpandToggle to show empty circle
* fixed Searchbar showing empty block when nothing was parsed
* made template related edges in GraphView red
* added SnippetMerger class that conducts the merging and snapping of TokenLocations to create snippets.
* added parameters for snippet snapping and merging to ApplicationSettings.
* added forEachStartTokenLocation() and forEachEndTokenLocation() to TokenLocationLine and TokenLocationFile.
* added < and > operators to TokenLocation that compare the location part (not the id).
* darkened red of error message in status bar
* added status error when not defined startupproject
* added tooltips for remaining search view buttons
* added filepath as tooptip for title in code view
* opening file when clicking title in code view
This change reduces the size of collapsed classes in the GraphView by removing the condensed view of the class access
nodes and using a single button next to the class name for collapsing and expanding instead.
* Created CxxNameResolver classes that take care of resolving the name hierarchies of clang components.
* NameResolvers store clang::Decls they ignore when traversing the hierarchy of DeclContext nodes. This approach is used to avoid infinite loops when parsing template parameters of a template class that depends on it's template parameters that depend on the template class....
<description>
This change makes the MessageQueue use Tasks for sending a Message to each MessageListener. Messages can define their
behavior by setting setSendAsTask(), the default is true. Only MessageStatus and MessageInterruptTasks are still sent on
the Messaging thread to allow for immediate effect.
This change also introduced the class SimpleTask which holds only a single perform() callback to override, for Tasks
that are finished in a single step. The class LambdaTask derives from SimpleTask and allows for passing a lambda as the
perform() callback.
This change introduces the TaskScheduler, which queues and processes Tasks on a separate thread. The Task class provides
a common interface for deriving all specific tasks. A task can split it's processing into multiple update calls. The
TaskScheduler will update a task until it is finished or interrupt it when necessary. TaskGroups can be used to bundle
multiple Tasks together. So far only TaskGroupSequential was implemented which runs the Tasks in the set order.
TaskParseCxx utilizes the CxxParser to parse each source file in a single update call. Parsing can be interrupted using
the ESC key. The Statusbar shows the parsing progress.
IMPORTANT: Windows users will have to execute clean, setup and cmake after this commit.
* added script to generate the program icon on windows.
* modified setup script to execute icon generator.
* modified CMakeLists to automatically generate the resource file required for windows builds and link it to the visual studio project.
* fixed test that was expecting a wrong result.
* refactored DataType class.
* added code to parse dependent type names.
* refactored utilityCxx a little.
This change adds the class GraphViewStyle that holds all values for defining the style of Nodes and Edges in the Graph.
The GraphViewStyleImpl interface and the QtGraphViewStyleImpl implementation are used to request Qt specific metrics,
like the charwidth of used fonts.
This change includes the following fixes:
* aggregation number always in the middle between nodes
* darker color for aggregation edge
* split off QGraphicsLineItem implementations from QtGraphEdge
* fixed some colors in color scheme
fortune cookie message = Sorge dich nicht! Du wirst bekommen was du brauchst.
Having only one access point to the Storage will simplify caching and parallelisation in the future. Also the two
accesses were never really independent of each other anyways, because in some cases they had to access both data sets.
This change facilitates loading Coati into Coati:
* removed error logs for TokenComponent setting
* fixed edge checks regarding template stuff
* separated CommonSettings to be inherited by Project- and ApplicationSettings for common settings
* pluralized SourcePaths setting in xml to allow for loading multiple directories and single files
* added CompilerFlags section to settings
* passing Compiler arguments to Parser via struct Parser::Arguments
* setting -isystem for all header search paths to avoid distinction between "" and <> headers
* parsing all headers from the SourcePaths, even the ones that are not included
* fixed parsing time measurement and display
* added utility.h file for general utility functionality
* Added tests for parameter parsing of template functions.
* Added missing tests for template parameters and arguments.
* Fixed warning when using the "typename" keyword in combination with template template arguments.
* Added code to handle template template default arguments and implemented test case.
* fixed search activating 2 nodes, because of NodeType in query
* fixed search results didn't contain results with same name
* added protected contents to sample and tictactoe code
* fixed CxxParserTests not logging errors, disabled logging for tests with errors
* fixed CxxParserTests not finding system headers, added TestSettings.xml and pass headers to Parser::parseFile()
* name of template class now also contains type information on the template parameters (e.g. typename, class, bool, const std::string&, etc.)
* implemented handling of non-type template parameters and template template parameters.
* implemented lots of tests for these features.
The utility class FilePath wraps an instance of boost::filesystem::path. Using FilePath allows for easy comparision of
file paths on different platforms using absolute or relative paths. Whereever file paths are compared, this wrapper
should be used.
This change adds the class FileRegister, that splits cpp and h files for the Parser and keeps track which headers have
been parsed yet. TranslationUnits get only created for cpp files now, and headers are parsed on the fly. This reduces
parse time by around half.
Fixed several issues with the overlaps resolving algorithm.
Introducing the cell padding in the last graph update caused some issues that are now resolved.
Coincidently the average number of algorithm iterations needed could be reduced, therefore the thing is now faster.
fortune cookie message = All the answers you need are right there in front of you!
Show the right colors in SearchView for the different NodeTypes or QueryNodeTyps
Note: add colors to your Applcationsettings like in the template
fortune cookie message = recognition will come from unexpected sources
This change adds the node NODE_FILE to the graph and the edge EDGE_INCLUDE, which determine the dependencies between
files by parsing the include preprocessor directive. File nodes can be searched and displayed in the GraphView as file
dependency graph. The filter "file" allows for selecting all files. The file depency information is used on project
refresh do determine which files need to get reparsed.
* Changed the TokenLocation of TemplateArguments to have at least a valid file name.
* Extended name of anonymous namespace by the containing file's name.
* Adjusted tests to reflect these changes.
* Fixed template parameters to be resolved to their correct names even if they are (re)definied outside of the class (e.g. for template method definitions).
* Added test code.
This change adds animated transitions to the QtGraphView using sequential and parallel animations. First vanishing nodes
and all edges get faded out, then the remaining nodes are moved and resized, lastly new nodes and all edges are faded in.
fortune cookie message = Sorge dich nicht! Du wirst bekommen, was du brauchst.
This change allows setting log levels on loggers. The apps ConsoleLogger was switched to only logging warnings and
errors the FileLogger still saves all types of logs.
Added CompositView, which implements both View and ViewLayout, for combining Views to one Widget. The layout can be set
to either horizontal or vertical. SearchView, RefreshView and UndoRedoView are combined to one DockWidget by the
ComponentManager now.
fortune cookie message = You'll get more secure and confident with your relationship with your co-workers.
* added FileManager that keeps track of the files analyzed by the project. It therefore checks if files get added, updated or deleted from the source- and include folders.
* refactored ASTVisitor::hasValidLocation() and implemented one version that search for the location in the currently parsed file and one that searches in all the project's source files.
* quickfix: switched order in FileManager so that header files are parsed first.
* created folder for helper classes used in tests.
* added test code for FileManager
* adjusted test code for Parser Tests.
* click in autocompletions list causes search
* right arrow for autocompletion
* changed query operator has from '>' to '.' and operator sub from '.' to '+'
* fixed autocompletions to fit to query
and
* fixed spectral layouting endless looping on unconnected node
* split active and visible setting for DummyNodes
* set and check visible on DummyEdges
* include fix for Mac in MatrixDynamicBase
* fix for disappearing style and crashes in GraphView
* cleaned up DummyNode
fortune cookie message = Du stehst kurz davor etwas Neues zu unternehmen.
* implemented parsing of template argument default parameters.
* added tests for default arguments of template parameters.
fortune cookie message = If you continually give, you will continually have.
This change adds the refresh component that allows for refreshing the source code via UI or shortcut. If automatic
refreshing is activated via the UI, the code is refreshed anytime the window gets focus. The contents of all the
updated source files get removed from Storage, before reparsing them. File dependencies are not respected yet.
* fixed: implicit template specializations will be found even if the original template definition is located in a different file (that gets included).
* implemented correct extraction of template parameters, template arguments and template specialization edges.
* added token locations for template parameters of partial template specializations.
* graph view displays full names of template parameters and template arguments.
* added test code for features
Implemented graph layouting, hybrid approach using spectral and force based layouting
fortune cookie message = Giving credit where it's due ensures loyalty to you.
* added sample code
* bigger and closed arrow for inheritance edge
* fixed code hover
* lightened scope colors
* fixed interrupted scope ranges to show all lines
* fixed edge creation to correctly show recursion
* prevent aggregation edge number from floating around
* scroll to top when changing codesnippets
This change saves the errors while parsing in the Storage and displays them in the CodeView. The error messages can be
read in the tooltip on hovering. The error count is visible in the status bar.
* HOT FIX: typename not always correctly split to name hierarchy
* show node main if available as first active token after parsing
* show all contents of namespaces if searched for
* added missing enum type and enum value usages
* added inheritance for structs
* removed MessageActivateToken and replaced it with MessageActivateTokens
* activating correct nodes for clicks in GraphView and CodeView
* no graph change when clicking an edge, but highlighting the edge instead
* only show active nodes without connections when multiple are activated
* fixed snippet sorting to put file that contains declarations on top
* set sizeHints on CodeView and SearchView to increase their initial size
fortune cookie message = Das Leben wird eine positive Wendung nehmen.
This change adds the creation of aggregation edges to the StorageGraph, which save the connection counts between
children in their parents. Aggregation edges are displayed using a straight line, that is thicker than more connections
it represents with the connection count in the middle.
This change fixes the clang cmake config to use release libs of clang, adds the TicTacToe source files to bin/app/src/
and adds time measuring to the parse call in Project.
This change puts the initial project loading into a MessageLoadProject so that the UI thread is still responding.
Also the StatusBar is now responding to more Messages and it's MessageListeners get added to the MessageQueue in
front of other listeners so that the StatusBar UI can announce actions.
* fixed clang warnings
* fixed cxxparser tests
* added missing checks for valid location in ASTVisitor callbacks for templates
* removed warning when statusbar view is removed
* made font of activated classes bold
* fixed search box not responding to autocompletion tab or enter press
* fixed search box not updated when active token changes
* Implemented basic parsing and storage of template parameter types, template classes and template functions
* Added template specialization edge.
* Implemented handling of partial specializations. Needs some more tweaking.
* Added a lot of test code for template stuff (still needs some cases to be tested)
* Removed the isTemplateParameterType from DataType, since it shouldn't be needed
* Function qualTypeToDataType uses clang PrintingPolicy to fetch the desired format.
* Merged a lot of code for getting the correct names and name hierarchies of definitions
* Removed TokenComponentDataType
* Merged both of the Storage::addTypeEdge(...) functions
* Added "-fno-delayed-template-parsing" flag to parser for parsing test code.
This change moves nested layouting of the Graph to GraphController. The DummyNodes get the correct size assigned, as
well as a visible flag. Mouse interaction with the Graph is now handled via the GraphController, which makes it
easier to undo them or reuse the changed state after graph updates.
This change styles the nodes in the QtGraphView. Since Stylesheets can't be applied to QGraphicsItems, the style
implementation was done in the source files. The containers for public/protected/private nodes are now subclassed from
QtGraphNode as QtGraphNodeAccess, including a helper class QtAccessArrow for indicating the opened/colapsed state.
* fitting boxes to name length
* highlighting active edges and nodes
* hover for edges and nodes
* single click selection for edges and nodes
* edges have transparent second line for easier clicking
* sub structure class members by public/protected/private
* hide completer when elements are updated
* show correct indices in sub query search
* move query elements forward when the cursor is at the end of the box
This change only shows matches in the autocompletion list that are prefiltered by the already entered query. The
autocompletion now also shows up after an operator of type . or > has been entered to show an alphabetical list of
possible tokens.
* replaced typenames (which have been stored as single sting where "::" indicated hierarchy levels) by a vector of strings where each element indicates a hierarchy level.
* introduced this change to the Storage method signatures, the ParseFunction and ParseVariable
* this way the SearchIndex creates a node for each level of the hierarchy.
This change switches the search box in the search view from QLineEdit to the derived QtSmartSearchBox implementation,
which is capeable of displaying tokens of the query as selectable button elements in different colors. It allows for all
the usual mouse and keyboard interactions that the QLineEdit offers.
fortune cookie message = Good news from someone dear is coming soon.
Graph view has now toggle-able subnodes. Subnodes connected to the active node are always visible.
fortune cookie message = You will enjoy true success in whatever you do.
This change allows for parsing of a real codebase by specifying the system header search paths in
ApplicationSettings.xml and the additional header search paths in ProjectSettings.xml.
Both settings files are now documented in ApplicationSettings_template.xml and ProjectSettings_template.xml, which show
all possible settings for each file.
This change also fixes some parsing edge cases that occured.
- QtSearchView was split into QtSearchView and QtSearchBox.
- QtSearchBox contains all Qt elements and can use them purely
- QtSearchView forwards calls from the SearchController to QtSearchBox
- Searching is initiated with MessageSearch to the SearchController
- Autocompletion is initiated with MessageSearchAutocomplete to the SearchController
- The search field is able to create filter queries by only giving autocompletions for the last token in the query
- For named tokens the search field adds their token ids to the query in the form of "A,25" for faster lookup
bug id = #21
This change refactors the GraphController to use the GraphAccess only with a single call that returns a Graph. The Graph
contains all information for visually representing the active Token with all parent and child nodes and edges. This
reduces coupling of the data and logic tiers.
Having the TestStorage as static field of the TestSuites caused problems in test startup, because the TestSuites are
instantiated globally and for some reason the programm has trouble allocating memory when initializing the SearchIndex.
This change stores each Token name in the separate singleton class Dictionary. The Dictionary gives each saved word an
Id and thereby avoids duplicated names. E.g if the constructor method "Graph::Graph" is stored then the word "Graph"
only appears once in memory.
The class SearchIndex is now responsible for the name hierarchy and is instantiated by the Storage. The SearchIndex
builds the name hierarchy using SearchNodes, each holding a Dictionary string reference of the name it holds. E.g if the
names "math::ceil" and "math::floor" are added to the SearchIndex then 3 nodes get created, the SearchNode "math" will
hold the two childs "ceil" and "floor".
The hierarchical graph creation functionality got split off from Graph into the new subclass StorageGraph. The
StorageGraph creates nodes with a passed SearchNode pointer of the name it represents in the SearchIndex. Thereby the
StorageGraph reuses the hierarchical information in the SearchIndex and can create nodes much quicker by avoiding node
searches and name comparisions.
The name information is now stored in the Nodes via the TokenComponentName class, which is subclassed into
TokenComponentNameReferenced and TokenComponentNameCached. The StorageClass creates nodes with the component
TokenComponentNameReferenced, which holds a pointer to the SearchNode instance holding the name. This allows for
retrieving the full name of the node, without using other Nodes int the Graph, which might not be present. If the Node
is copied then the component changes to a TokenComponentNameCached, which holds the full name as a string, so the
memory in the Storage doesn't have to be accessed anymore.
TokenComponentSignature is now only holding an Id of the signature string saved in the Dictionary, which speeds up the
signature comparison. A follow-up will change saving the whole signature as string to reusing the wordIds it is
consisting of.
Lastly the SearchIndex holds basic fuzzy search functionality. A passed query gets compared down the SearchNode
hierarchy as long as matches for each letter are found. Matches must contain all letters of the query. The search is
case-insensitive. If letters are found in front positions, next to each other or written in uppercase they are weighed
higher in the match ranking. The character ':' is also interpreted and found, although the '::' delimiter is not stored.
E.g. the query "m:l" used on the example above will return both "math::floor" and "math::ceil", but "floor" is ranked
higher because the 'l' appears closer to the start.
This change allows the user to use a simple query syntax in the search field to filter the results using names,
operators and predefined filters.
The class QueryTree is capable of parsing a simple query language consiting of QueryNodes, subclassed as
QueryOperators, QueryCommands and QueryTokens. QueryToken identifies a Token by name. QueryCommand represents predefined
filters. QueryOperator defines the syntactic relationship.
QueryNode examples:
"A" -> QueryToken that identifies the Token named A
class -> QueryCommand that filters all Nodes that are classes
. -> QueryOperator that concatenates filters
QueryExamples:
"A".class -> All Tokens named A that are classes
method.(private|protected) -> All methods that are private or protected
"A".member -> All members of A
"A":field.!const -> All fields of A that are not const (':' can be used instead of '.member')
QueryTree checks for correct operator precedence and validity of the query.
All QueryCommands can be found in data/query/QueryCommand.cpp
All QueryOperators can be found in data/query/QueryOperator.cpp
The parsed QueryTree is then passed to the class GraphFilterConductor, which is capable of applying the query on a
Graph. Each part of the query gets assigned a GraphFilter that filters the input Graph in order of node precedence. The
class SubGraph is used as intermediate container, only holding bare pointers to Nodes and Edges. The output Graph holds
all Nodes that match the query.
MessageActivateTokens is used to show all of the results in the CodeView. The GraphView currently only shows the first
Node of the results.
bug id = #6
This change adds the StorageTestSuite that checks whether the ParserClient callbacks cause the right things to be saved.
In order to allow for easier data retrieval the ParserClient callbacks now return the Id of the inserted Token, which
can then be used to retrieve the Token pointer.
This change parses type usages in function and method bodies and saves them with an edge of type EDGE_TYPE_USAGE. This
also includes base class types used in initialization lists of derived class constructors.
This change switches function and method parsing to use the structure ParseFunction, which holds all important values of
the function. This structure is then used also in call and usage parsing to allow distinction between signatures.
With this change typedefs have two TokenLocations, one for the TypeName and one for the UnderlyingType. When the
typedef is used with a variable the typedef is referenced now and not the underlying type.
This change parses the TokenLocations of scopes of classes, structs, functions, methods, namespaces and enums and
saves them in Storage with type LOCATION_SCOPE. The TokenLocations of these nodes are just their name declarations now.
Scope locations are displayed in blue color in the code view.
* Annotations updated when active token Id changes
* FileName in window title
* Maximize button not shown
* Scrolls roughly to the maximized location.
* Maximizing also on double click in code snippet
This change saves only the short names in the node, getName() no only gives the short version. The fullName is assembled
from the parent nodes on demand using the getFullName() method.
This change moves the EdgeComponent system to Token and redefines them as TokenComponents.
* Performance tests on the side showed that using typeid() in getComponent() is faster than dynamically casting every
pointer.
* The hasComponent() method was deliberately left out in the implementation, because it used to look up the right
component and discard it again. This check for presence can also be achieved by just using getComponent().
* The method addComponent() is protected in Token, so that Node and Edge can guard which Component is set on which type
by implementing specific setters for each Component.
* The GraphTestSuite was extended to include tests for the Component implementation (and missing tests for locationIds
in Token and copying of Node and Edge were added).
Changed gui wording in lib and app to view wording, since MVC views do not necessarily deal with GUIs. And moved files
so structure is similar in lib and app.
For easier use with int and other basic types, new logging macros where introduced that use stringstream internally, so
heterogeneous data can be logged using the following syntax:
LOG_INFO_STREAM(<< "hello " << 42 << '!');
* Implemented functions for parsing usages of fields and global variables in ASTBodyVisitor and ASTVisitor.
* Storage adds TokenLocatons and Edges for usages.
* Added tests for parsing usage of fields and global variables.
fortune cookie message = Fame, riches and love are yours for the taking.
The shortcut is added to the new Find menubar item and dispatches a MessageFind, which is handled by the
SearchController. QtSearchView assures via the ViewLayout that the view is visible when setting focus on the search box.
FontMetric specific settings are now applied in the showEvent method, which is executed right before showing the Widget
where the font setting from the stylesheet is already present.
* updated spacings and colors in QtCodeView and QtCodeFile
* using Source Code Pro Regular (had to remove light cut from the Fonts folder because it got always used)
* updated syntax highlighted colors to fit design
* changed colors of highlights
* indent all QtCodeSnippets by maximum line number in QtCodeFile
* fixed syntax highlighting to work with sole multiline comment end
fortune cookie message = Your present plans are going to succeed.
The graph views edges are now clickable. On click the edge and its connected nodes will be displayed.
fortune cookie message = Be prepared to accept an exciting opportunity in the future.
This change makes it possible to create a MacOSX app bundle of either Debug or Release builds:
* The script setup/MacOSX/bundle_install.sh is configured by cmake to be placed in the apps Debug and Release folders.
* Execution of bundle_install.sh creates the bundle file structure and copies binaries, libs, frameworks and resources
into it.
* The script uses install_name_tool to reconfigure the dynamic library paths and framework paths of the binary, so they
won't point to the system libraries, but to the locations in the bundle instead.
* An Info.plist file is also configured by cmake and placed in the bundle by the script. Note: It is named
bundle_info.plist in the Debug and Release folders after configuration so the start of the binary during developement
does not use it.
* When opening the bundle it is assured via setup() in src/app/platform_includes/includesMac.h that the working
directory of the binary is set to .app/Contents/Resources so the paths to the data files are evaluated correctly.
fortune cookie message = Dein Einsatz wird in Kürze belohnt werden.
* removed const m_dimensions in VectorBase and replaced it with N
* changed signature of DummyNode::operator=
* fixed warning for DummyNode wrongly declared as class
Added simple layouting functions (implementations for simple grid and circle layout).
Also the graph displays class members and connections (edges) to or from them "correctly" now.
fortune cookie message = Your exotic ideas will lead you to many exciting new adventures.
This change save typedef expressions to the Storage using a node of type Node::NODE_TYPEDEF and an edge of type
Edge::EDGE_TYPEDEF_OF that leads to the original type, with the qualifiers saved as DataType on the edge.
This change introduces the structure ParseTypeUsage for passing information about type usage from the parser to it's
client. It is used for passing data type and location information about return types and parameter types, which now
get saved to the corresponding edges in the Storage.
* Added Array and Reference Modifier for DataType.
* Created basic tests for DataType.
* Fixed bug where qualifiers of modifiers were not copied correctly.
* Fixed bug where DataType modifiers were applied in the wrong (reverse) order.
Function and method nodes are now distincted by their signature when added to the graph. A different node for each
overloaded declaration is created with the signature as member field.
fortune cookie message = Deine harte Arbeit wird schnell belohnt werden.
- ASTVisitor uses DataType instead of string to pass on type information.
- Edges can have EdgeComponents.
- EdgeComponentDataType stores information that modifies the type a ...TYPE_OF edge points to.
- EdgeComponentDataType can create a complete DataType using it's internal information.
- Fixed a bug where DataTypeModifiers did not apply their qualifiers.
Moved Storage instantiation into the Project. Storage access is still handled via the GraphAccess and LocationAccess
interfaces, but the ComponentFactory now only serves proxy implementations of these access classes. Project is
responsible for setting the subject of the proxies to the current Storage instance.
Added MenuBar action Project->New, which opens a directory dialog for choosing the new source directory to parse the
files from. The ConfigManager can be created empty now and setValue calls can be used to set values.
fortune cookie message = Your troubles will be faded by the luck you will soon have.
This change adds a MenuBar to QtMainWindow and adds an action for opening a different project settings XML file. The
message MessageLoadProject is thereon used to load a new project in the Application and clearing the Storage.
* activate token of most nested location
* show parts of locations with start or end outside of snippet
* (fix for multiple parsing of function and method bodies)
This change parses calls in CxxParser by using the class ASTBodyVisitor which traverses the Stmts and Exprs in the body
of the function and method declarations.
The calls include:
* calls in functions
* calls in methods
* constructor calls
* constructor calls in initialization lists
* implicit constructor calls of fields
* calls to operators
* global constructor calls of global variables
* global function calls for global variables
The calls don't include:
* destructor calls
* some calls to implicit constructors and copy constructors
fortune cookie message = Don't be afraid to take that big step.
By now the graph displays the active node, without neighbours or anything, and clicking a node sets the active token id to the clicked nodes id.
fortune cookie message = Be prepared to accept an exciting opportunity in the future.
This change removes the local std::string variable in the log macros due to possible name collisions with macro callers
in clang. Responsibility of this change is held by Malte.
This change parses inheritance in CxxParser and saves them as EDGE_INHERITANCE in the Storage's graph. Tests for correct
inheritance recognition were added to CxxParserTestSuite.
fortune cookie message = If you continually give, you will continually have.
- Implemented autocompletion for the QtSearchView.
- Added: MessageFinishedParsing that will be dispatched by the Project after the parsing step is done.
- Moved the initial MessageActivateToken from the Project to the Application.
- Added getNamesForNodesWithNamePrefix() to GraphAccess.
- Added a basic SearchView (+ controller and qt implementation)
- SearchView can set the active element of other views by dispatching a message.
- SearchView adjusts its text when the active element is changed by other views.
- Added wrappers for Button and EditBox that allow the usage of simple callback functions.
This change refactors parts of the QtCodeSnippet and places some settings into ApplicationSettings.
Additionally Application- and ProjectSettings were abstracted to the common base class Settings, which takes care of the
getting and setting values of the Configmanager member.
- Renamed MessageActivateTokenLocation to MessageActivateToken, since it uses the Token id (not the TokenLocation id).
- Implemented CodeController to add one snippet for each location where the activated token occurred.
- Added Semaphore to QtThreadedFunctor to prevent an evil race-condition.
- CodeView dispatches message when a token is clicked.
Added some features to QTCodeSnippet:
* Code is now annotated using information from TokenLocationFile
* Start of line numbering can be set
* Clicking on an annotation in the code gives the corresponding Token Id
fortune cookie message = Your kindness will be repaid.
- Replaced CodeAccess with LocationAccess.
- Implemented basic functionality for CodeController.
- Added messaging to the project.
- Added QtThreadedFunctor to enable our Messaging thread to invoke calls on the QtThread.
- Added GraphController.
fortune cookie message = Good news from someone dear is coming soon.
This change switches the view creation architecture from element based creation in the lib to view based creation in the
app. This avoids wrapping of the Qt elements and instead defines interfaces for the different View implementations.
* GuiElementFactory was renamed to GuiFactory and is now responsible for View creation instead of Element creation.
* WidgetWrappers are now used on Views to access their root widget.
* Removed DummyView and DummyController.
* Removed GuiElement and corresponding QtElement files.
* Removed ViewManager, functionality is now handled via MainView.
* Added MainView, where Views are registered via the ViewLayout interface.
* Added QtMainWindow, derived from QMainWindow and instantiated by QtMainView, which displays the Views as QDockWidgets.
fortune cookie message = All your hard work will soon pay off.
This change adds structures for saving TokenLocations in the Storage and improves the location derival in the
ASTVisitor:
* TokenLocation saves an Id of a Token. For each Token two TokenLocations are created for both start- and endpoint. They
keep a reference to each other and know whether they are start or end.
* TokenLocationLine saves all TokenLocations in a line and the lineNumber.
* TokenLocationFile saves all TokenLocationLines of a specific file and the filePath.
* TokenLocationCollection saves multiple TokenLocationFiles.
* TokenLocation can derive it's lineNumber and filePath through references to first TokenLocationLine and from there to
TokenLocationFile.
We decided to stop code reviews for now, because they are too time consuming. We will consider reintroducing them when
the project is at a more mature stage.
Changed old 2d vector template to a generic vector class that can have any amount of dimensions. Added implementations for 2d and 4d and typedefs for int and float.
Updated the unit tests accordingly
fortune cookie message = pray for what you want but work for the things you need
This change removes the ConsoleLogger from the TestSuiteFixture and registers a PlainFileLogger instead, which saves
logs during testing into the file bin/test/data/log/test_log.txt.
This change adds graph data structures Token, Node, Edge and Graph for holding code semantics in the Storage and fills
them with the parsed data. Things to note:
* Token is the base for the classes Edge and Node and holds a unique id.
* Graph is the owner of all Nodes and Edges.
* A Node saves the Edges it is endpoint of.
* An Edge saves the Nodes that are endpoints.
* Nodes and Edges are not subclassed. The different Node and Edge types are distinguished by the enums NodeType and
EdgeType. Type specific fields are therefore available for all types, a check when assigning them prohibits faulty
usage.
* Storage uses the Graph::getNodeHierarchy method, which is not strict about knowing type names beforehand. E.g. if A::B
is passed in and A doesn't exist so far, then node A gets created as NODE_UNDEFINED with B as it's child.
* The Graph methods addNodeAsPlainCopy and addEdgeAsPlainCopy can be used to create a sub-graph without reference to the
original, but with the same ids.
review id = 19
fortune cookie message = Your cheerful outlook is one of your best assets.
Added CodeView interface and QtCodeView implementation for showing syntax highlighted code snippets. This change avoids
using the GuiElement hierarchy by using directly using QtClasses and creating QtCodeView in the QtElementFactory.
modified tests to check different cases more independently.
review id = 23
fortune cookie message = You will have a peace of mind when you talk to an old friend. (Du wirst einem alten Freund begegnen und deine Ruhe finden.)
- added callbacks that will be called when clang encounters typedefs.
- added tests for this feature.
review id = 20
fortune cookie message = New ideas could be profitable.
Detects most elements in the code, currently classes, structs, global variables, fields, functions, methods, namespaces
and enums. The structs ParseLocation and ParseVariable are helper objects to transmit information to the ParserClient.
Currently the parsed information only gets logged in the Storage.
review id = 15
fortune cookie message = In dreams and in life, nothing is impossible.
Added boost components: filesystem and system to the project
environment varibable BOOST_155_DIR targeting the boost_1_55_0 folder is needed
review id = 14
fortune cookie message = when you speak honestly and openly, others truly listen to you
To use the messaging system the classes Message and MessageListener need to get derived. The Singelton MessageQueue is
responsible for managing listeners and sending messages in a threadsafe way.
review id = 9
fortune cookie message = You are on your track to starting something new.
- added abstract class tree for gui elements.
- implemented concrete gui classes that wrap the qt specific content.
- implemented first DummyView which plays nice with the ViewManager.
review id = 12
Enabled logging in unit tests.
Seperated LogManager into LogManager (singleton stuff) and LogManagerImplementation (actual logging stuff)
Implemented simple FileLogger
review id = 13
This change adds CxxParser a class for parsing cxx files. It holds basic functionality to detect classes using the clang
libTooling classes. CxxParser is derived from a common Parser interface, which can later be used for other language
parsers. The Parser forwards detected elements to a ParserClient interface, which is implemented by the Storage class.
The detected classes only get logged so far.
review id = 6
* bd3c63a - (HEAD, origin/class, class) applied second review <Eberhard Graether> (4 minutes ago)
* 3db7d93 - added missing files <Eberhard Graether> (4 days ago)
* 6d09f4e - applied review <Eberhard Graether> (4 days ago)
* d9d3e64 - fixed CMake files <Eberhard Graether> (4 days ago)
* f22a771 - only one Storage class <Eberhard Graether> (4 days ago)
* 6ee9553 - added tests <Eberhard Graether> (4 days ago)
* 0a525e8 - fixed includes <Eberhard Graether> (4 days ago)
* ca3fbb8 - added StorageManager as ParserClient <Eberhard Graether> (4 days ago)
* fa185a5 - renamed ParseLocation to ParseObject <Eberhard Graether> (4 days ago)
* 8b89808 - pass location with ParseLocation object <Eberhard Graether> (4 days ago)
* 548ae3c - renamed clang AST classes <Eberhard Graether> (4 days ago)
* d3b2caf - cleaned includes <Eberhard Graether> (4 days ago)
* f7fdf33 - removed using namespace stmts <Eberhard Graether> (4 days ago)
* 5378ef6 - using ParserClient <Eberhard Graether> (4 days ago)
* b9ce2f6 - working with test file <Eberhard Graether> (4 days ago)
* 5585abc - added CxxParser and ParserClient <Eberhard Graether> (4 days ago)
Clang is usually compiled without runtime type information, so all files that use clang classes can't have rtti. This
change creates a separate CLANG_FILES variable which holds all clang files, so they can be compiled without rtti.
review id = 11
Adds clang and llvm include directories and libraries to the build. The
environment variable CLANG_DIR is used to define the location of llvm
and clang. A simple clang tool implementation was added to assure proper
operation of the libraries.
- switched to !/bin/bash because of issues with echo -e on Mac OS
- improved git-sync-master.sh to also rebase on current feature branch
- improved logs in git-publish.sh
- changed build scripts to execute from /bin
Added cmake build instructions for app, lib and test.
Included Qt via environment variable QT_DIR.
Included CxxTest via environment variable CXX_TEST_DIR.
Added scripts for setup, syncing master and publishing features.
Added dlls for QT on Windows.
This change adds the commit message template in file
.git_commit_message.txt. It will be used after once running:
$ git config commit.template .git_commit_message.txt
reviewer=
review url=
2014-02-04 00:12:58 +01:00
653 changed files with 229757 additions and 1 deletions
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.