Add support for autodetecting URLs and making hyperlinks

.github/actions/spell-check/dictionary/apis.txt

@@ -45,6 +45,7 @@ rx
 serializer
 SIZENS
 spsc
+sregex
 STDCPP
 syscall
 tmp

src/buffer/out/textBuffer.cpp

@@ -35,7 +35,8 @@ TextBuffer::TextBuffer(const COORD screenBufferSize,
     _unicodeStorage{},
     _renderTarget{ renderTarget },
     _size{},
-    _currentHyperlinkId{ 1 }
+    _currentHyperlinkId{ 1 },
+    _currentPatternId{ 0 }
 {
     // initialize ROWs
     for (size_t i = 0; i < static_cast<size_t>(screenBufferSize.Y); ++i)
@@ -2353,3 +2354,86 @@ void TextBuffer::CopyHyperlinkMaps(const TextBuffer& other)
     _hyperlinkMap = other._hyperlinkMap;
     _hyperlinkCustomIdMap = other._hyperlinkCustomIdMap;
 }
+
+// Method Description:
+// - Adds a regex pattern we should search for
+// - The searching does not happen here, we only search when asked to by TerminalCore
+// Arguments:
+// - The regex pattern
+// Return value:
+// - An ID that the caller should associate with the given pattern
+const size_t TextBuffer::AddPatternRecognizer(const std::string regexString)
+{
+    ++_currentPatternId;
+    _IdsAndPatterns.emplace(std::make_pair(_currentPatternId, regexString));
+    return _currentPatternId;
+}
+
+// Method Description:
+// - Finds patterns within the requested region of the text buffer
+// Arguments:
+// - The firstRow to start searching from
+// - The lastRow to search
+// Return value:
+// - A vector containing 3-element tuples where
+//      tuple<0>: ID of the pattern
+//      tuple<1>: start coordinate of the pattern
+//      tuple<2>: end coordinate of the pattern
+const std::vector<std::tuple<size_t, COORD, COORD>> TextBuffer::UpdatePatterns(const size_t firstRow, const size_t lastRow) const
+{
+    std::vector<std::tuple<size_t, COORD, COORD>> result;
+
+    std::wstring concatAll;
+    const auto rowSize = GetRowByOffset(0).size();
+    concatAll.reserve(rowSize * (lastRow - firstRow + 1));
+
+    // to deal with text that spans multiple lines, we will first concatenate
+    // all the text into one string and find the pattern in that string
+    // later, we convert the locations of the patterns back into (row, col) coords
+    for (auto i = firstRow; i <= lastRow; ++i)
+    {
+        auto row = GetRowByOffset(i);
+        concatAll += row.GetCharRow().GetText();
+    }
+
+    // convert the string into something the regex iterator can work with
+    const auto constAll = til::u16u8(concatAll);
+
+    // for each pattern we know of, iterate through the string
+    for (auto idAndPattern : _IdsAndPatterns)
+    {
+        std::regex regexObj{ idAndPattern.second };
+
+        // search through the run with our regex object
+        auto words_begin = std::sregex_iterator(constAll.begin(), constAll.end(), regexObj);
+        auto words_end = std::sregex_iterator();
+
+        size_t lenUpToThis = 0;
+        for (auto i = words_begin; i != words_end; ++i)
+        {
+            // record the locations and convert them back to coords
+            // when we find a match, the prefix is text that is between this
+            // match and the previous match, and the suffix is the text that
+            // it between this match and the next match
+            // we will use that, along with the size of the match, to determine
+            // the locations
+            const auto prefixSize = i->prefix().str().size();
+            const auto start = lenUpToThis + prefixSize;
+            const auto end = start + i->str().size();
+            lenUpToThis = end;
+
+            // store the locations as (col, row) coordinates
+            // NOTE: these are VIEWPORT coordinates, not buffer coordinates
+            // Keeping these as viewport coordinates for now because its the renderer
+            // that actually uses these coordinates and the renderer works in viewport coords
+            const auto startRow = gsl::narrow<SHORT>(start / rowSize);
+            const auto startCol = gsl::narrow<SHORT>(start % rowSize);
+            const auto endRow = gsl::narrow<SHORT>(end / rowSize);
+            const auto endCol = gsl::narrow<SHORT>(end % rowSize);
+            const COORD startCoord{ startCol, startRow };
+            const COORD endCoord{ endCol, endRow };
+            result.push_back(std::make_tuple(idAndPattern.first, startCoord, endCoord));
+        }
+    }
+    return result;
+}

src/buffer/out/textBuffer.hpp

@@ -60,6 +60,8 @@ filling in the last row, and updating the screen.
 
 #include "../renderer/inc/IRenderTarget.hpp"
 
+#include <regex>
+
 class TextBuffer final
 {
 public:
@@ -182,6 +184,9 @@ class TextBuffer final
                           const std::optional<Microsoft::Console::Types::Viewport> lastCharacterViewport,
                           std::optional<std::reference_wrapper<PositionInformation>> positionInfo);
 
+    const size_t AddPatternRecognizer(const std::string regexString);
+    const std::vector<std::tuple<size_t, COORD, COORD>> UpdatePatterns(const size_t firstRow, const size_t lastRow) const;
+
 private:
     void _UpdateSize();
     Microsoft::Console::Types::Viewport _size;
@@ -229,6 +234,9 @@ class TextBuffer final
 
     void _PruneHyperlinks();
 
+    std::unordered_map<size_t, std::string> _IdsAndPatterns;
+    size_t _currentPatternId;
+
 #ifdef UNIT_TESTING
     friend class TextBufferTests;
     friend class UiaTextRangeTests;

src/cascadia/TerminalControl/TermControl.cpp

@@ -35,6 +35,9 @@ constexpr const auto ScrollBarUpdateInterval = std::chrono::milliseconds(8);
 // The minimum delay between updating the TSF input control.
 constexpr const auto TsfRedrawInterval = std::chrono::milliseconds(100);
 
+// The minimum delay between updating the locations of regex patterns
+constexpr const auto UpdatePatternLocationsInterval = std::chrono::milliseconds(50);
+
 DEFINE_ENUM_FLAG_OPERATORS(winrt::Microsoft::Terminal::TerminalControl::CopyFormat);
 
 namespace winrt::Microsoft::Terminal::TerminalControl::implementation
@@ -103,6 +106,7 @@ namespace winrt::Microsoft::Terminal::TerminalControl::implementation
         // This event is explicitly revoked in the destructor: does not need weak_ref
         auto onReceiveOutputFn = [this](const hstring str) {
             _terminal->Write(str);
+            _updatePatternLocations->Run();
         };
         _connectionOutputEventToken = _connection.TerminalOutput(onReceiveOutputFn);
 
@@ -142,6 +146,16 @@ namespace winrt::Microsoft::Terminal::TerminalControl::implementation
             TsfRedrawInterval,
             Dispatcher());
 
+        _updatePatternLocations = std::make_shared<ThrottledFunc<>>(
+            [weakThis = get_weak()]() {
+                if (auto control{ weakThis.get() })
+                {
+                    control->UpdatePatternLocations();
+                }
+            },
+            TsfRedrawInterval,
+            Dispatcher());
+
         _updateScrollBar = std::make_shared<ThrottledFunc<ScrollBarUpdate>>(
             [weakThis = get_weak()](const auto& update) {
                 if (auto control{ weakThis.get() })
@@ -1477,6 +1491,15 @@ namespace winrt::Microsoft::Terminal::TerminalControl::implementation
         return _DoMouseWheel(location, modifiers, delta, state);
     }
 
+    // Method Description:
+    // - Tell TerminalCore to update its knowledge about the locations of visible regex patterns
+    // - We should call this (through the throttled function) when something causes the visible
+    //   region to change, such as when new text enters the buffer or the viewport is scrolled
+    void TermControl::UpdatePatternLocations()
+    {
+        _terminal->UpdatePatterns();
+    }
+
     // Method Description:
     // - Adjust the opacity of the acrylic background in response to a mouse
     //   scrolling event.

src/cascadia/TerminalControl/TermControl.h

@@ -118,6 +118,8 @@ namespace winrt::Microsoft::Terminal::TerminalControl::implementation
 
         bool OnMouseWheel(const Windows::Foundation::Point location, const int32_t delta, const bool leftButtonDown, const bool midButtonDown, const bool rightButtonDown);
 
+        void UpdatePatternLocations();
+
         ~TermControl();
 
         Windows::UI::Xaml::Automation::Peers::AutomationPeer OnCreateAutomationPeer();
@@ -179,6 +181,8 @@ namespace winrt::Microsoft::Terminal::TerminalControl::implementation
 
         std::shared_ptr<ThrottledFunc<>> _tsfTryRedrawCanvas;
 
+        std::shared_ptr<ThrottledFunc<>> _updatePatternLocations;
+
         struct ScrollBarUpdate
         {
             std::optional<double> newValue;

src/cascadia/TerminalCore/Terminal.cpp

@@ -77,6 +77,9 @@ void Terminal::Create(COORD viewportSize, SHORT scrollbackLines, IRenderTarget&
     const TextAttribute attr{};
     const UINT cursorSize = 12;
     _buffer = std::make_unique<TextBuffer>(bufferSize, attr, cursorSize, renderTarget);
+    // Add regex pattern recognizers to the buffer
+    // For now, we only add the URI regex pattern
+    _hyperlinkPatternId = _buffer->AddPatternRecognizer(R"([a-z]+[:.].*?(?=\s))");
 }
 
 // Method Description:
@@ -960,6 +963,45 @@ void Terminal::_NotifyTerminalCursorPositionChanged() noexcept
     }
 }
 
+// Method Description:
+// - Implements the logic to determine if a location is in a given region
+// Arguments:
+// - The location
+// - The start and end coordinates of the region
+// Return value:
+// - True if the location is within those coordinates, false otherwise
+bool Microsoft::Terminal::Core::Terminal::_IsLocationWithinCoordinates(const COORD location, const COORD first, const COORD second) const noexcept
+{
+    if (first.Y == second.Y)
+    {
+        const auto sameRow = location.Y == first.Y;
+        const auto inRange = (first.X <= location.X && location.X < second.X);
+        return (sameRow && inRange);
+    }
+    else
+    {
+        // check first row
+        if (location.Y == first.Y && (first.X <= location.X))
+        {
+            return true;
+        }
+        // check rows in between first row and last row
+        for (auto curRow = first.Y + 1; curRow < second.Y; ++curRow)
+        {
+            if (location.Y == curRow)
+            {
+                return true;
+            }
+        }
+        // check last row
+        if (location.Y == second.Y && location.X < second.X)
+        {
+            return true;
+        }
+    }
+    return false;
+}
+
 void Terminal::SetWriteInputCallback(std::function<void(std::wstring&)> pfn) noexcept
 {
     _pfnWriteInput.swap(pfn);
@@ -1032,6 +1074,15 @@ bool Terminal::IsCursorBlinkingAllowed() const noexcept
     return cursor.IsBlinkingAllowed();
 }
 
+// Method Description:
+// - Update our internal knowledge about where regex patterns are on the screen
+// - This is called by TerminalControl (through a throttled function) when the visible
+//   region changes (for example by text entering the buffer or scrolling)
+void Microsoft::Terminal::Core::Terminal::UpdatePatterns() noexcept
+{
+    _patternsAndLocations = _buffer->UpdatePatterns(_VisibleStartIndex(), _VisibleEndIndex());
+}
+
 const std::optional<til::color> Terminal::GetTabColor() const noexcept
 {
     return _tabColor;

src/cascadia/TerminalCore/Terminal.hpp

@@ -160,6 +160,7 @@ class Microsoft::Terminal::Core::Terminal final :
     const bool IsGridLineDrawingAllowed() noexcept override;
     const std::wstring GetHyperlinkUri(uint16_t id) const noexcept override;
     const std::wstring GetHyperlinkCustomId(uint16_t id) const noexcept override;
+    const size_t GetPatternId(const COORD location) const noexcept override;
 #pragma endregion
 
 #pragma region IUiaData
@@ -185,6 +186,8 @@ class Microsoft::Terminal::Core::Terminal final :
     void SetCursorOn(const bool isOn);
     bool IsCursorBlinkingAllowed() const noexcept;
 
+    void UpdatePatterns() noexcept;
+
     const std::optional<til::color> GetTabColor() const noexcept;
 
 #pragma region TextSelection
@@ -229,6 +232,9 @@ class Microsoft::Terminal::Core::Terminal final :
     bool _altGrAliasing;
     bool _suppressApplicationTitle;
 
+    size_t _hyperlinkPatternId;
+    std::vector<std::tuple<size_t, COORD, COORD>> _patternsAndLocations;
+
 #pragma region Text Selection
     // a selection is represented as a range between two COORDs (start and end)
     // the pivot is the COORD that remains selected when you extend a selection in any direction
@@ -302,6 +308,8 @@ class Microsoft::Terminal::Core::Terminal final :
 
     void _NotifyTerminalCursorPositionChanged() noexcept;
 
+    bool _IsLocationWithinCoordinates(const COORD location, const COORD first, const COORD second) const noexcept;
+
 #pragma region TextSelection
     // These methods are defined in TerminalSelection.cpp
     std::vector<SMALL_RECT> _GetSelectionRects() const noexcept;

src/cascadia/TerminalCore/terminalrenderdata.cpp

@@ -131,6 +131,24 @@ const std::wstring Microsoft::Terminal::Core::Terminal::GetHyperlinkCustomId(uin
     return _buffer->GetCustomIdFromId(id);
 }
 
+// Method Description:
+// - Gets the regex pattern id of a location
+// Arguments:
+// - The location
+// Return value:
+// - The pattern ID of the location
+const size_t Microsoft::Terminal::Core::Terminal::GetPatternId(const COORD location) const noexcept
+{
+    for (auto found : _patternsAndLocations)
+    {
+        if (_IsLocationWithinCoordinates(location, std::get<1>(found), std::get<2>(found)))
+        {
+            return std::get<0>(found);
+        }
+    }
+    return 0;
+}
+
 std::vector<Microsoft::Console::Types::Viewport> Terminal::GetSelectionRects() noexcept
 try
 {

src/host/renderData.cpp

@@ -348,6 +348,12 @@ const std::wstring RenderData::GetHyperlinkCustomId(uint16_t id) const noexcept
     return gci.GetActiveOutputBuffer().GetTextBuffer().GetCustomIdFromId(id);
 }
 
+// For now, we ignore regex patterns in conhost
+const size_t RenderData::GetPatternId(const COORD /*location*/) const noexcept
+{
+    return 0;
+}
+
 // Routine Description:
 // - Converts a text attribute into the RGB values that should be presented, applying
 //   relevant table translation information and preferences.

src/host/renderData.hpp

@@ -58,6 +58,8 @@ class RenderData final :
 
     const std::wstring GetHyperlinkUri(uint16_t id) const noexcept override;
     const std::wstring GetHyperlinkCustomId(uint16_t id) const noexcept override;
+
+    const size_t GetPatternId(const COORD location) const noexcept override;
 #pragma endregion
 
 #pragma region IUiaData

src/host/ut_host/VtIoTests.cpp

@@ -401,6 +401,11 @@ class MockRenderData : public IRenderData, IUiaData
     {
         return {};
     }
+
+    const size_t GetPatternId(const COORD /*location*/) const noexcept
+    {
+        return 0;
+    }
 };
 
 void VtIoTests::RendererDtorAndThread()