[vcpkg] Track parser row/col state in Paragraph (renamed from RawParagraph) (#9987)

7 files changed, 42 insertions, 37 deletions

diff --git a/toolsrc/include/vcpkg-test/util.h b/toolsrc/include/vcpkg-test/util.h
index d8ecc51ea..c15846d05 100644
--- a/toolsrc/include/vcpkg-test/util.h
+++ b/toolsrc/include/vcpkg-test/util.h
@@ -23,6 +23,18 @@ namespace vcpkg::Test
         const std::vector<std::pair<const char*, const char*>>& features = {},
         const std::vector<const char*>& default_features = {});
 
+    inline auto test_parse_control_file(const std::vector<std::unordered_map<std::string, std::string>>& v)
+    {
+        std::vector<vcpkg::Parse::Paragraph> pghs;
+        for (auto&& p : v)
+        {
+            pghs.emplace_back();
+            for (auto&& kv : p)
+                pghs.back().emplace(kv.first, std::make_pair(kv.second, vcpkg::Parse::TextRowCol{}));
+        }
+        return vcpkg::SourceControlFile::parse_control_file("", std::move(pghs));
+    }
+
     std::unique_ptr<vcpkg::StatusParagraph> make_status_pgh(const char* name,
                                                             const char* depends = "",
                                                             const char* default_features = "",
diff --git a/toolsrc/include/vcpkg/binaryparagraph.h b/toolsrc/include/vcpkg/binaryparagraph.h
index 05c42dee4..edb2b6a13 100644
--- a/toolsrc/include/vcpkg/binaryparagraph.h
+++ b/toolsrc/include/vcpkg/binaryparagraph.h
@@ -12,7 +12,7 @@ namespace vcpkg
     struct BinaryParagraph
     {
         BinaryParagraph();
-        explicit BinaryParagraph(Parse::RawParagraph fields);
+        explicit BinaryParagraph(Parse::Paragraph fields);
         BinaryParagraph(const SourceParagraph& spgh,
                         Triplet triplet,
                         const std::string& abi_tag,
diff --git a/toolsrc/include/vcpkg/paragraphparser.h b/toolsrc/include/vcpkg/paragraphparser.h
index 86c3d0228..cc4637402 100644
--- a/toolsrc/include/vcpkg/paragraphparser.h
+++ b/toolsrc/include/vcpkg/paragraphparser.h
@@ -2,6 +2,7 @@
 
 #include <vcpkg/base/expected.h>
 #include <vcpkg/packagespec.h>
+#include <vcpkg/textrowcol.h>
 
 #include <memory>
 #include <string>
@@ -21,25 +22,30 @@ namespace vcpkg::Parse
     template<class P>
     using ParseExpected = vcpkg::ExpectedT<std::unique_ptr<P>, std::unique_ptr<ParseControlErrorInfo>>;
 
-    using RawParagraph = std::unordered_map<std::string, std::string>;
+    using Paragraph = std::unordered_map<std::string, std::pair<std::string, TextRowCol>>;
 
     struct ParagraphParser
     {
-        ParagraphParser(RawParagraph&& fields) : fields(std::move(fields)) {}
+        ParagraphParser(Paragraph&& fields) : fields(std::move(fields)) {}
 
         void required_field(const std::string& fieldname, std::string& out);
-        std::string optional_field(const std::string& fieldname) const;
+        std::string optional_field(const std::string& fieldname);
+        void required_field(const std::string& fieldname, std::pair<std::string&, TextRowCol&> out);
+        void optional_field(const std::string& fieldname, std::pair<std::string&, TextRowCol&> out);
         std::unique_ptr<ParseControlErrorInfo> error_info(const std::string& name) const;
 
     private:
-        RawParagraph&& fields;
+        Paragraph&& fields;
         std::vector<std::string> missing_fields;
     };
 
     ExpectedS<std::vector<std::string>> parse_default_features_list(const std::string& str,
-                                                                    CStringView origin = "<unknown>");
+                                                                    CStringView origin = "<unknown>",
+                                                                    TextRowCol textrowcol = {});
     ExpectedS<std::vector<ParsedQualifiedSpecifier>> parse_qualified_specifier_list(const std::string& str,
-                                                                                    CStringView origin = "<unknown>");
+                                                                                    CStringView origin = "<unknown>",
+                                                                                    TextRowCol textrowcol = {});
     ExpectedS<std::vector<Dependency>> parse_dependencies_list(const std::string& str,
-                                                               CStringView origin = "<unknown>");
+                                                               CStringView origin = "<unknown>",
+                                                               TextRowCol textrowcol = {});
 }
diff --git a/toolsrc/include/vcpkg/paragraphs.h b/toolsrc/include/vcpkg/paragraphs.h
index cffe85af7..02e87a76d 100644
--- a/toolsrc/include/vcpkg/paragraphs.h
+++ b/toolsrc/include/vcpkg/paragraphs.h
@@ -8,11 +8,11 @@
 
 namespace vcpkg::Paragraphs
 {
-    using RawParagraph = Parse::RawParagraph;
+    using Paragraph = Parse::Paragraph;
 
-    ExpectedS<RawParagraph> get_single_paragraph(const Files::Filesystem& fs, const fs::path& control_path);
-    ExpectedS<std::vector<RawParagraph>> get_paragraphs(const Files::Filesystem& fs, const fs::path& control_path);
-    ExpectedS<std::vector<RawParagraph>> parse_paragraphs(const std::string& str, const std::string& origin);
+    ExpectedS<Paragraph> get_single_paragraph(const Files::Filesystem& fs, const fs::path& control_path);
+    ExpectedS<std::vector<Paragraph>> get_paragraphs(const Files::Filesystem& fs, const fs::path& control_path);
+    ExpectedS<std::vector<Paragraph>> parse_paragraphs(const std::string& str, const std::string& origin);
 
     Parse::ParseExpected<SourceControlFile> try_load_port(const Files::Filesystem& fs, const fs::path& control_path);
 
diff --git a/toolsrc/include/vcpkg/parse.h b/toolsrc/include/vcpkg/parse.h
index 3214305bf..539bb2ed7 100644
--- a/toolsrc/include/vcpkg/parse.h
+++ b/toolsrc/include/vcpkg/parse.h
@@ -18,14 +18,20 @@ namespace vcpkg::Parse
 
     struct ParseError : IParseError
     {
-        ParseError(std::string origin, int row, int column, std::string line, std::string message)
-            : origin(std::move(origin)), row(row), column(column), line(std::move(line)), message(std::move(message))
+        ParseError(std::string origin, int row, int column, int caret_col, std::string line, std::string message)
+            : origin(std::move(origin))
+            , row(row)
+            , column(column)
+            , caret_col(caret_col)
+            , line(std::move(line))
+            , message(std::move(message))
         {
         }
 
         const std::string origin;
         const int row;
         const int column;
+        const int caret_col;
         const std::string line;
         const std::string message;
 
@@ -105,27 +111,8 @@ namespace vcpkg::Parse
         const char* it() const { return m_it; }
         char cur() const { return *m_it; }
         SourceLoc cur_loc() const { return {m_it, row, column}; }
-        char next()
-        {
-            char ch = *m_it;
-            // See https://www.gnu.org/prep/standards/standards.html#Errors
-            if (ch == '\t')
-                column = (column + 7) / 8 * 8 + 1; // round to next 8-width tab stop
-            else if (ch == '\n')
-            {
-                row++;
-                column = 1;
-            }
-            else if (ch == '\0')
-            {
-                return '\0';
-            }
-            else
-            {
-                ++column;
-            }
-            return *++m_it;
-        }
+        TextRowCol cur_rowcol() const { return {row, column}; }
+        char next();
         bool at_eof() const { return *m_it == 0; }
 
         void add_error(std::string message) { add_error(std::move(message), cur_loc()); }
diff --git a/toolsrc/include/vcpkg/sourceparagraph.h b/toolsrc/include/vcpkg/sourceparagraph.h
index bdb75fc58..08c306abf 100644
--- a/toolsrc/include/vcpkg/sourceparagraph.h
+++ b/toolsrc/include/vcpkg/sourceparagraph.h
@@ -70,7 +70,7 @@ namespace vcpkg
         }
 
         static Parse::ParseExpected<SourceControlFile> parse_control_file(
-            const fs::path& path_to_control, std::vector<Parse::RawParagraph>&& control_paragraphs);
+            const fs::path& path_to_control, std::vector<Parse::Paragraph>&& control_paragraphs);
 
         std::unique_ptr<SourceParagraph> core_paragraph;
         std::vector<std::unique_ptr<FeatureParagraph>> feature_paragraphs;
diff --git a/toolsrc/include/vcpkg/statusparagraph.h b/toolsrc/include/vcpkg/statusparagraph.h
index ec850607d..2c6d34865 100644
--- a/toolsrc/include/vcpkg/statusparagraph.h
+++ b/toolsrc/include/vcpkg/statusparagraph.h
@@ -30,7 +30,7 @@ namespace vcpkg
     struct StatusParagraph
     {
         StatusParagraph() noexcept;
-        explicit StatusParagraph(Parse::RawParagraph&& fields);
+        explicit StatusParagraph(Parse::Paragraph&& fields);
 
         bool is_installed() const { return want == Want::INSTALL && state == InstallState::INSTALLED; }