[Support] Add MIR YAML parsing infrastructure

Introduce source ranges, deterministic text diagnostics, and an independently authored YAML stream parser for the fixed LLVM MIR container shapes. Record the compatibility matrix and Stage 6 validation state.

1 files changed, 91 insertions, 0 deletions

diff --git a/backend/lib/Support/TextDiagnostic.cpp b/backend/lib/Support/TextDiagnostic.cpp
new file mode 100644
index 0000000..1c027c3
--- /dev/null
+++ b/backend/lib/Support/TextDiagnostic.cpp
@@ -0,0 +1,91 @@
+// SPDX-License-Identifier: GPL-3.0-or-later
+
+#include "rcc/Support/TextDiagnostic.h"
+
+#include <algorithm>
+#include <iterator>
+#include <sstream>
+#include <utility>
+
+namespace rcc {
+
+std::string_view getDiagnosticSeverityName(DiagnosticSeverity Severity) noexcept {
+  switch (Severity) {
+  case DiagnosticSeverity::Note:
+    return "note";
+  case DiagnosticSeverity::Warning:
+    return "warning";
+  case DiagnosticSeverity::Error:
+    return "error";
+  }
+
+  return "unknown";
+}
+
+TextSource::TextSource(std::string Name, std::string Text)
+    : Name(std::move(Name)), Text(std::move(Text)), LineOffsets{0} {
+  for (std::size_t Index = 0; Index < this->Text.size(); ++Index) {
+    if (this->Text[Index] == '\n')
+      LineOffsets.push_back(Index + 1);
+  }
+}
+
+TextPosition TextSource::positionAt(std::size_t Offset) const noexcept {
+  Offset = std::min(Offset, Text.size());
+  const auto LineIterator = std::upper_bound(LineOffsets.begin(), LineOffsets.end(), Offset);
+  const std::size_t LineIndex =
+      static_cast<std::size_t>(std::distance(LineOffsets.begin(), LineIterator) - 1);
+  return TextPosition{
+      .Offset = Offset,
+      .Line = LineIndex + 1,
+      .Column = Offset - LineOffsets[LineIndex] + 1,
+  };
+}
+
+TextRange TextSource::range(std::size_t Begin, std::size_t End) const noexcept {
+  Begin = std::min(Begin, Text.size());
+  End = std::clamp(End, Begin, Text.size());
+  return TextRange{.Begin = positionAt(Begin), .End = positionAt(End)};
+}
+
+std::string_view TextSource::line(std::size_t OneBasedLine) const noexcept {
+  if (OneBasedLine == 0 || OneBasedLine > LineOffsets.size())
+    return {};
+
+  const std::size_t Begin = LineOffsets[OneBasedLine - 1];
+  std::size_t End = OneBasedLine < LineOffsets.size() ? LineOffsets[OneBasedLine] - 1 : Text.size();
+  if (End > Begin && Text[End - 1] == '\r')
+    --End;
+  return std::string_view(Text).substr(Begin, End - Begin);
+}
+
+std::string formatTextDiagnostic(const TextDiagnostic &Diagnostic, const TextSource &Source) {
+  std::ostringstream Output;
+  const TextPosition Position = Diagnostic.Range.Begin;
+  const std::string_view SourceName =
+      Diagnostic.SourceName.empty() ? Source.name() : std::string_view(Diagnostic.SourceName);
+
+  Output << SourceName << ':' << Position.Line << ':' << Position.Column << ": "
+         << getDiagnosticSeverityName(Diagnostic.Severity);
+  if (!Diagnostic.Category.empty())
+    Output << '[' << Diagnostic.Category << ']';
+  Output << ": " << Diagnostic.Message;
+
+  const std::string_view SourceLine = Source.line(Position.Line);
+  if (SourceLine.empty())
+    return Output.str();
+
+  Output << '\n' << SourceLine << '\n';
+  const std::size_t PrefixWidth = Position.Column > 0 ? Position.Column - 1 : 0;
+  for (std::size_t Index = 0; Index < PrefixWidth && Index < SourceLine.size(); ++Index)
+    Output << (SourceLine[Index] == '\t' ? '\t' : ' ');
+
+  Output << '^';
+  if (Diagnostic.Range.End.Line == Position.Line &&
+      Diagnostic.Range.End.Column > Position.Column + 1) {
+    Output << std::string(Diagnostic.Range.End.Column - Position.Column - 1, '~');
+  }
+  return Output.str();
+}
+
+} // namespace rcc