Fix minor bugs and add language detection.

README.md

@@ -59,4 +59,5 @@ It should support:
 ### TODO:
 
 - Make "g" command work.
-- properly handle escapes for %q ' etc in ruby parser (rn everything is escapable.)
+- Switch to proper posix style Shard objects (to distinguish 1 empty line from no lines.)
+- Use proper ruby binding context to keep locals across <pre>`</pre> function calls.

include/bed.h

@@ -20,6 +20,8 @@ struct BEd {
   internal::functions::Function eof_op;
   std::array<std::optional<internal::functions::Suffix>, 26> suffixes;
   internal::theme::Theme theme;
+  std::unordered_map<std::string, std::string> extention_to_language;
+  std::unordered_map<std::string, std::string> filename_to_language;
   std::unordered_map<std::string, internal::syntax::Language *> languages;
   internal::io::IO io;
   internal::vase::AppendStorage append{"/tmp"};

include/internal/buffer/decl.h

@@ -1,6 +1,7 @@
 #pragma once
 
 #include "definitions.h"
+#include "internal/syntax/c/parser.h"
 #include "internal/syntax/parser.h"
 #include "internal/syntax/ruby/parser.h"
 #include "internal/theme/theme.h"
@@ -32,7 +33,7 @@ struct Buffer {
   virtual bool waste() = 0;
   virtual uint64_t lines() = 0;
   virtual uint64_t bytes() = 0;
-  virtual void set_filename(std::filesystem::path path) = 0;
+  virtual void set_filename(BEd &ctx, std::filesystem::path path) = 0;
   virtual std::filesystem::path filename() = 0;
   virtual void saved_hook() = 0;
   virtual void load(BEd &ctx, vase::Shard *text) = 0;

include/internal/buffer/types/clip.h

@@ -14,7 +14,7 @@ struct ClipBuffer : Buffer {
   uint64_t lines() override;
   uint64_t bytes() override;
   void load(BEd &ctx, vase::Shard *text) override;
-  void set_filename(std::filesystem::path path) override;
+  void set_filename(BEd &ctx, std::filesystem::path path) override;
   std::filesystem::path filename() override;
   vase::Shard *copy(uint64_t start_line, uint64_t end_line) override;
   void substitute(

include/internal/buffer/types/generic.h

@@ -34,7 +34,7 @@ struct GenericBuffer : ShardBuffer {
   void saved_hook() override;
   bool waste() override;
   void load(BEd &ctx, vase::Shard *text) override;
-  void set_filename(std::filesystem::path path) override;
+  void set_filename(BEd &ctx, std::filesystem::path path) override;
   std::filesystem::path filename() override;
   void substitute(
     BEd &ctx, uint64_t start_line, uint64_t end_line,

include/internal/buffer/types/readonly.h

@@ -19,7 +19,7 @@ struct ReadonlyBuffer : ShardBuffer {
   void load(BEd &, vase::Shard *) override {
     throw ed_error("This buffer is read-only.");
   }
-  void set_filename(std::filesystem::path) override {}
+  void set_filename(BEd &, std::filesystem::path) override {}
   std::filesystem::path filename() override {
     return {};
   }

include/internal/syntax/c/decl.h

@@ -0,0 +1,30 @@
+#pragma once
+
+#include "../decl.h"
+#include "pch.h"
+#include "tries.h"
+
+namespace bed::internal::syntax::c {
+struct alignas(2) CState {
+  struct CInternalState {
+    uint16_t brace_level;
+    enum : uint8_t {
+      NONE,
+      STRING,
+      CHAR,
+      MULTI_COMMENT,
+      INLINE_COMMENT
+    } state;
+    uint8_t flags = 0;
+  };
+
+  uint8_t top;
+  // the stack (RubyInternalState * top)
+
+  inline CInternalState *stack() {
+    return (CInternalState *)(this + 1);
+  }
+};
+
+Language lang_c();
+} // namespace bed::internal::syntax::c

include/internal/syntax/c/parser.h

@@ -0,0 +1,65 @@
+#pragma once
+
+#include "decl.h"
+#include "pch.h"
+
+namespace bed::internal::syntax::c {
+struct CParser {
+  void **v_state;
+  CState *state;
+  std::string_view line;
+  bool still_empty = true;
+  uint32_t i = 0;
+
+  CParser(void **v_state, std::string_view line)
+      : v_state(v_state), state((CState *)*v_state), line(line) {}
+
+  uint32_t len() const {
+    return line.size();
+  }
+
+  CState::CInternalState &current() {
+    return state->stack()[state->top - 1];
+  }
+
+  char peek(int32_t offset = 0) {
+    uint32_t pos = i + offset;
+    return pos < line.size() ? line[pos] : '\0';
+  }
+
+  std::string_view peek_str(uint32_t len) {
+    return line.substr(i, len);
+  }
+
+  void advance() {
+    ++i;
+  }
+
+  void advance(uint32_t n) {
+    i += n;
+  }
+
+  void push_state() {
+    state->top++;
+    state = (CState *)realloc(
+      state,
+      sizeof(CState)
+        + sizeof(CState::CInternalState) * state->top
+    );
+    current() = {
+      .brace_level = 1,
+      .state = CState::CInternalState::NONE,
+      .flags = 0
+    };
+  }
+
+  void pop_state() {
+    state->top--;
+  }
+};
+
+void c_parse(
+  void **v_state, std::string_view line, bool first_line,
+  std::vector<io::Token> *tokens, std::vector<ParseEvent> *events
+);
+} // namespace bed::internal::syntax::c

include/internal/syntax/c/tries.h

@@ -0,0 +1,34 @@
+#pragma once
+
+#include "internal/trie/trie.h"
+#include "pch.h"
+
+namespace bed::internal::syntax::c {
+const static std::vector<std::string> types = {
+  "int",
+  "char",
+  "long",
+  "void",
+  "auto"
+};
+
+const static std::vector<std::string> type_modifier = {
+  "signed",
+  "unsigned",
+  "long"
+};
+
+const static std::vector<std::string> math_operators = {
+  "+",
+  "-",
+  "*"
+};
+
+struct CTries {
+  trie::Trie<void> types_trie;
+  CTries() {
+    for (auto &keyword : types)
+      types_trie.insert(keyword);
+  }
+};
+} // namespace bed::internal::syntax::c

include/internal/syntax/ruby/decl.h

@@ -24,7 +24,8 @@ struct alignas(2) RubyState {
       DEF_NAME = 0b10,
       MODULE_NAME = 0b11
     };
-    static constexpr uint8_t NEWLINE = 1 << 5;
+    static constexpr uint8_t NEWLINE = 1 << 4;
+    static constexpr uint8_t ALLOW_ESCAPE = 1 << 5;
     static constexpr uint8_t ALLOW_INTERPOLATION = 1 << 6;
     static constexpr uint8_t EXPECTING_EXPRESSION = 1 << 7;
     uint8_t flags = 0;

include/internal/vase/shard.h

@@ -6,6 +6,9 @@
 #include "storage/storage.h"
 
 namespace bed::internal::vase {
+// TODO: rewrite with invariant, text ends with a newline, as right now both 1 real and empty line and no text mean teh same thing,
+//       which causes a ton of issues.
+
 struct Shard {
   enum struct Kind : uint8_t {
     Branch,

src/bed/bed.cc

@@ -47,6 +47,14 @@ BEd::BEd(std::vector<std::string> args)
   internal::functions::Suffix::register_suffixes(*this);
   internal::scripting::register_basic(*this);
   languages["ruby"] = new internal::syntax::Language(internal::syntax::ruby::lang_ruby());
+  extention_to_language[".rb"] = "ruby";
+  languages["c"] = new internal::syntax::Language(internal::syntax::c::lang_c());
+  extention_to_language[".h"] = "c";
+  extention_to_language[".hh"] = "c";
+  extention_to_language[".h++"] = "c";
+  extention_to_language[".c"] = "c";
+  extention_to_language[".cc"] = "c";
+  extention_to_language[".c++"] = "c";
   buffers["clip"] = new internal::buffer::ClipBuffer("clip");
   current() = {"default", 0};
   try {

src/internal/buffer/generic.cc

@@ -249,11 +249,17 @@ void GenericBuffer::load(BEd &ctx, vase::Shard *text) {
     ctx.prev().end = text->lines + 1;
   }
   ctx.current() = {name, lines()};
+  syntax::Language *lang = parse.lang;
   syntax::release(parse);
-  parse = syntax::make_parser(root, lines(), nullptr);
+  if (save_path.has_extension()) {
+    auto it = ctx.extention_to_language.find(save_path.extension().string());
+    if (it != ctx.extention_to_language.end())
+      lang = ctx.languages[it->second];
+  }
+  parse = syntax::make_parser(root, lines(), lang);
 }
 
-void GenericBuffer::set_filename(std::filesystem::path path) {
+void GenericBuffer::set_filename(BEd &, std::filesystem::path path) {
   save_path = path;
 }
 

src/internal/buffer/special/clip.cc

@@ -47,7 +47,7 @@ void ClipBuffer::load(BEd &ctx, vase::Shard *text) {
   }
 }
 
-void ClipBuffer::set_filename(std::filesystem::path) {}
+void ClipBuffer::set_filename(BEd &, std::filesystem::path) {}
 
 std::filesystem::path ClipBuffer::filename() {
   return "";

src/internal/functions/posix.cc

@@ -145,7 +145,7 @@ void Function::register_posix(BEd &ctx) {
         if (std::holds_alternative<std::filesystem::path>(arg)) {
           auto path = std::get<std::filesystem::path>(arg);
           s = vase::Shard::from_file(path);
-          buf.set_filename(path);
+          buf.set_filename(ctx, path);
         } else if (std::holds_alternative<ShellArg>(arg)) {
           auto cmd = std::get<ShellArg>(arg).cmd;
           ctx.escape_command(cmd, buf.filename().string());
@@ -191,7 +191,7 @@ void Function::register_posix(BEd &ctx) {
         if (std::holds_alternative<std::filesystem::path>(arg)) {
           auto path = std::get<std::filesystem::path>(arg);
           s = vase::Shard::from_file(path);
-          buf.set_filename(path);
+          buf.set_filename(ctx, path);
         } else if (std::holds_alternative<ShellArg>(arg)) {
           auto cmd = std::get<ShellArg>(arg).cmd;
           ctx.escape_command(cmd, buf.filename().string());
@@ -234,7 +234,7 @@ void Function::register_posix(BEd &ctx) {
         auto &addr = std::get<std::string>(addr_);
         auto &buf = ctx.buffer(addr);
         if (std::holds_alternative<std::filesystem::path>(arg))
-          buf.set_filename(std::get<std::filesystem::path>(arg));
+          buf.set_filename(ctx, std::get<std::filesystem::path>(arg));
         else if (std::holds_alternative<ShellArg>(arg))
           throw ed_error("Can't save shell command as save path.");
         if (buf.filename().empty())
@@ -590,7 +590,7 @@ void Function::register_posix(BEd &ctx) {
           auto path = std::get<std::filesystem::path>(arg);
           s = vase::Shard::from_file(path);
           if (buf.filename().empty())
-            buf.set_filename(path);
+            buf.set_filename(ctx, path);
         } else if (std::holds_alternative<ShellArg>(arg)) {
           auto cmd = std::get<ShellArg>(arg).cmd;
           ctx.escape_command(cmd, buf.filename().string());
@@ -740,7 +740,7 @@ void Function::register_posix(BEd &ctx) {
           if (std::holds_alternative<std::filesystem::path>(arg)) {
             auto path = std::get<std::filesystem::path>(arg);
             vase::write_file(path, text);
-            buf.set_filename(path);
+            buf.set_filename(ctx, path);
           } else if (std::holds_alternative<ShellArg>(arg)) {
             auto cmd = std::get<ShellArg>(arg).cmd;
             ctx.escape_command(cmd, buf.filename().string());

src/internal/syntax/c/core.cc

@@ -0,0 +1,37 @@
+#include "internal/syntax/c/parser.h"
+
+namespace bed::internal::syntax::c {
+Language lang_c() {
+  return Language{
+    .none_state = []() {
+      CState *st = (CState *)malloc(
+        sizeof(CState)
+        + sizeof(CState::CInternalState)
+      );
+      st->top = 1;
+      st->stack()[0] = {
+        .brace_level = 1,
+        .state = CState::CInternalState::NONE,
+        .flags = 0
+      };
+      return st; },
+    .parse = c_parse,
+    .copy = [](void *v_i_st) {
+      CState *i_st = (CState *)v_i_st;
+      uint32_t bytes = sizeof(CState)
+                       + sizeof(CState::CInternalState) * i_st->top;
+      CState *o_st = (CState *)malloc(bytes);
+      memcpy(o_st, i_st, bytes);
+      return o_st; },
+    .equal = [](void *v_a_st, void *v_b_st) {
+      CState *a_st = (CState *)v_a_st;
+      uint32_t a_bytes = sizeof(CState)
+                         + sizeof(CState::CInternalState) * a_st->top;
+      CState *b_st = (CState *)v_b_st;
+      uint32_t b_bytes = sizeof(CState)
+                         + sizeof(CState::CInternalState) * b_st->top;
+      return a_bytes == b_bytes && memcmp(a_st, b_st, a_bytes) == 0; },
+    .destroy = [](void *v_st) { free(v_st); },
+  };
+}
+} // namespace bed::internal::syntax::c

src/internal/syntax/c/parse.cc

@@ -0,0 +1,396 @@
+#include "internal/syntax/c/parser.h"
+
+namespace bed::internal::syntax::c {
+inline bool is_hex(char c) {
+  return ('0' <= c && c <= '9')
+         || ('a' <= c && c <= 'f')
+         || ('A' <= c && c <= 'F');
+};
+
+inline bool identifier_start_char(char c) {
+  return (c & 0x80)
+         || ('a' <= c && c <= 'z')
+         || ('A' <= c && c <= 'Z')
+         || c == '_';
+}
+
+inline bool identifier_char(char c) {
+  return (c & 0x80)
+         || ('a' <= c && c <= 'z')
+         || ('A' <= c && c <= 'Z')
+         || ('0' <= c && c <= '9')
+         || c == '_';
+}
+
+void handle_multiline_comment(CParser &p, std::vector<io::Token> *tokens, std::vector<ParseEvent> *events) {
+  uint32_t start = p.i;
+  while (p.i < p.len()) {
+    if (p.peek() == '*' && p.peek(1) == '/') {
+      p.i += 2;
+      p.current().state = CState::CInternalState::NONE;
+      tokens->push_back({start, p.i, io::Token::Comment});
+      events->push_back(true);
+      return;
+    }
+    p.advance();
+  }
+  tokens->push_back({start, p.i, io::Token::Comment});
+}
+
+void handle_inline_comment(CParser &p, std::vector<io::Token> *tokens, std::vector<ParseEvent> *) {
+  uint32_t start = p.i;
+  p.i = p.len();
+  tokens->push_back({start, p.i, io::Token::Comment});
+  if (p.line.empty() || p.line.back() != '\\')
+    p.current().state = CState::CInternalState::NONE;
+}
+
+bool handle_escapes(CParser &p, std::vector<io::Token> *tokens, uint32_t &start, bool string) {
+  if (p.peek() != '\\')
+    return false;
+  if (string)
+    tokens->push_back({start, p.i, io::Token::String});
+  else
+    tokens->push_back({start, p.i, io::Token::Char});
+  start = p.i;
+  p.advance();
+  if (p.peek() == 'x') {
+    p.advance();
+    while (is_hex(p.peek()))
+      p.advance();
+  } else if (p.peek() == 'u') {
+    p.advance();
+    for (int i = 0; i < 4 && is_hex(p.peek()); i++)
+      p.advance();
+  } else if (p.peek() == 'U') {
+    p.advance();
+    for (int i = 0; i < 8 && is_hex(p.peek()); i++)
+      p.advance();
+  } else if ('0' <= p.peek() && p.peek() <= '7') {
+    p.advance();
+    if ('0' <= p.peek() && p.peek() <= '7')
+      p.advance();
+    if ('0' <= p.peek() && p.peek() <= '7')
+      p.advance();
+  } else {
+    p.advance();
+  }
+  tokens->push_back({start, p.i, io::Token::Escape});
+  start = p.i;
+  return true;
+}
+
+void handle_string(CParser &p, std::vector<io::Token> *tokens, bool string) {
+  uint32_t start = p.i;
+  while (p.i < p.len()) {
+    if (handle_escapes(p, tokens, start, string))
+      continue;
+    if (p.peek() == (string ? '"' : '\'')) {
+      p.advance();
+      if (string)
+        tokens->push_back({start, p.i, io::Token::String});
+      else
+        tokens->push_back({start, p.i, io::Token::Char});
+      p.current().state = CState::CInternalState::NONE;
+      return;
+    }
+    p.advance();
+  }
+  if (p.i >= p.len()) {
+    if (string)
+      tokens->push_back({start, p.len(), io::Token::String});
+    else
+      tokens->push_back({start, p.len(), io::Token::Char});
+    if (p.line.empty() || p.line.back() != '\\')
+      p.current().state = CState::CInternalState::NONE;
+  }
+}
+
+void handle_number(CParser &p, std::vector<io::Token> *tokens) {
+  bool integer = true;
+  uint32_t start = p.i;
+  if (p.peek() == '0') {
+    p.advance();
+    if (p.peek() == 'x' || p.peek() == 'X') {
+      p.advance();
+      if (!(is_hex(p.peek()) || p.peek() == '.'))
+        return;
+      while (is_hex(p.peek())) {
+        p.advance();
+        if (p.peek() == '\'') {
+          p.advance();
+          if (!(is_hex(p.peek())))
+            return;
+        }
+      }
+      bool need_p = false;
+      if (p.peek() == '.') {
+        need_p = true;
+        p.advance();
+      }
+      while (is_hex(p.peek())) {
+        p.advance();
+        if (p.peek() == '\'') {
+          p.advance();
+          if (!(is_hex(p.peek())))
+            return;
+        }
+      }
+      if (p.peek() == 'p' || p.peek() == 'P') {
+        integer = false;
+        p.advance();
+        if (p.peek() == '-' || p.peek() == '+')
+          p.advance();
+        if (!('0' <= p.peek() && p.peek() <= '9'))
+          return;
+        p.advance();
+        while ('0' <= p.peek() && p.peek() <= '9') {
+          p.advance();
+          if (p.peek() == '\'') {
+            p.advance();
+            if (!('0' <= p.peek() && p.peek() <= '9'))
+              return;
+          }
+        }
+      } else if (need_p) {
+        return;
+      }
+    } else if (p.peek() == 'b' || p.peek() == 'B') {
+      p.advance();
+      if (p.peek() != '0' && p.peek() != '1')
+        return;
+      p.advance();
+      while (p.peek() == '0' || p.peek() == '1') {
+        p.advance();
+        if (p.peek() == '\'') {
+          p.advance();
+          if (!(p.peek() == '0' || p.peek() == '1'))
+            return;
+        }
+      }
+    } else {
+      while ('0' <= p.peek() && p.peek() <= '7') {
+        p.advance();
+        if (p.peek() == '\'') {
+          p.advance();
+          if (!('0' <= p.peek() && p.peek() <= '7'))
+            return;
+        }
+      }
+      if (p.peek() == '.') {
+        integer = false;
+        p.advance();
+      }
+      while ('0' <= p.peek() && p.peek() <= '7') {
+        p.advance();
+        if (p.peek() == '\'') {
+          p.advance();
+          if (!('0' <= p.peek() && p.peek() <= '7'))
+            return;
+        }
+      }
+      if (p.peek() == 'e' || p.peek() == 'E') {
+        integer = false;
+        p.advance();
+        if (p.peek() == '-' || p.peek() == '+')
+          p.advance();
+        if (!('0' <= p.peek() && p.peek() <= '9'))
+          return;
+        p.advance();
+        while ('0' <= p.peek() && p.peek() <= '9') {
+          p.advance();
+          if (p.peek() == '\'') {
+            p.advance();
+            if (!('0' <= p.peek() && p.peek() <= '9'))
+              return;
+          }
+        }
+      }
+    }
+  } else if (p.peek() == '.') {
+    integer = false;
+    p.advance();
+    goto after_decimal;
+  } else {
+    p.advance();
+    while ('0' <= p.peek() && p.peek() <= '9') {
+      p.advance();
+      if (p.peek() == '\'') {
+        p.advance();
+        if (!('0' <= p.peek() && p.peek() <= '9'))
+          return;
+      }
+    }
+    if (p.peek() == '.') {
+      integer = false;
+      p.advance();
+    }
+  after_decimal:
+    while ('0' <= p.peek() && p.peek() <= '9') {
+      p.advance();
+      if (p.peek() == '\'') {
+        p.advance();
+        if (!('0' <= p.peek() && p.peek() <= '9'))
+          return;
+      }
+    }
+    if (p.peek() == 'e' || p.peek() == 'E') {
+      integer = false;
+      p.advance();
+      if (p.peek() == '-' || p.peek() == '+')
+        p.advance();
+      if (!('0' <= p.peek() && p.peek() <= '9'))
+        return;
+      p.advance();
+      while ('0' <= p.peek() && p.peek() <= '9') {
+        p.advance();
+        if (p.peek() == '\'') {
+          p.advance();
+          if (!('0' <= p.peek() && p.peek() <= '9'))
+            return;
+        }
+      }
+    }
+  }
+  if (integer) {
+    if (p.peek() == 'u' || p.peek() == 'U') {
+      p.advance();
+      if (p.peek() == 'w') {
+        p.advance();
+        if (p.peek() != 'b')
+          return;
+        p.advance();
+      }
+      if (p.peek() == 'W') {
+        p.advance();
+        if (p.peek() != 'B')
+          return;
+        p.advance();
+      }
+      if (p.peek() == 'l' || p.peek() == 'L')
+        p.advance();
+      if (p.peek() == 'l' || p.peek() == 'L')
+        p.advance();
+    } else if (p.peek() == 'l' || p.peek() == 'L') {
+      p.advance();
+      if (p.peek() == 'l' || p.peek() == 'L')
+        p.advance();
+      if (p.peek() == 'u' || p.peek() == 'U')
+        p.advance();
+    } else if (p.peek_str(2) == "wb" || p.peek_str(2) == "WB") {
+      p.advance(2);
+    } else if (p.peek() == 'f' || p.peek() == 'F' || p.peek() == 'l' || p.peek() == 'L') {
+      p.advance();
+    } else if ((p.peek() == 'd' || p.peek() == 'D')
+               && (p.peek(1) == 'f' || p.peek(1) == 'F' || p.peek(1) == 'd' || p.peek(1) == 'D' || p.peek(1) == 'l' || p.peek(1) == 'L')) {
+      p.advance(2);
+    }
+  } else {
+    if (p.peek() == 'f' || p.peek() == 'F' || p.peek() == 'l' || p.peek() == 'L')
+      p.advance();
+    else if ((p.peek() == 'd' || p.peek() == 'D')
+             && (p.peek(1) == 'f' || p.peek(1) == 'F' || p.peek(1) == 'd' || p.peek(1) == 'D' || p.peek(1) == 'l' || p.peek(1) == 'L'))
+      p.advance(2);
+  }
+  if (identifier_char(p.peek()) || p.peek() == '.')
+    return;
+  tokens->push_back({start, p.i, io::Token::Number});
+}
+
+void handle_syntax(CParser &p, std::vector<io::Token> *tokens, std::vector<ParseEvent> *events) {
+  if (p.peek() == '/' && p.peek(1) == '/') {
+    p.current().state = CState::CInternalState::INLINE_COMMENT;
+    return;
+  }
+  if (p.peek() == '/' && p.peek(1) == '*') {
+    p.current().state = CState::CInternalState::MULTI_COMMENT;
+    events->push_back(false);
+    return;
+  }
+  switch (p.peek()) {
+  case '"':
+    p.current().state = CState::CInternalState::STRING;
+    p.advance();
+    tokens->push_back({p.i - 1, p.i, io::Token::String});
+    return;
+  case '\'':
+    p.current().state = CState::CInternalState::CHAR;
+    p.advance();
+    tokens->push_back({p.i - 1, p.i, io::Token::Char});
+    return;
+  case 'u': {
+    uint32_t j = 1;
+    if (p.peek(j) == '8')
+      j++;
+    if (p.peek(j) == '"') {
+      p.current().state = CState::CInternalState::STRING;
+      tokens->push_back({p.i, p.i + j, io::Token::Annotation});
+      p.advance(j + 1);
+      tokens->push_back({p.i - 1, p.i, io::Token::String});
+      return;
+    } else if (p.peek(j) == '\'') {
+      p.current().state = CState::CInternalState::CHAR;
+      tokens->push_back({p.i, p.i + j, io::Token::Annotation});
+      p.advance(j + 1);
+      tokens->push_back({p.i - 1, p.i, io::Token::Char});
+      return;
+    }
+    goto def;
+  }
+  case 'U':
+  case 'L': {
+    if (p.peek(1) == '"') {
+      p.current().state = CState::CInternalState::STRING;
+      tokens->push_back({p.i, p.i + 1, io::Token::Annotation});
+      p.advance(2);
+      tokens->push_back({p.i - 1, p.i, io::Token::String});
+      return;
+    } else if (p.peek(1) == '\'') {
+      p.current().state = CState::CInternalState::CHAR;
+      tokens->push_back({p.i, p.i + 1, io::Token::Annotation});
+      p.advance(2);
+      tokens->push_back({p.i - 1, p.i, io::Token::Char});
+      return;
+    }
+    goto def;
+  }
+  default:
+  def:
+    if (('0' <= p.peek() && p.peek() <= '9') || p.peek() == '.') {
+      handle_number(p, tokens);
+      return;
+    }
+    p.advance();
+  }
+}
+
+void c_parse(
+  void **v_state,
+  std::string_view line,
+  bool,
+  std::vector<io::Token> *tokens,
+  std::vector<ParseEvent> *events
+) {
+  CParser p(v_state, line);
+  while (p.i < p.len()) {
+    if (p.current().state == CState::CInternalState::MULTI_COMMENT) {
+      handle_multiline_comment(p, tokens, events);
+      continue;
+    }
+    if (p.current().state == CState::CInternalState::INLINE_COMMENT) {
+      handle_inline_comment(p, tokens, events);
+      return;
+    }
+    if (p.current().state == CState::CInternalState::STRING) {
+      handle_string(p, tokens, true);
+      continue;
+    }
+    if (p.current().state == CState::CInternalState::CHAR) {
+      handle_string(p, tokens, false);
+      continue;
+    }
+    handle_syntax(p, tokens, events);
+    continue;
+  }
+}
+} // namespace bed::internal::syntax::c

src/internal/syntax/ruby/parse.cc

@@ -35,73 +35,84 @@ inline uint8_t utf8_codepoint_width(unsigned char c) {
 }
 
 bool handle_escapes(RubyParser &p, std::vector<io::Token> *tokens, uint32_t &start, bool string = true) {
-  // TODO: properly handle escapes for %q ' etc.
+  if (p.peek() != '\\')
-  if (p.peek() == '\\') {
+    return false;
-    if (string)
+  if (!(p.current().flags & RubyState::RubyInternalState::ALLOW_ESCAPE)) {
-      tokens->push_back({start, p.i, io::Token::String});
-    else
-      tokens->push_back({start, p.i, io::Token::Regexp});
-    start = p.i;
     p.advance();
-    if (p.peek() == 'x') {
+    if (p.peek() != '\'' && p.peek() != '\\')
-      p.advance();
+      return false;
-      if (is_hex(p.peek()))
+    if (string)
-        p.advance();
+      tokens->push_back({start, p.i - 1, io::Token::String});
-      if (is_hex(p.peek()))
+    else
-        p.advance();
+      tokens->push_back({start, p.i - 1, io::Token::Regexp});
-    } else if (p.peek() == 'u') {
+    p.advance();
-      p.advance();
+    tokens->push_back({p.i - 2, p.i, io::Token::Escape});
-      if (p.peek() == '{') {
-        p.advance();
-        while (p.peek() != '}' && p.peek() != '\0')
-          p.advance();
-        if (p.peek() == '}')
-          p.advance();
-      } else {
-        if (is_hex(p.peek()))
-          p.advance();
-        if (is_hex(p.peek()))
-          p.advance();
-        if (is_hex(p.peek()))
-          p.advance();
-        if (is_hex(p.peek()))
-          p.advance();
-      }
-    } else if ('0' <= p.peek() && p.peek() <= '7') {
-      p.advance();
-      if ('0' <= p.peek() && p.peek() <= '7')
-        p.advance();
-      if ('0' <= p.peek() && p.peek() <= '7')
-        p.advance();
-    } else if (p.peek() == 'c') {
-      p.advance();
-      if (p.peek() != '\\')
-        p.advance();
-    } else if (p.peek() == 'M' || p.peek() == 'C') {
-      p.advance();
-      if (p.peek() == '-') {
-        p.advance();
-        if (p.peek() != '\\')
-          p.advance();
-      }
-    } else if (p.peek() == 'N') {
-      p.advance();
-      if (p.peek() == '{') {
-        p.advance();
-        while (p.peek() != '}' && p.peek() != '\0')
-          p.advance();
-        if (p.peek() == '}')
-          p.advance();
-      }
-    } else {
-      p.advance();
-    }
-    tokens->push_back({start, p.i, io::Token::Escape});
     start = p.i;
     return true;
   }
-  return false;
+  if (string)
-};
+    tokens->push_back({start, p.i, io::Token::String});
+  else
+    tokens->push_back({start, p.i, io::Token::Regexp});
+  start = p.i;
+  p.advance();
+  if (p.peek() == 'x') {
+    p.advance();
+    if (is_hex(p.peek()))
+      p.advance();
+    if (is_hex(p.peek()))
+      p.advance();
+  } else if (p.peek() == 'u') {
+    p.advance();
+    if (p.peek() == '{') {
+      p.advance();
+      while (p.peek() != '}' && p.peek() != '\0')
+        p.advance();
+      if (p.peek() == '}')
+        p.advance();
+    } else {
+      if (is_hex(p.peek()))
+        p.advance();
+      if (is_hex(p.peek()))
+        p.advance();
+      if (is_hex(p.peek()))
+        p.advance();
+      if (is_hex(p.peek()))
+        p.advance();
+    }
+  } else if ('0' <= p.peek() && p.peek() <= '7') {
+    p.advance();
+    if ('0' <= p.peek() && p.peek() <= '7')
+      p.advance();
+    if ('0' <= p.peek() && p.peek() <= '7')
+      p.advance();
+  } else if (p.peek() == 'c') {
+    p.advance();
+    if (p.peek() != '\\')
+      p.advance();
+  } else if (p.peek() == 'M' || p.peek() == 'C') {
+    p.advance();
+    if (p.peek() == '-') {
+      p.advance();
+      if (p.peek() != '\\')
+        p.advance();
+    }
+  } else if (p.peek() == 'N') {
+    p.advance();
+    if (p.peek() == '{') {
+      p.advance();
+      while (p.peek() != '}' && p.peek() != '\0')
+        p.advance();
+      if (p.peek() == '}')
+        p.advance();
+    }
+  } else {
+    p.advance();
+  }
+  tokens->push_back({start, p.i, io::Token::Escape});
+  start = p.i;
+  return true;
+}
 
 bool handle_heredoc(RubyParser &p, std::vector<io::Token> *tokens, std::vector<ParseEvent> *events) {
   uint8_t *heredocs = p.state->heredocs();
@@ -669,6 +680,7 @@ bool handle_syntax(RubyParser &p, std::vector<io::Token> *tokens, std::vector<Pa
     p.current().delim_start = '\'';
     p.current().delim_end = '\'';
     p.current().flags &= ~RubyState::RubyInternalState::ALLOW_INTERPOLATION;
+    p.current().flags &= ~RubyState::RubyInternalState::ALLOW_ESCAPE;
     p.advance();
     return false;
   }
@@ -679,6 +691,7 @@ bool handle_syntax(RubyParser &p, std::vector<io::Token> *tokens, std::vector<Pa
     p.current().delim_start = '"';
     p.current().delim_end = '"';
     p.current().flags |= RubyState::RubyInternalState::ALLOW_INTERPOLATION;
+    p.current().flags |= RubyState::RubyInternalState::ALLOW_ESCAPE;
     p.advance();
     return false;
   }
@@ -704,6 +717,7 @@ bool handle_syntax(RubyParser &p, std::vector<io::Token> *tokens, std::vector<Pa
     char delim_start = '\0';
     char delim_end = '\0';
     bool allow_interp = true;
+    bool allow_escape = true;
     int prefix_len = 1;
     bool is_regexp = false;
     switch (type) {
@@ -717,6 +731,7 @@ bool handle_syntax(RubyParser &p, std::vector<io::Token> *tokens, std::vector<Pa
     case 'I':
     case 'W':
       allow_interp = true;
+      allow_escape = true;
       prefix_len = 2;
       break;
     case 'w':
@@ -724,10 +739,12 @@ bool handle_syntax(RubyParser &p, std::vector<io::Token> *tokens, std::vector<Pa
     case 'i':
     case 's':
       allow_interp = false;
+      allow_escape = false;
       prefix_len = 2;
       break;
     default:
       allow_interp = true;
+      allow_escape = true;
       prefix_len = 1;
       break;
     }
@@ -769,6 +786,8 @@ bool handle_syntax(RubyParser &p, std::vector<io::Token> *tokens, std::vector<Pa
     p.current().delim_end = delim_end;
     if (allow_interp)
       p.current().flags |= RubyState::RubyInternalState::ALLOW_INTERPOLATION;
+    if (allow_escape)
+      p.current().flags |= RubyState::RubyInternalState::ALLOW_ESCAPE;
     p.current().lit_brace_level = 1;
     p.advance(prefix_len + 1);
     return false;
@@ -981,6 +1000,7 @@ bool handle_syntax(RubyParser &p, std::vector<io::Token> *tokens, std::vector<Pa
           p.current().flags &= ~RubyState::RubyInternalState::EXPECTING_EXPRESSION;
           tokens->push_back({p.i, p.i + j, io::Token::Function});
           p.advance(j);
+          return false;
         }
         uint32_t start = p.i;
         if (p.peek(j) == ':') {

src/internal/vase/vase.cc

@@ -179,7 +179,10 @@ Shard *insert(AppendStorage *ap, Shard *root, Shard *text, uint64_t line) {
   if (!root) {
     if (line != 0)
       throw ed_error("line out of range");
-    Shard::retain(text);
+    if (text)
+      Shard::retain(text);
+    else
+      text = newline(ap);
     return text;
   }
   if (line > root->lines + 1)