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

@@ -33,7 +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/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

@@ -47,7 +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/parse.cc

@@ -106,6 +106,197 @@ void handle_string(CParser &p, std::vector<io::Token> *tokens, bool string) {
   }
 }
 
+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;
@@ -165,82 +356,8 @@ void handle_syntax(CParser &p, std::vector<io::Token> *tokens, std::vector<Parse
   }
   default:
   def:
-    if ('0' <= p.peek() && p.peek() <= '9') {
+    if (('0' <= p.peek() && p.peek() <= '9') || p.peek() == '.') {
-      // TODO: add floats/doubles
+      handle_number(p, tokens);
-      uint32_t start = p.i;
-      if (p.peek() == '0') {
-        p.advance();
-        if (p.peek() == 'x' || p.peek() == 'X') {
-          p.advance();
-          while (is_hex(p.peek())) {
-            p.advance();
-            if (p.peek() == '\'') {
-              p.advance();
-              if (!(is_hex(p.peek())))
-                return;
-            }
-          }
-        } else if (p.peek() == 'b' || p.peek() == 'B') {
-          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;
-            }
-          }
-        }
-      } 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() == '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 (identifier_char(p.peek())) {
-        return;
-      }
-      tokens->push_back({start, p.i, io::Token::Number});
       return;
     }
     p.advance();

src/internal/syntax/ruby/parse.cc

@@ -1000,6 +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");
+    if (text)
       Shard::retain(text);
+    else
+      text = newline(ap);
     return text;
   }
   if (line > root->lines + 1)