(System)Verilog: escaped identifiers (LRM 5.6.1)

Add support for escaped identifiers in Verilog and SystemVerilog:
`\` + zero or more non-whitespace characters + whitespace.

Note that the `\` itself isn't part of the identifier, and that `\foo`
is the same as just `foo` (unlike in VHDL), but that identifiers
identical to keywords such as `\begin` are also allowed.
This definition also theoretically allows an empty identifier, `\`,
which ctags currently ignores, issuing a notice.

Escaped identifiers are defined in the standard in section
3.7.1 (Verilog) / 5.6.1 (SystemVerilog).

Units/parser-verilog.r/verilog-escaped-id.d/args.ctags

@@ -0,0 +1 @@
+--sort=no

Units/parser-verilog.r/verilog-escaped-id.d/expected.tags

@@ -0,0 +1,11 @@
+themodule#(x=42)	input.v	/^module \\themodule#(x=42) ($/;"	m
+clk,rst,	input.v	/^    input wire \\clk,rst, , \\d ,$/;"	p	module:themodule#(x=42)
+d	input.v	/^    input wire \\clk,rst, , \\d ,$/;"	p	module:themodule#(x=42)
+1+1=2	input.v	/^    output reg \\1+1=2$/;"	p	module:themodule#(x=42)
+\\	input.v	/^localparam \\\\ = \\ ;$/;"	c	module:themodule#(x=42)
+\\\\	input.v	/^localparam \\\\\\ = \\\\ ;$/;"	c	module:themodule#(x=42)
+r\\n	input.v	/^localparam \\r\\n = \\\\\\ ;$/;"	c	module:themodule#(x=42)
+\\r\\n	input.v	/^localparam \\\\r\\n = \\r\\n ;$/;"	c	module:themodule#(x=42)
+\\\\r\\n	input.v	/^localparam \\\\\\r\\n = \\\\r\\n ;$/;"	c	module:themodule#(x=42)
+end	input.v	/^always @(posedge \\clk,rst, ) begin : \\end$/;"	b	module:themodule#(x=42)
+\\end	input.v	/^    if (\\\\\\r\\n ) begin : \\\\end$/;"	b	block:themodule#(x=42).end

Units/parser-verilog.r/verilog-escaped-id.d/input.v

@@ -0,0 +1,19 @@
+module \themodule#(x=42) (
+    input wire \clk,rst, , \d ,
+    output reg \1+1=2
+);
+
+localparam \ = 1;  // null identifier
+localparam \\ = \ ;
+localparam \\\ = \\ ;
+localparam \r\n = \\\ ;
+localparam \\r\n = \r\n ;
+localparam \\\r\n = \\r\n ;
+
+always @(posedge \clk,rst, ) begin : \end
+    if (\\\r\n ) begin : \\end
+        \1+1=2 <= d;
+    end
+end
+
+endmodule

parsers/verilog.c

@@ -705,10 +705,10 @@ static int vGetc (void)
 	return _vGetc (false);
 }
 
-// Is the first charactor in an identifier? [a-zA-Z_`]
+// Is the first character in an identifier? [a-zA-Z_`\\]
 static bool isWordToken (const int c)
 {
-	return (isalpha (c) || c == '_' || c == '`');
+	return (isalpha (c) || c == '_' || c == '`' || c == '\\');
 }
 
 // Is a charactor in an identifier? [a-zA-Z0-9_`$]
@@ -873,12 +873,26 @@ static int _readWordToken (tokenInfo *const token, int c, bool skip)
 	Assert (isWordToken (c));
 
 	clearToken (token);
-	do
+	if (c == '\\')
 	{
-		vStringPut (token->name, c);
-		c = vGetc ();
-	} while (isIdentifierCharacter (c));
-	_updateKind (token);
+		// Escaped identifier (may be empty)
+		c = vGetc (); // skip leading '\'
+		while (isgraph (c))
+		{
+			vStringPut (token->name, c);
+			c = vGetc ();
+		}
+		token->kind = K_IDENTIFIER;
+	}
+	else
+	{
+		do
+		{
+			vStringPut (token->name, c);
+			c = vGetc ();
+		} while (isIdentifierCharacter (c));
+		_updateKind (token);
+	}
 
 	if (skip)
 		return skipWhite (c);