Disable EQUS expansion for raw symbols (by lexing them as strings)

man/rgbasm.5

@@ -1104,7 +1104,9 @@ Additionally, label names can contain up to a single dot
 .Ql \&. ,
 which may not be the first character.
 .Pp
-A symbol cannot have the same name as a reserved keyword, unless it is prefixed by a hash
+A symbol cannot have the same name as a reserved keyword, unless it is a
+.Dq raw symbol
+prefixed by a hash
 .Sq # .
 For example,
 .Ql #load
@@ -1387,6 +1389,8 @@ This expansion is disabled in a few contexts:
 and
 .Ql MACRO name
 will not expand string constants in their names.
+Expansion is also disabled for raw string constant symbols (string constant symbols prefixed by a hash
+.Sq # ) .
 .Bd -literal -offset indent
 DEF COUNTREG EQUS "[hl+]"
     ld a, COUNTREG

src/asm/lexer.cpp

@@ -1908,17 +1908,21 @@ static Token yylex_NORMAL() {
 				// `token` is either a `SYMBOL` or a `LOCAL`, and both have a `std::string` value.
 				assume(token.value.holds<std::string>());
 
-				// Local symbols cannot be string expansions
+				// Raw symbols and local symbols cannot be string expansions
 				if (token.type == T_(SYMBOL) && lexerState->expandStrings) {
 					// Attempt string expansion
 					Symbol const *sym = sym_FindExactSymbol(token.value.get<std::string>());
 
 					if (sym && sym->type == SYM_EQUS) {
-						std::shared_ptr<std::string> str = sym->getEqus();
-
-						assume(str);
-						beginExpansion(str, sym->name);
-						continue; // Restart, reading from the new buffer
+						if (raw) {
+							token.type = T_(STR_SYMBOL);
+						} else {
+							std::shared_ptr<std::string> str = sym->getEqus();
+
+							assume(str);
+							beginExpansion(str, sym->name);
+							continue; // Restart, reading from the new buffer
+						}
 					}
 				}
 
@@ -1934,7 +1938,7 @@ static Token yylex_NORMAL() {
 				// ':' character *immediately* follows the identifier. Thus, at the beginning of a
 				// line, "Label:" and "mac:" are treated as label definitions, but "Label :" and
 				// "mac :" are treated as macro invocations.
-				if (token.type == T_(SYMBOL) && peek() == ':') {
+				if ((token.type == T_(SYMBOL) || token.type == T_(STR_SYMBOL)) && peek() == ':') {
 					token.type = T_(LABEL);
 				}
 

src/asm/parser.y

@@ -314,6 +314,7 @@
 %token <int32_t> NUMBER "number"
 %token <std::string> STRING "string"
 %token <std::string> SYMBOL "symbol"
+%token <std::string> STR_SYMBOL "string symbol"
 %token <std::string> LABEL "label"
 %token <std::string> LOCAL "local label"
 %token <std::string> ANON "anonymous label"
@@ -354,10 +355,14 @@
 %type <std::string> def_rl
 %type <std::string> def_equs
 %type <std::string> redef_equs
+%type <std::string> symbol
 %type <std::string> scoped_sym
 // `scoped_sym_no_anon` exists because anonymous labels usually count as "scoped symbols", but some
 // contexts treat anonymous labels and other labels/symbols differently, e.g. `purge` or `export`.
 %type <std::string> scoped_sym_no_anon
+// `scoped_sym_no_str` exists because string symbols usually count as symbols, but some contexts
+// treat string symbols and other symbols differently, e.g. `relocexpr_no_str`.
+%type <std::string> scoped_sym_no_str
 
 // SM83 instruction parameters
 %type <int32_t> reg_r
@@ -514,7 +519,7 @@ endc:
 def_id:
 	OP_DEF {
 		lexer_ToggleStringExpansion(false);
-	} SYMBOL {
+	} symbol {
 		lexer_ToggleStringExpansion(true);
 		$$ = std::move($3);
 	}
@@ -523,13 +528,17 @@ def_id:
 redef_id:
 	POP_REDEF {
 		lexer_ToggleStringExpansion(false);
-	} SYMBOL {
+	} symbol {
 		lexer_ToggleStringExpansion(true);
 		$$ = std::move($3);
 	}
 ;
 
-scoped_sym_no_anon: SYMBOL | LABEL | LOCAL;
+symbol: SYMBOL | STR_SYMBOL;
+
+scoped_sym_no_str: SYMBOL | LABEL | LOCAL | ANON;
+
+scoped_sym_no_anon: symbol | LABEL | LOCAL;
 
 scoped_sym: scoped_sym_no_anon | ANON;
 
@@ -558,7 +567,7 @@ label:
 ;
 
 macro:
-	SYMBOL {
+	symbol {
 		// Parsing 'macro_args' will restore the lexer's normal mode
 		lexer_SetMode(LEXER_RAW);
 	} macro_args {
@@ -844,7 +853,7 @@ rept:
 for:
 	POP_FOR {
 		lexer_ToggleStringExpansion(false);
-	} SYMBOL {
+	} symbol {
 		lexer_ToggleStringExpansion(true);
 	} COMMA for_args NEWLINE capture_rept endofline {
 		if ($8.span.ptr) {
@@ -888,7 +897,7 @@ break:
 def_macro:
 	POP_MACRO {
 		lexer_ToggleStringExpansion(false);
-	} SYMBOL {
+	} symbol {
 		lexer_ToggleStringExpansion(true);
 	} NEWLINE capture_macro endofline {
 		if ($6.span.ptr) {
@@ -1153,16 +1162,16 @@ charmap_args:
 ;
 
 newcharmap:
-	POP_NEWCHARMAP SYMBOL {
+	POP_NEWCHARMAP symbol {
 		charmap_New($2, nullptr);
 	}
-	| POP_NEWCHARMAP SYMBOL COMMA SYMBOL {
+	| POP_NEWCHARMAP symbol COMMA symbol {
 		charmap_New($2, &$4);
 	}
 ;
 
 setcharmap:
-	POP_SETCHARMAP SYMBOL {
+	POP_SETCHARMAP symbol {
 		charmap_Set($2);
 	}
 ;
@@ -1174,7 +1183,7 @@ pushc:
 ;
 
 pushc_setcharmap:
-	POP_PUSHC SYMBOL {
+	POP_PUSHC symbol {
 		charmap_Push();
 		charmap_Set($2);
 	}
@@ -1307,7 +1316,7 @@ relocexpr:
 ;
 
 relocexpr_no_str:
-	scoped_sym {
+	scoped_sym_no_str {
 		$$.makeSymbol($1);
 	}
 	| NUMBER {
@@ -1400,7 +1409,7 @@ relocexpr_no_str:
 	| OP_ISCONST LPAREN relocexpr RPAREN {
 		$$.makeNumber($3.isKnown());
 	}
-	| OP_BANK LPAREN scoped_sym RPAREN {
+	| OP_BANK LPAREN scoped_sym_no_str RPAREN {
 		// '@' is also a SYMBOL; it is handled here
 		$$.makeBankSymbol($3);
 	}
@@ -1529,6 +1538,13 @@ string:
 	STRING {
 		$$ = std::move($1);
 	}
+	| STR_SYMBOL {
+		if (Symbol *sym = sym_FindExactSymbol($1); sym && sym->type == SYM_EQUS) {
+			$$ = *sym->getEqus();
+		} else {
+			::error("'%s' is not a string symbol", $1.c_str());
+		}
+	}
 	| OP_STRSUB LPAREN string COMMA iconst COMMA uconst RPAREN {
 		size_t len = strlenUTF8($3, false);
 		uint32_t pos = adjustNegativePos($5, len, "STRSUB");

test/asm/anon-label-bad.err

@@ -5,7 +5,7 @@ error: anon-label-bad.asm(6):
 error: anon-label-bad.asm(9):
     syntax error, unexpected anonymous label
 error: anon-label-bad.asm(10):
-    syntax error, unexpected anonymous label, expecting symbol or label or local label
+    syntax error, unexpected anonymous label, expecting symbol or string symbol or label or local label
 error: anon-label-bad.asm(22):
     syntax error, unexpected ::
 error: Assembly aborted (5 errors)!

test/asm/def-scoped.err

@@ -1,7 +1,7 @@
 error: def-scoped.asm(10):
-    syntax error, unexpected local label, expecting symbol
+    syntax error, unexpected local label, expecting symbol or string symbol
 error: def-scoped.asm(13):
-    syntax error, unexpected local label, expecting symbol
+    syntax error, unexpected local label, expecting symbol or string symbol
 error: def-scoped.asm(16):
-    syntax error, unexpected local label, expecting symbol
+    syntax error, unexpected local label, expecting symbol or string symbol
 error: Assembly aborted (3 errors)!

test/asm/label-macro-arg.err

@@ -1,15 +1,15 @@
 error: label-macro-arg.asm(38) -> label-macro-arg.asm::test_char(25):
-    syntax error, unexpected local label, expecting symbol
+    syntax error, unexpected local label, expecting symbol or string symbol
 while expanding symbol "VAR_DEF"
 error: label-macro-arg.asm(38) -> label-macro-arg.asm::test_char(26):
-    syntax error, unexpected local label, expecting symbol
+    syntax error, unexpected local label, expecting symbol or string symbol
 error: label-macro-arg.asm(38) -> label-macro-arg.asm::test_char(29):
     Interpolated symbol "sizeof_.something" does not exist
 error: label-macro-arg.asm(39) -> label-macro-arg.asm::test_char(25):
-    syntax error, unexpected label, expecting symbol
+    syntax error, unexpected label, expecting symbol or string symbol
 while expanding symbol "VAR_DEF"
 error: label-macro-arg.asm(39) -> label-macro-arg.asm::test_char(26):
-    syntax error, unexpected label, expecting symbol
+    syntax error, unexpected label, expecting symbol or string symbol
 error: label-macro-arg.asm(39) -> label-macro-arg.asm::test_char(29):
     Invalid format spec 'sizeof_'
 error: label-macro-arg.asm(39) -> label-macro-arg.asm::test_char(29):

test/asm/raw-string-symbol-errors.asm

@@ -0,0 +1,29 @@
+def n equ 42
+def s equs "hello"
+macro m
+endm
+
+assert (#n) == 42
+assert (#s) == $656c6c6f
+assert (#m) == 0
+assert (#u) == 0
+
+assert strlen(#n) == 0
+assert strlen(#s) == 5
+assert strlen(#m) == 0
+assert strlen(#u) == 0
+
+def d_n = (#n)
+def d_s = (#s)
+def d_m = (#m)
+def d_u = (#u)
+
+def s_n equs #n
+def s_s equs #s
+def s_m equs #m
+def s_u equs #u
+
+purge #s
+purge #s
+assert (#s) == 0
+assert strlen(#s) == 0

test/asm/raw-string-symbol-errors.err

@@ -0,0 +1,27 @@
+warning: raw-string-symbol-errors.asm(7): [-Wobsolete]
+    Treating multi-unit strings as numbers is deprecated
+error: raw-string-symbol-errors.asm(8):
+    'm' is not a numeric symbol
+error: raw-string-symbol-errors.asm(11):
+    syntax error, unexpected symbol
+error: raw-string-symbol-errors.asm(13):
+    syntax error, unexpected symbol
+error: raw-string-symbol-errors.asm(14):
+    syntax error, unexpected symbol
+warning: raw-string-symbol-errors.asm(17): [-Wobsolete]
+    Treating multi-unit strings as numbers is deprecated
+error: raw-string-symbol-errors.asm(18):
+    'm' is not a numeric symbol
+error: raw-string-symbol-errors.asm(19):
+    Expected constant expression: 'u' is not constant at assembly time
+error: raw-string-symbol-errors.asm(21):
+    syntax error, unexpected symbol
+error: raw-string-symbol-errors.asm(23):
+    syntax error, unexpected symbol
+error: raw-string-symbol-errors.asm(24):
+    syntax error, unexpected symbol
+error: raw-string-symbol-errors.asm(27):
+    's' was already purged
+error: raw-string-symbol-errors.asm(29):
+    syntax error, unexpected symbol
+error: Assembly aborted (11 errors)!

test/asm/raw-string-symbols.asm

@@ -0,0 +1,34 @@
+opt Wno-unmapped-char
+
+def hello equs "world"
+def name equs "hello"
+println "{name}"
+println #name
+assert !strcmp(strsub(#name, 1, 4), "hell")
+assert strlen(#hello) == charlen(#hello)
+assert strlen("{hello}") == 5
+
+def multi equs """the	quick
+brown	fox"""
+println #multi
+
+def char equs "A"
+def n = #char
+println n
+def n = (#char)
+println n
+def n = 1 + #char
+println n
+assert #char == $41
+
+def fmt equs "%s %s %d"
+println strfmt(#fmt, #name, #hello, (#char))
+
+purge #name
+assert !def(name) && !def(#name) && def(hello)
+
+section "test", rom0
+#label:
+db #hello
+dw #hello
+dw BANK(#label), #label

test/asm/raw-string-symbols.out

@@ -0,0 +1,8 @@
+hello
+hello
+the	quick
+brown	fox
+$41
+$41
+$42
+hello world 65

test/asm/syntax-error-lexer-mode.err

@@ -1,5 +1,5 @@
 error: syntax-error-lexer-mode.asm(7):
     syntax error, unexpected af
 error: syntax-error-lexer-mode.asm(11):
-    syntax error, unexpected af, expecting symbol
+    syntax error, unexpected af, expecting symbol or string symbol
 error: Assembly aborted (2 errors)!