Fix =>, yield, and throw with respect to newlines.

[cycript.git] / Cycript.l.in

diff --git a/Cycript.l.in b/Cycript.l.in
index 77a4d277d56be3c180210279d182f1abed51f2b2..8c183a42387318cb863ec7e2a56611ae6396884b 100644 (file)
--- a/Cycript.l.in
+++ b/Cycript.l.in
@@ -1,28 +1,35 @@
 /* Cycript - Optimizing JavaScript Compiler/Runtime
 /* Cycript - Optimizing JavaScript Compiler/Runtime

- * Copyright (C) 2009-2013  Jay Freeman (saurik)
+ * Copyright (C) 2009-2015  Jay Freeman (saurik)

 */
 
 */
 

-/* GNU General Public License, Version 3 {{{ */
+/* GNU Affero General Public License, Version 3 {{{ */

 /*
 /*

- * Cycript is free software: you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published
- * by the Free Software Foundation, either version 3 of the License,
- * or (at your option) any later version.
- *
- * Cycript is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of

  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with Cycript.  If not, see <http://www.gnu.org/licenses/>.
+ * GNU Affero General Public License for more details.
+
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.

 **/
 /* }}} */
 
 /* XXX: supposedly I will be screwed on very very long multi-line comments and need to replace these with a manual lexer. http://websrv.cs.fsu.edu/~engelen/courses/COP5621/Pr2.pdf */
 
 %{
 **/
 /* }}} */
 
 /* XXX: supposedly I will be screwed on very very long multi-line comments and need to replace these with a manual lexer. http://websrv.cs.fsu.edu/~engelen/courses/COP5621/Pr2.pdf */
 
 %{

-#define YYLTYPE cy::location
+
+#if defined(__clang__)
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Wunknown-pragmas"
+#pragma clang diagnostic ignored "-Wdeprecated-register"
+#endif
+
+#define YYLTYPE CYLocation

 #include "Cycript.tab.hh"
 typedef cy::parser::token tk;
 
 #include "Cycript.tab.hh"
 typedef cy::parser::token tk;
 


@@ -73,16 +80,16 @@ typedef cy::parser::token tk;
             left = yyleng - (nl - yytext) - 1; \
             nl = reinterpret_cast<const char *>(memchr(nl + 1, '\n', left)); \
         } while (nl != NULL); \
             left = yyleng - (nl - yytext) - 1; \
             nl = reinterpret_cast<const char *>(memchr(nl + 1, '\n', left)); \
         } while (nl != NULL); \

+        yylloc->step(); \

         yylloc->end.lines(lines); \
         yylloc->end.columns(left); \
         yylloc->end.lines(lines); \
         yylloc->end.columns(left); \

-        yylloc->step(); \

         more \
     } else L \
 }
 
 #define L { \
     yylloc->step(); \
         more \
     } else L \
 }
 
 #define L { \
     yylloc->step(); \

-    yylloc->columns(yyleng); \
+    yylloc->end.columns(yyleng); \

 }
 
 #define M { \
 }
 
 #define M { \


@@ -91,6 +98,14 @@ typedef cy::parser::token tk;
     } \
 }
 
     } \
 }
 

+#define E(message) { \
+    CYDriver::Error error; \
+    error.location_ = *yylloc; \
+    error.message_ = "syntax error, " message; \
+    yyextra->errors_.push_back(error); \
+    yyterminate(); \
+}
+

 int H(char c) {
     if (c >= '0' && c <= '9')
         return c - '0';
 int H(char c) {
     if (c >= '0' && c <= '9')
         return c - '0';


@@ -101,6 +116,46 @@ int H(char c) {
     return -1;
 }
 
     return -1;
 }
 

+static void U(char *&local, unsigned point) {
+    if (false) {
+    } else if (point < 0x000080) {
+        *local++ = point;
+    } else if (point < 0x000800) {
+        *local++ = 0xc0 | point >> 0x06 & 0x1f;
+        goto one;
+    } else if (point < 0x010000) {
+        *local++ = 0xe0 | point >> 0x0c & 0x0f;
+        goto two;
+    } else if (point < 0x110000) {
+        *local++ = 0xf0 | point >> 0x12 & 0x07;
+        *local++ = 0x80 | point >> 0x0c & 0x3f;
+      two:
+        *local++ = 0x80 | point >> 0x06 & 0x3f;
+      one:
+        *local++ = 0x80 | point >> 0x00 & 0x3f;
+    } else _assert(false);
+}
+
+static void U(char *&local, const char *text, yy_size_t &i) {
+    unsigned point;
+
+    char next(text[++i]);
+    if (next != '{') {
+        point = H(text[i + 0]) << 12 | H(text[i + 1]) << 8 | H(text[i + 2]) << 4 | H(text[i + 3]);
+        i += 3;
+    } else {
+        point = 0;
+        for (;;) {
+            next = text[++i];
+            if (next == '}')
+                break;
+            point = (point << 4) | H(next);
+        }
+    }
+
+    U(local, point);
+}
+

 #define YY_INPUT(data, value, size) { \
     if (yyextra->data_.eof()) \
         value = YY_NULL; \
 #define YY_INPUT(data, value, size) { \
     if (yyextra->data_.eof()) \
         value = YY_NULL; \


@@ -116,24 +171,60 @@ int H(char c) {
 %option prefix="cy"
 %option bison-bridge
 %option bison-locations
 %option prefix="cy"
 %option bison-bridge
 %option bison-locations

+%option nodefault

 %option noyywrap
 %option noyywrap

-%option yylineno
+%option noyylineno

 %option nounput
 %option nounput

-%option interactive
+%option nounistd
+%option 8bit
+%option backup
+%option batch
+%option never-interactive
+%option pointer

 %option reentrant
 %option stack
 
 %option reentrant
 %option stack
 

-Exponent [eE][+-]?[0-9]+
-Escape   \\[\\'"bfnrtv]|\\0|\\x[0-9a-fA-F]{2}|\\u[0-9a-fA-F]{4}|\\\n
+%option full
+%option ecs
+
+U1 [\x00-\x7f]
+U0 [\x80-\xbf]
+U2 [\xc2-\xdf]
+U3 [\xe0-\xef]
+U4 [\xf0-\xf4]
+
+HexDigit [0-9a-fA-F]
+LineTerminatorSequence \r?\n|\r|\xe2\x80[\xa8\xa9]
+WhiteSpace [\x09\x0b\x0c\x20]|\xc2\xa0|\xef\xbb\xbf
+UnicodeEscape \\u({HexDigit}{4}|\{{HexDigit}+\})
+
+OctalEscape \\[1-7]|\\[4-7][0-7]|\\[0-3][0-7][0-7]?
+StringEscape \\['"\\bfnrtv]|\\0|{OctalEscape}|\\x{HexDigit}{2}|{UnicodeEscape}
+StringExtra {StringEscape}|\\{LineTerminatorSequence}
+SingleString ([^'\\\n]|{StringExtra})*
+DoubleString ([^"\\\n]|{StringExtra})*
+StringPrefix '{SingleString}|\"{DoubleString}
+
+@include UnicodeIDStart.l
+@include UnicodeIDContinue.l
+
+IdentifierMore [$_]

 
 

-IdentifierStart [a-zA-Z$_]
-IdentifierPart  [a-zA-Z$_0-9]
+UnicodeStart {IdentifierMore}|{UnicodeIDStart}
+UnicodePart {IdentifierMore}|\xe2\x80[\x8c\x8d]|{UnicodeIDContinue}
+UnicodeFail {U2}|{U3}|{U3}{U0}|{U4}|{U4}{U0}|{U4}{U0}{U0}
+UnicodeScrap {UnicodePart}*{UnicodeFail}?
+
+IdentifierStart {UnicodeStart}|{UnicodeEscape}
+IdentifierPart {UnicodePart}|{UnicodeEscape}
+IdentifierFail {UnicodeFail}|\\(u({HexDigit}{0,3}|\{{HexDigit}*))?
+IdentifierScrap {IdentifierPart}*{IdentifierFail}?

 
 NonTerminator [^\n]
 BackslashSequence \\{NonTerminator}
 RegularExpressionFirstChar [^\n*\\/]|{BackslashSequence}
 RegularExpressionChar [^\n\\/]|{BackslashSequence}
 
 NonTerminator [^\n]
 BackslashSequence \\{NonTerminator}
 RegularExpressionFirstChar [^\n*\\/]|{BackslashSequence}
 RegularExpressionChar [^\n\\/]|{BackslashSequence}

-RegularExpressionFlags {IdentifierPart}*
+RegularExpressionFlags {UnicodePart}*

 RegularExpressionChars {RegularExpressionChar}*
 RegularExpressionBody {RegularExpressionFirstChar}{RegularExpressionChars}
 
 RegularExpressionChars {RegularExpressionChar}*
 RegularExpressionBody {RegularExpressionFirstChar}{RegularExpressionChars}
 


@@ -154,6 +245,10 @@ XMLName {XMLNameStart}{XMLNamePart}*
 %%
 
 <RegExp>\/{RegularExpressionBody}\/{RegularExpressionFlags} L C I(literal, RegEx(Y), tk::RegularExpressionLiteral, hi::Constant);
 %%
 
 <RegExp>\/{RegularExpressionBody}\/{RegularExpressionFlags} L C I(literal, RegEx(Y), tk::RegularExpressionLiteral, hi::Constant);

+<RegExp>\/{RegularExpressionBody}\/{RegularExpressionFlags}{UnicodeFail} L E("invalid flags")
+<RegExp>\/{RegularExpressionBody}?\\? L E("unterminated regex")
+
+#![^\n]* L M

 
 \/\/[^\n]* L M
 
 
 \/\/[^\n]* L M
 


@@ -161,6 +256,7 @@ XMLName {XMLNameStart}{XMLNamePart}*
        /* XXX: unify these two rules using !? */
 \/\*!([^*]|[\r\n]|(\*+([^*/]|[\r\n])))*\*+\/ V() C I(comment, Comment(Y), tk::Comment, hi::Comment);
 \/\*([^*]|[\r\n]|(\*+([^*/]|[\r\n])))*\*+\/ V(N) M
        /* XXX: unify these two rules using !? */
 \/\*!([^*]|[\r\n]|(\*+([^*/]|[\r\n])))*\*+\/ V() C I(comment, Comment(Y), tk::Comment, hi::Comment);
 \/\*([^*]|[\r\n]|(\*+([^*/]|[\r\n])))*\*+\/ V(N) M

+\/\*([^*]|[\r\n]|(\*+([^*/]|[\r\n])))*\** V() E("invalid comment")

 
 @begin E4X
 <RegExp>"<>"      L F(tk::LeftRight, hi::Structure);
 
 @begin E4X
 <RegExp>"<>"      L F(tk::LeftRight, hi::Structure);


@@ -185,6 +281,7 @@ XMLName {XMLNameStart}{XMLNamePart}*
 @end
 
 "..."  L C F(tk::PeriodPeriodPeriod, hi::Meta);
 @end
 
 "..."  L C F(tk::PeriodPeriodPeriod, hi::Meta);

+".."   L E("invalid operator")

 
 @begin E4X
 "::"   L C F(tk::ColonColon, hi::Operator);
 
 @begin E4X
 "::"   L C F(tk::ColonColon, hi::Operator);


@@ -193,6 +290,7 @@ XMLName {XMLNameStart}{XMLNamePart}*
 
 @begin E4X ObjectiveC
 "@"    L C F(tk::At, hi::Operator);
 
 @begin E4X ObjectiveC
 "@"    L C F(tk::At, hi::Operator);

+"#"    L C F(tk::Pound, hi::Operator);

 @end
 
 "&"    L C F(tk::Ampersand, hi::Operator);
 @end
 
 "&"    L C F(tk::Ampersand, hi::Operator);


@@ -203,7 +301,7 @@ XMLName {XMLNameStart}{XMLNamePart}*
 "="    L C F(tk::Equal, hi::Operator);
 "=="   L C F(tk::EqualEqual, hi::Operator);
 "==="  L C F(tk::EqualEqualEqual, hi::Operator);
 "="    L C F(tk::Equal, hi::Operator);
 "=="   L C F(tk::EqualEqual, hi::Operator);
 "==="  L C F(tk::EqualEqualEqual, hi::Operator);

-"=>"   L C F(tk::EqualRight, hi::Operator);
+"=>"   L C F(yylval->newline_ ? tk::EqualRight_ : tk::EqualRight, hi::Operator);

 "!"    L C F(tk::Exclamation, hi::Operator);
 "!="   L C F(tk::ExclamationEqual, hi::Operator);
 "!=="  L C F(tk::ExclamationEqualEqual, hi::Operator);
 "!"    L C F(tk::Exclamation, hi::Operator);
 "!="   L C F(tk::ExclamationEqual, hi::Operator);
 "!=="  L C F(tk::ExclamationEqualEqual, hi::Operator);


@@ -259,15 +357,27 @@ XMLName {XMLNameStart}{XMLNamePart}*
 
 @begin C
 "typedef"         L C I(identifier, Identifier("typedef"), tk::Typedef, hi::Meta);
 
 @begin C
 "typedef"         L C I(identifier, Identifier("typedef"), tk::Typedef, hi::Meta);

+"unsigned"        L C I(identifier, Identifier("unsigned"), tk::Unsigned, hi::Type);
+"signed"          L C I(identifier, Identifier("signed"), tk::Signed, hi::Type);
+"extern"          L C I(identifier, Identifier("extern"), tk::Extern, hi::Type);

 @end
 
 @end
 

-@begin ObjectiveC
+@begin C

 "@encode"         L C F(tk::AtEncode, hi::Meta);
 "@encode"         L C F(tk::AtEncode, hi::Meta);

+@end
+
+@begin ObjectiveC

 "@end"            L C F(tk::AtEnd, hi::Meta);
 "@implementation" L C F(yyextra->no_.AtImplementation ? tk::AtImplementation_ : tk::AtImplementation, hi::Meta);
 "@end"            L C F(tk::AtEnd, hi::Meta);
 "@implementation" L C F(yyextra->no_.AtImplementation ? tk::AtImplementation_ : tk::AtImplementation, hi::Meta);

-"@import"         L C F(tk::AtImport, hi::Meta);
+"@import"         L C F(tk::AtImport, hi::Special);

 "@selector"       L C F(tk::AtSelector, hi::Meta);
 
 "@selector"       L C F(tk::AtSelector, hi::Meta);
 

+"@null"           L C F(tk::AtNull, hi::Constant);
+"@YES"            L C F(tk::AtYes, hi::Constant);
+"@NO"             L C F(tk::AtNo, hi::Constant);
+"@true"           L C F(tk::AtTrue, hi::Constant);
+"@false"          L C F(tk::AtFalse, hi::Constant);
+

 "NULL"         L C I(identifier, Identifier("NULL"), tk::Identifier_, hi::Constant);
 "nil"          L C I(identifier, Identifier("nil"), tk::Identifier_, hi::Constant);
 "YES"          L C I(identifier, Identifier("YES"), tk::Yes, hi::Constant);
 "NULL"         L C I(identifier, Identifier("NULL"), tk::Identifier_, hi::Constant);
 "nil"          L C I(identifier, Identifier("nil"), tk::Identifier_, hi::Constant);
 "YES"          L C I(identifier, Identifier("YES"), tk::Yes, hi::Constant);


@@ -349,7 +459,7 @@ XMLName {XMLNameStart}{XMLNamePart}*
 "volatile"     L C I(identifier, Identifier("volatile"), tk::Volatile, hi::Meta);
 
 "let"          L C I(identifier, Identifier("let"), tk::Let, hi::Meta);
 "volatile"     L C I(identifier, Identifier("volatile"), tk::Volatile, hi::Meta);
 
 "let"          L C I(identifier, Identifier("let"), tk::Let, hi::Meta);

-"yield"        L C I(identifier, Identifier("yield"), tk::Yield, hi::Control);
+"yield"        L R I(identifier, Identifier("yield"), tk::Yield, hi::Control);

 
 "each"         L C I(identifier, Identifier("each"), tk::Each, hi::Control);
 "of"           L C I(identifier, Identifier("of"), tk::Of, hi::Operator);
 
 "each"         L C I(identifier, Identifier("each"), tk::Each, hi::Control);
 "of"           L C I(identifier, Identifier("of"), tk::Of, hi::Operator);


@@ -359,15 +469,38 @@ XMLName {XMLNameStart}{XMLNamePart}*
 "xml"          L C I(identifier, Identifier("xml"), tk::XML, hi::Meta);
 @end
 
 "xml"          L C I(identifier, Identifier("xml"), tk::XML, hi::Meta);
 @end
 

-{IdentifierStart}{IdentifierPart}* L C I(identifier, Identifier(Y), tk::Identifier_, hi::Identifier);
+{UnicodeStart}{UnicodePart}* L C I(identifier, Identifier(Y), tk::Identifier_, hi::Identifier);

 
 

-(\.[0-9]+|(0|[1-9][0-9]*)(\.[0-9]*)?){Exponent}? L C I(number, Number(strtod(yytext, NULL)), tk::NumericLiteral, hi::Constant);
+{IdentifierStart}{IdentifierPart}* L C {
+    char *value(A char[yyleng + 1]);
+    char *local(value);
+
+    for (yy_size_t i(0), e(yyleng); i != e; ++i) {
+        char next(yytext[i]);
+        if (next != '\\')
+            *local++ = next;
+        else
+            U(local, yytext, ++i);
+    }
+
+    *local = '\0';
+    I(identifier, Identifier(value), tk::Identifier_, hi::Identifier);
+}
+
+({IdentifierStart}{IdentifierPart}*)?{IdentifierFail} L E("invalid identifier")

 
 

-0[xX][0-9a-fA-F]+ L C I(number, Number(strtoull(yytext + 2, NULL, 16)), tk::NumericLiteral, hi::Constant);

 0[0-7]+ L C I(number, Number(strtoull(yytext + 1, NULL, 8)), tk::NumericLiteral, hi::Constant);
 0[0-7]+ L C I(number, Number(strtoull(yytext + 1, NULL, 8)), tk::NumericLiteral, hi::Constant);

+0[0-9]+ L C I(number, Number(strtoull(yytext + 1, NULL, 10)), tk::NumericLiteral, hi::Constant);
+
+0[xX][0-9a-fA-F]+ L C I(number, Number(strtoull(yytext + 2, NULL, 16)), tk::NumericLiteral, hi::Constant);
+0[oO][0-7]+ L C I(number, Number(strtoull(yytext + 2, NULL, 8)), tk::NumericLiteral, hi::Constant);

 0[bB][0-1]+ L C I(number, Number(strtoull(yytext + 2, NULL, 2)), tk::NumericLiteral, hi::Constant);
 
 0[bB][0-1]+ L C I(number, Number(strtoull(yytext + 2, NULL, 2)), tk::NumericLiteral, hi::Constant);
 

-\"([^"\\\n]|{Escape})*\"|'([^'\\\n]|{Escape})*' L C {
+(\.[0-9]+|(0|[1-9][0-9]*)(\.[0-9]*)?)([eE][+-]?[0-9]+)? L C I(number, Number(strtod(yytext, NULL)), tk::NumericLiteral, hi::Constant);
+(\.[0-9]+|(0|[1-9][0-9]*)(\.[0-9]*)?)[eE][+-]?{IdentifierScrap} L E("invalid exponent")
+(\.?[0-9]|(0|[1-9][0-9]*)\.){IdentifierScrap} L E("invalid number")
+
+'{SingleString}'|\"{DoubleString}\" L C {

     char *value(A char[yyleng]);
     char *local(value);
 
     char *value(A char[yyleng]);
     char *local(value);
 


@@ -375,8 +508,13 @@ XMLName {XMLNameStart}{XMLNamePart}*
         char next(yytext[i]);
 
         if (yytext[i] == '\\')
         char next(yytext[i]);
 
         if (yytext[i] == '\\')

-            switch (next = yytext[++i]) {
-                case '\n': continue;
+            // XXX: support more line continuation characters
+            if (false) line: {
+                yylloc->end.lines(1);
+                yylloc->end.columns(yyleng - i);
+            } else switch (next = yytext[++i]) {
+                case '\n': goto line;
+

                 case '\\': next = '\\'; break;
                 case '\'': next = '\''; break;
                 case '"': next = '"'; break;
                 case '\\': next = '\\'; break;
                 case '\'': next = '\''; break;
                 case '"': next = '"'; break;


@@ -386,12 +524,31 @@ XMLName {XMLNameStart}{XMLNamePart}*
                 case 'r': next = '\r'; break;
                 case 't': next = '\t'; break;
                 case 'v': next = '\v'; break;
                 case 'r': next = '\r'; break;
                 case 't': next = '\t'; break;
                 case 'v': next = '\v'; break;

-                case '0': next = '\0'; break;
+
+                case '0': case '1': case '2': case '3':
+                    if (yytext[i + 1] < '0' || yytext[i + 1] > '7')
+                        next = H(yytext[i]), i += 0;
+                    else if (yytext[i + 2] < '0' || yytext[i + 2] > '7')
+                        next = H(yytext[i]) << 3 | H(yytext[i + 1]), i += 1;
+                    else
+                        next = H(yytext[i]) << 6 | H(yytext[i + 1]) << 3 | H(yytext[i + 2]), i += 2;
+                break;
+
+                case '4': case '5': case '6': case '7':
+                    if (yytext[i + 1] < '0' || yytext[i + 1] > '7')
+                        next = H(yytext[i]), i += 0;
+                    else
+                        next = H(yytext[i]) << 3 | H(yytext[i + 1]), i += 1;
+                break;

 
                 case 'x':
 
                 case 'x':

-                    next = H(yytext[i + 1]) << 4 | H(yytext[i + 2]);
+                    U(local, H(yytext[i + 1]) << 4 | H(yytext[i + 2]));

                     i += 2;
                     i += 2;

-                break;
+                continue;
+
+                case 'u':
+                    U(local, yytext, i);
+                continue;

             }
 
         *local++ = next;
             }
 
         *local++ = next;


@@ -401,19 +558,18 @@ XMLName {XMLNameStart}{XMLNamePart}*
     I(string, String(value, local - value), tk::StringLiteral, hi::Constant);
 }
 
     I(string, String(value, local - value), tk::StringLiteral, hi::Constant);
 }
 

-\r?\n|\r|\xe2\x80[\xa8\xa9] yylloc->end.lines(); yylloc->step(); N
+{StringPrefix}\\(x.{0,2}|u([^{].{0,3}|\{[^}]*)?|{UnicodeFail})? L E("invalid escape")
+{StringPrefix} L E("invalid string")
+
+{LineTerminatorSequence} yylloc->step(); yylloc->end.lines(); N

 
 

-[ \t] L
+{WhiteSpace} L

 
 <<EOF>> if (yyextra->auto_) { yyextra->auto_ = false; F(tk::AutoComplete, hi::Nothing); } L yyterminate();
 
 
 <<EOF>> if (yyextra->auto_) { yyextra->auto_ = false; F(tk::AutoComplete, hi::Nothing); } L yyterminate();
 

-. L {
-    CYDriver::Error error;
-    error.location_ = *yylloc;
-    error.message_ = "syntax error, unknown token";
-    yyextra->errors_.push_back(error);
-    yyterminate();
-}
+@({UnicodeStart}{UnicodeScrap}|{UnicodeFail}) L E("invalid keyword")
+
+. L E("invalid character")

 
 %%
 
 
 %%
 


@@ -482,3 +638,7 @@ void CYDriver::PushCondition(Condition condition) {
 void CYDriver::PopCondition() {
     yy_pop_state(scanner_);
 }
 void CYDriver::PopCondition() {
     yy_pop_state(scanner_);
 }

+
+#if defined(__clang__)
+#pragma clang diagnostic pop
+#endif