PR 5902: Parse errors
Eric Blake
ebb9@email.byu.edu
Mon Mar 11 20:24:00 GMT 2002
I committed these two patches, both mainline and 3.1, on approval from
Alexandre.
-------- Original Message --------
From: Eric Blake <ebb9@email.byu.edu>
Subject: Re: PR 5902: Parse errors
To: apbianco@redhat.com
CC: java-patches@gcc.gnu.org
Alexandre Petit-Bianco wrote:
>
> I wanted to give it a little more testing, but the patch appears to be
> malformed (first line 11, then 28... probably a paste that went
> wrong.) Could you please re-post something? Two different patches
> would be really convenient.
Here goes. I'm attaching the patches this time, instead of inlining
quoting them. Patch one fixed PR 5902, patch 2 is comment formatting.
2002-03-11 Eric Blake <ebb9@email.byu.edu>
* lex.c: Adjust comments to GNU standards.
2002-03-11 Eric Blake <ebb9@email.byu.edu>
Fix for PR java/5902:
* lex.c (java_lex): Fix parsing of literals.
--
This signature intentionally left boring.
Eric Blake ebb9@email.byu.edu
BYU student, free software programmer
-------------- next part --------------
Index: lex.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/java/lex.c,v
retrieving revision 1.81
diff -u -r1.81 lex.c
--- lex.c 2002/03/03 21:10:09 1.81
+++ lex.c 2002/03/11 19:32:39
@@ -1,5 +1,5 @@
/* Language lexer for the GNU compiler for the Java(TM) language.
- Copyright (C) 1997, 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
+ Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002 Free Software Foundation, Inc.
Contributed by Alexandre Petit-Bianco (apbianco@cygnus.com)
This file is part of GNU CC.
@@ -990,7 +990,7 @@
/* End borrowed section */
char literal_token [256];
int literal_index = 0, radix = 10, long_suffix = 0, overflow = 0, bytes;
- int found_hex_digits = 0;
+ int found_hex_digits = 0, found_non_octal_digits = 0;
int i;
#ifndef JC1_LITE
int number_beginning = ctxp->c_line->current;
@@ -1052,18 +1052,20 @@
}
/* Parse the first part of the literal, until we find something
which is not a number. */
- while ((radix == 10 && JAVA_ASCII_DIGIT (c)) ||
- (radix == 16 && JAVA_ASCII_HEXDIGIT (c)) ||
- (radix == 8 && JAVA_ASCII_OCTDIGIT (c)))
+ while ((radix == 16 && JAVA_ASCII_HEXDIGIT (c)) ||
+ JAVA_ASCII_DIGIT (c))
{
/* We store in a string (in case it turns out to be a FP) and in
PARTS if we have to process a integer literal. */
int numeric = hex_value (c);
int count;
- /* Remember when we find a valid hexadecimal digit */
+ /* Remember when we find a valid hexadecimal digit. */
if (radix == 16)
found_hex_digits = 1;
+ /* Remember when we find an invalid octal digit. */
+ else if (radix == 8 && !JAVA_ASCII_OCTDIGIT (c))
+ found_non_octal_digits = 1;
literal_token [literal_index++] = c;
/* This section of code if borrowed from gcc/c-lex.c */
@@ -1184,19 +1186,14 @@
}
} /* JAVA_ASCCI_FPCHAR (c) */
+ /* Here we get back to converting the integral literal. */
if (radix == 16 && ! found_hex_digits)
java_lex_error
("0x must be followed by at least one hexadecimal digit", 0);
-
- /* Here we get back to converting the integral literal. */
- if (c == 'L' || c == 'l')
+ else if (radix == 8 && found_non_octal_digits)
+ java_lex_error ("Octal literal contains digit out of range", 0);
+ else if (c == 'L' || c == 'l')
long_suffix = 1;
- else if (radix == 16 && JAVA_ASCII_LETTER (c))
- java_lex_error ("Digit out of range in hexadecimal literal", 0);
- else if (radix == 8 && JAVA_ASCII_DIGIT (c))
- java_lex_error ("Digit out of range in octal literal", 0);
- else if (radix == 16 && !literal_index)
- java_lex_error ("No digit specified for hexadecimal literal", 0);
else
java_unget_unicode ();
-------------- next part --------------
--- lex.c.~ Mon Mar 11 12:33:06 2002
+++ lex.c Mon Mar 11 12:49:58 2002
@@ -38,7 +38,7 @@
#include "flags.h"
#include "chartables.h"
-/* Function declaration */
+/* Function declarations. */
static char *java_sprint_unicode PARAMS ((struct java_line *, int));
static void java_unicode_2_utf8 PARAMS ((unicode_t));
static void java_lex_error PARAMS ((const char *, int));
@@ -191,7 +191,7 @@
free (ctxp->p_line);
}
ctxp->p_line = ctxp->c_line;
- ctxp->c_line = NULL; /* Reallocated */
+ ctxp->c_line = NULL; /* Reallocated. */
}
if (!ctxp->c_line)
@@ -606,7 +606,7 @@
static int
java_get_unicode ()
{
- /* It's time to read a line when... */
+ /* It's time to read a line when... */
if (!ctxp->c_line || ctxp->c_line->current == ctxp->c_line->size)
{
int c;
@@ -649,7 +649,7 @@
}
/* Parse the end of a C style comment.
- * C is the first character following the '/' and '*'. */
+ * C is the first character following the '/' and '*'. */
static void
java_parse_end_comment (c)
int c;
@@ -669,7 +669,7 @@
return;
case '/':
return;
- case '*': /* reparse only '*' */
+ case '*': /* Reparse only '*'. */
java_unget_unicode ();
}
}
@@ -705,9 +705,9 @@
java_lex_error ("Comment not terminated at end of input", 0);
if (seen_star && (c == '/'))
- return 1; /* Goto step1 in caller */
+ return 1; /* Goto step1 in caller. */
- /* We're parsing @deprecated */
+ /* We're parsing `@deprecated'. */
if (valid_tag && (c == '@'))
{
char tag [11];
@@ -910,7 +910,7 @@
/* Translation of the Unicode escape in the raw stream of Unicode
characters. Takes care of line terminator. */
step1:
- /* Skip white spaces: SP, TAB and FF or ULT */
+ /* Skip white spaces: SP, TAB and FF or ULT. */
for (c = java_get_unicode ();
c == '\n' || JAVA_WHITE_SPACE_P (c); c = java_get_unicode ())
if (c == '\n')
@@ -921,15 +921,16 @@
ctxp->elc.col = (ctxp->elc.col < 0 ? 0 : ctxp->elc.col);
- if (c == 0x1a) /* CTRL-Z */
+ if (c == 0x1a) /* CTRL-Z. */
{
if ((c = java_get_unicode ()) == UEOF)
- return 0; /* Ok here */
+ return 0; /* Ok here. */
else
- java_unget_unicode (); /* Caught later, at the end of the function */
+ java_unget_unicode (); /* Caught later, at the end of the
+ function. */
}
- /* Handle EOF here */
- if (c == UEOF) /* Should probably do something here... */
+ /* Handle EOF here. */
+ if (c == UEOF) /* Should probably do something here... */
return 0;
/* Take care of eventual comments. */
@@ -959,7 +960,7 @@
if ((c = java_get_unicode ()) == '*')
{
if ((c = java_get_unicode ()) == '/')
- goto step1; /* Empy documentation comment */
+ goto step1; /* Empty documentation comment. */
else if (java_parse_doc_section (c))
goto step1;
}
@@ -980,14 +981,14 @@
if (ctxp->elc.col < 0)
abort ();
- /* Numeric literals */
+ /* Numeric literals. */
if (JAVA_ASCII_DIGIT (c) || (c == '.'))
{
- /* This section of code is borrowed from gcc/c-lex.c */
+ /* This section of code is borrowed from gcc/c-lex.c. */
#define TOTAL_PARTS ((HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR) * 2 + 2)
int parts[TOTAL_PARTS];
HOST_WIDE_INT high, low;
- /* End borrowed section */
+ /* End borrowed section. */
char literal_token [256];
int literal_index = 0, radix = 10, long_suffix = 0, overflow = 0, bytes;
int found_hex_digits = 0, found_non_octal_digits = 0;
@@ -997,7 +998,7 @@
tree value;
#endif
- /* We might have a . separator instead of a FP like .[0-9]* */
+ /* We might have a . separator instead of a FP like .[0-9]*. */
if (c == '.')
{
unicode_t peep = java_sneak_unicode ();
@@ -1024,13 +1025,13 @@
radix = 8;
else if (c == '.')
{
- /* Push the '.' back and prepare for a FP parsing... */
+ /* Push the '.' back and prepare for a FP parsing... */
java_unget_unicode ();
c = '0';
}
else
{
- /* We have a zero literal: 0, 0{f,F}, 0{d,D} */
+ /* We have a zero literal: 0, 0{l,L}, 0{f,F}, 0{d,D}. */
JAVA_LEX_LIT ("0", 10);
switch (c)
{
@@ -1060,15 +1061,15 @@
int numeric = hex_value (c);
int count;
- /* Remember when we find a valid hexadecimal digit. */
+ /* Remember when we find a valid hexadecimal digit. */
if (radix == 16)
found_hex_digits = 1;
- /* Remember when we find an invalid octal digit. */
+ /* Remember when we find an invalid octal digit. */
else if (radix == 8 && !JAVA_ASCII_OCTDIGIT (c))
found_non_octal_digits = 1;
literal_token [literal_index++] = c;
- /* This section of code if borrowed from gcc/c-lex.c */
+ /* This section of code if borrowed from gcc/c-lex.c. */
for (count = 0; count < TOTAL_PARTS; count++)
{
parts[count] *= radix;
@@ -1094,7 +1095,7 @@
int seen_digit = (literal_index ? 1 : 0);
int seen_exponent = 0;
int fflag = 0; /* 1 for {f,F}, 0 for {d,D}. FP literal are
- double unless specified. */
+ double unless specified. */
/* It is ok if the radix is 8 because this just means we've
seen a leading `0'. However, radix==16 is invalid. */
@@ -1120,9 +1121,10 @@
{
if (stage < 2)
{
- /* {E,e} must have seen at list a digit */
+ /* {E,e} must have seen at least a digit. */
if (!seen_digit)
- java_lex_error ("Invalid FP literal", 0);
+ java_lex_error
+ ("Invalid FP literal, mantissa must have digit", 0);
seen_digit = 0;
seen_exponent = 1;
stage = 2;
@@ -1135,7 +1137,7 @@
if ( c == 'f' || c == 'F' || c == 'd' || c == 'D')
{
fflag = ((c == 'd') || (c == 'D')) ? 0 : 1;
- stage = 4; /* So we fall through */
+ stage = 4; /* So we fall through. */
}
if ((c=='-' || c =='+') && stage == 2)
@@ -1160,12 +1162,13 @@
#ifndef JC1_LITE
struct jpa_args a;
#endif
- if (stage != 4) /* Don't push back fF/dD */
+ if (stage != 4) /* Don't push back fF/dD. */
java_unget_unicode ();
/* An exponent (if any) must have seen a digit. */
if (seen_exponent && !seen_digit)
- java_lex_error ("Invalid FP literal", 0);
+ java_lex_error
+ ("Invalid FP literal, exponent must have digit", 0);
literal_token [literal_index] = '\0';
JAVA_LEX_LIT (literal_token, radix);
@@ -1184,7 +1187,7 @@
#endif
}
}
- } /* JAVA_ASCCI_FPCHAR (c) */
+ } /* JAVA_ASCII_FPCHAR (c) */
/* Here we get back to converting the integral literal. */
if (radix == 16 && ! found_hex_digits)
@@ -1198,10 +1201,10 @@
java_unget_unicode ();
#ifdef JAVA_LEX_DEBUG
- literal_token [literal_index] = '\0'; /* So JAVA_LEX_LIT is safe. */
+ literal_token [literal_index] = '\0'; /* So JAVA_LEX_LIT is safe. */
JAVA_LEX_LIT (literal_token, radix);
#endif
- /* This section of code is borrowed from gcc/c-lex.c */
+ /* This section of code is borrowed from gcc/c-lex.c. */
if (!overflow)
{
bytes = GET_TYPE_PRECISION (long_type_node);
@@ -1222,13 +1225,13 @@
}
/* End borrowed section. */
- /* Range checking */
+ /* Range checking. */
if (long_suffix)
{
/* 9223372036854775808L is valid if operand of a '-'. Otherwise
9223372036854775807L is the biggest `long' literal that can be
- expressed using a 10 radix. For other radixes, everything that
- fits withing 64 bits is OK. */
+ expressed using a 10 radix. For other radices, everything that
+ fits withing 64 bits is OK. */
int hb = (high >> 31);
if (overflow || (hb && low && radix == 10)
|| (hb && high & 0x7fffffff && radix == 10))
@@ -1238,9 +1241,9 @@
{
/* 2147483648 is valid if operand of a '-'. Otherwise,
2147483647 is the biggest `int' literal that can be
- expressed using a 10 radix. For other radixes, everything
+ expressed using a 10 radix. For other radices, everything
that fits within 32 bits is OK. As all literals are
- signed, we sign extend here. */
+ signed, we sign extend here. */
int hb = (low >> 31) & 0x1;
if (overflow || high || (hb && low & 0x7fffffff && radix == 10))
JAVA_INTEGRAL_RANGE_ERROR ("Numeric overflow for `int' literal");
@@ -1257,7 +1260,7 @@
return INT_LIT_TK;
}
- /* Character literals */
+ /* Character literals. */
if (c == '\'')
{
int char_lit;
@@ -1278,14 +1281,14 @@
java_lex_error ("Syntax error in character literal", 0);
if (char_lit == JAVA_CHAR_ERROR)
- char_lit = 0; /* We silently convert it to zero */
+ char_lit = 0; /* We silently convert it to zero. */
JAVA_LEX_CHAR_LIT (char_lit);
SET_LVAL_NODE_TYPE (build_int_2 (char_lit, 0), char_type_node);
return CHAR_LIT_TK;
}
- /* String literals */
+ /* String literals. */
if (c == '"')
{
int no_error;
@@ -1303,9 +1306,9 @@
}
java_unicode_2_utf8 (c);
}
- if (c == '\n' || c == UEOF) /* ULT */
+ if (c == '\n' || c == UEOF) /* ULT. */
{
- lineno--; /* Refer to the line the terminator was seen */
+ lineno--; /* Refer to the line the where terminator was seen. */
java_lex_error ("String not terminated at end of line", 0);
lineno++;
}
@@ -1314,7 +1317,8 @@
string = obstack_finish (&temporary_obstack);
#ifndef JC1_LITE
if (!no_error || (c != '"'))
- java_lval->node = error_mark_node; /* Requires futher testing FIXME */
+ java_lval->node = error_mark_node; /* FIXME: Requires futher
+ testing. */
else
java_lval->node = build_string (strlen (string), string);
#endif
@@ -1322,7 +1326,7 @@
return STRING_LIT_TK;
}
- /* Separator */
+ /* Separator. */
switch (c)
{
case '(':
@@ -1361,7 +1365,7 @@
/* return DOT_TK; */
}
- /* Operators */
+ /* Operators. */
switch (c)
{
case '=':
@@ -1541,7 +1545,7 @@
BUILD_OPERATOR (NOT_TK);
}
- /* Keyword, boolean literal or null literal */
+ /* Keyword, boolean literal or null literal. */
for (first_unicode = c, all_ascii = 1, ascii_index = 0;
JAVA_PART_CHAR_P (c); c = java_get_unicode ())
{
@@ -1597,7 +1601,7 @@
SET_LVAL_NODE (char_type_node);
return INTEGRAL_TK;
- /* Keyword based literals */
+ /* Keyword based literals. */
case TRUE_TK:
case FALSE_TK:
SET_LVAL_NODE ((kw->token == TRUE_TK ?
@@ -1608,7 +1612,7 @@
return NULL_TK;
/* Some keyword we want to retain information on the location
- they where found */
+ they where found. */
case CASE_TK:
case DEFAULT_TK:
case SUPER_TK:
@@ -1628,7 +1632,7 @@
}
}
- /* We may have an ID here */
+ /* We may have an ID here. */
if (JAVA_START_CHAR_P (first_unicode))
{
JAVA_LEX_ID (string);
@@ -1636,7 +1640,7 @@
return ID_TK;
}
- /* Everything else is an invalid character in the input */
+ /* Everything else is an invalid character in the input. */
{
char lex_error_buffer [128];
sprintf (lex_error_buffer, "Invalid character `%s' in input",
@@ -1690,7 +1694,7 @@
obstack_1grow (&temporary_obstack,
(unsigned char)(0x80 | (unicode & 0x3f)));
}
- else /* Range 0x800-0xffff */
+ else /* Range 0x800-0xffff. */
{
obstack_1grow (&temporary_obstack,
(unsigned char)(0xe0 | (unicode & 0xf000) >> 12));
@@ -1707,7 +1711,7 @@
tree node;
{
node = build_expr_wfl (node, ctxp->filename, ctxp->elc.line, ctxp->elc.col);
- /* Prevent java_complete_lhs from short-circuiting node (if constant). */
+ /* Prevent java_complete_lhs from short-circuiting node (if constant). */
TREE_TYPE (node) = NULL_TREE;
return node;
}
@@ -1722,7 +1726,7 @@
ctxp->elc.line = ctxp->c_line->lineno;
ctxp->elc.col = ctxp->c_line->char_col-1+forward;
- /* Might be caught in the middle of some error report */
+ /* Might be caught in the middle of some error report. */
ctxp->java_error_flag = 0;
java_error (NULL);
java_error (msg);
@@ -1759,11 +1763,11 @@
#ifdef JC1_LITE
return 0;
#else
- /* Dumb implementation. Doesn't try to cache or optimize things. */
- /* First line of the file is line 1, first column is 1 */
+ /* Dumb implementation. Doesn't try to cache or optimize things. */
+ /* First line of the file is line 1, first column is 1. */
- /* COL == -1 means, at the CR/LF in LINE */
- /* COL == -2 means, at the first non space char in LINE */
+ /* COL == -1 means, at the CR/LF in LINE. */
+ /* COL == -2 means, at the first non space char in LINE. */
FILE *fp;
int c, ccol, cline = 1;
@@ -1787,7 +1791,7 @@
cline++;
}
- /* Gather the chars of the current line in a buffer */
+ /* Gather the chars of the current line in a buffer. */
for (;;)
{
c = getc (fp);
@@ -1812,11 +1816,11 @@
else
first_non_space = 0;
- /* Place the '^' a the right position */
+ /* Place the '^' a the right position. */
base = obstack_base (&temporary_obstack);
for (ccol = 1; ccol <= col+3; ccol++)
{
- /* Compute \t when reaching first_non_space */
+ /* Compute \t when reaching first_non_space. */
char c = (first_non_space ?
(base [ccol-1] == '\t' ? '\t' : ' ') : ' ');
obstack_1grow (&temporary_obstack, c);
@@ -1981,8 +1985,7 @@
if (r == 0)
{
int i;
- /* We've found a match if all the remaining characters are
- `$'. */
+ /* We've found a match if all the remaining characters are `$'. */
for (i = min_length; i < length && name[i] == '$'; ++i)
;
if (i == length)
More information about the Gcc-patches
mailing list