binutils-gdb/gdb/ada-lex.l
Tom Tromey 63fc2437de Implement real literal extension for Ada
Sometimes it is convenient to be able to specify the exact bits of a
floating-point literal.  For example, you may want to set a
floating-point register to a denormalized value, or to a particular
NaN.

In C, you can do this by combining the "{}" cast with an array
literal, like:

    (gdb) p {double}{0x576488BDD2AE9FFE}
    $1 = 9.8765449999999996e+112

This patch adds a somewhat similar idea to Ada.  It extends the lexer
to allow "l" and "f" suffixes in a based literal.  The "f" indicates a
floating-point literal, and the "l"s control the size of the
floating-point type.

Note that this differs from Ada's based real literals.  I believe
those can also be used to control the bits of a floating-point value,
but they are a bit more cumbersome to use (simplest is binary but
that's also very lengthy).  Also, these aren't implemented in GDB.

I chose not to allow this extension to work with based integer
literals with exponents.  That didn't seem very useful.
2022-03-07 08:27:38 -07:00

712 lines
19 KiB
C++

/* FLEX lexer for Ada expressions, for GDB. -*- c++ -*-
Copyright (C) 1994-2022 Free Software Foundation, Inc.
This file is part of GDB.
This program 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.
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
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>. */
/*----------------------------------------------------------------------*/
/* The converted version of this file is to be included in ada-exp.y, */
/* the Ada parser for gdb. The function yylex obtains characters from */
/* the global pointer lexptr. It returns a syntactic category for */
/* each successive token and places a semantic value into yylval */
/* (ada-lval), defined by the parser. */
DIG [0-9]
NUM10 ({DIG}({DIG}|_)*)
HEXDIG [0-9a-f]
NUM16 ({HEXDIG}({HEXDIG}|_)*)
OCTDIG [0-7]
LETTER [a-z_]
ID ({LETTER}({LETTER}|{DIG}|[\x80-\xff])*|"<"{LETTER}({LETTER}|{DIG})*">")
WHITE [ \t\n]
TICK ("'"{WHITE}*)
GRAPHIC [a-z0-9 #&'()*+,-./:;<>=_|!$%?@\[\]\\^`{}~]
OPER ([-+*/=<>&]|"<="|">="|"**"|"/="|"and"|"or"|"xor"|"not"|"mod"|"rem"|"abs")
EXP (e[+-]{NUM10})
POSEXP (e"+"?{NUM10})
%{
#include "diagnostics.h"
/* Some old versions of flex generate code that uses the "register" keyword,
which clang warns about. This was observed for example with flex 2.5.35,
as shipped with macOS 10.12. The same happens with flex 2.5.37 and g++ 11
which defaults to ISO C++17, that does not allow register storage class
specifiers. */
DIAGNOSTIC_PUSH
DIAGNOSTIC_IGNORE_DEPRECATED_REGISTER
#define NUMERAL_WIDTH 256
#define LONGEST_SIGN ((ULONGEST) 1 << (sizeof(LONGEST) * HOST_CHAR_BIT - 1))
/* Temporary staging for numeric literals. */
static char numbuf[NUMERAL_WIDTH];
static void canonicalizeNumeral (char *s1, const char *);
static struct stoken processString (const char*, int);
static int processInt (struct parser_state *, const char *, const char *,
const char *);
static int processReal (struct parser_state *, const char *);
static struct stoken processId (const char *, int);
static int processAttribute (const char *);
static int find_dot_all (const char *);
static void rewind_to_char (int);
#undef YY_DECL
#define YY_DECL static int yylex ( void )
/* Flex generates a static function "input" which is not used.
Defining YY_NO_INPUT comments it out. */
#define YY_NO_INPUT
#undef YY_INPUT
#define YY_INPUT(BUF, RESULT, MAX_SIZE) \
if ( *pstate->lexptr == '\000' ) \
(RESULT) = YY_NULL; \
else \
{ \
*(BUF) = *pstate->lexptr; \
(RESULT) = 1; \
pstate->lexptr += 1; \
}
static int find_dot_all (const char *);
/* Depth of parentheses. */
static int paren_depth;
%}
%option case-insensitive interactive nodefault noyywrap
%s BEFORE_QUAL_QUOTE
%%
{WHITE} { }
"--".* { yyterminate(); }
{NUM10}{POSEXP} {
canonicalizeNumeral (numbuf, yytext);
char *e_ptr = strrchr (numbuf, 'e');
*e_ptr = '\0';
return processInt (pstate, nullptr, numbuf, e_ptr + 1);
}
{NUM10} {
canonicalizeNumeral (numbuf, yytext);
return processInt (pstate, NULL, numbuf, NULL);
}
{NUM10}"#"{HEXDIG}({HEXDIG}|_)*"#"{POSEXP} {
canonicalizeNumeral (numbuf, yytext);
char *e_ptr = strrchr (numbuf, 'e');
*e_ptr = '\0';
return processInt (pstate, numbuf,
strchr (numbuf, '#') + 1,
e_ptr + 1);
}
/* The "llf" is a gdb extension to allow a floating-point
constant to be written in some other base. The
floating-point number is formed by reinterpreting the
bytes, allowing direct control over the bits. */
{NUM10}(l{0,2}f)?"#"{HEXDIG}({HEXDIG}|_)*"#" {
canonicalizeNumeral (numbuf, yytext);
return processInt (pstate, numbuf, strchr (numbuf, '#') + 1,
NULL);
}
"0x"{HEXDIG}+ {
canonicalizeNumeral (numbuf, yytext+2);
return processInt (pstate, "16#", numbuf, NULL);
}
{NUM10}"."{NUM10}{EXP} {
canonicalizeNumeral (numbuf, yytext);
return processReal (pstate, numbuf);
}
{NUM10}"."{NUM10} {
canonicalizeNumeral (numbuf, yytext);
return processReal (pstate, numbuf);
}
{NUM10}"#"{NUM16}"."{NUM16}"#"{EXP} {
error (_("Based real literals not implemented yet."));
}
{NUM10}"#"{NUM16}"."{NUM16}"#" {
error (_("Based real literals not implemented yet."));
}
<INITIAL>"'"({GRAPHIC}|\")"'" {
yylval.typed_val.val = yytext[1];
yylval.typed_val.type = type_for_char (pstate, yytext[1]);
return CHARLIT;
}
<INITIAL>"'[\""{HEXDIG}{2,}"\"]'" {
ULONGEST v = strtoulst (yytext+3, nullptr, 16);
yylval.typed_val.val = v;
yylval.typed_val.type = type_for_char (pstate, v);
return CHARLIT;
}
/* Note that we don't handle bracket sequences of more than 2
digits here. Currently there's no support for wide or
wide-wide strings. */
\"({GRAPHIC}|"[\""({HEXDIG}{2,}|\")"\"]")*\" {
yylval.sval = processString (yytext+1, yyleng-2);
return STRING;
}
\" {
error (_("ill-formed or non-terminated string literal"));
}
if {
rewind_to_char ('i');
return 0;
}
task {
rewind_to_char ('t');
return 0;
}
thread{WHITE}+{DIG} {
/* This keyword signals the end of the expression and
will be processed separately. */
rewind_to_char ('t');
return 0;
}
/* ADA KEYWORDS */
abs { return ABS; }
and { return _AND_; }
else { return ELSE; }
in { return IN; }
mod { return MOD; }
new { return NEW; }
not { return NOT; }
null { return NULL_PTR; }
or { return OR; }
others { return OTHERS; }
rem { return REM; }
then { return THEN; }
xor { return XOR; }
/* BOOLEAN "KEYWORDS" */
/* True and False are not keywords in Ada, but rather enumeration constants.
However, the boolean type is no longer represented as an enum, so True
and False are no longer defined in symbol tables. We compromise by
making them keywords (when bare). */
true { return TRUEKEYWORD; }
false { return FALSEKEYWORD; }
/* ATTRIBUTES */
{TICK}[a-z][a-z_]+ { BEGIN INITIAL; return processAttribute (yytext+1); }
/* PUNCTUATION */
"=>" { return ARROW; }
".." { return DOTDOT; }
"**" { return STARSTAR; }
":=" { return ASSIGN; }
"/=" { return NOTEQUAL; }
"<=" { return LEQ; }
">=" { return GEQ; }
<BEFORE_QUAL_QUOTE>"'" { BEGIN INITIAL; return '\''; }
[-&*+./:<>=|;\[\]] { return yytext[0]; }
"," { if (paren_depth == 0 && pstate->comma_terminates)
{
rewind_to_char (',');
return 0;
}
else
return ',';
}
"(" { paren_depth += 1; return '('; }
")" { if (paren_depth == 0)
{
rewind_to_char (')');
return 0;
}
else
{
paren_depth -= 1;
return ')';
}
}
"."{WHITE}*all { return DOT_ALL; }
"."{WHITE}*{ID} {
yylval.sval = processId (yytext+1, yyleng-1);
return DOT_ID;
}
{ID}({WHITE}*"."{WHITE}*({ID}|\"{OPER}\"))*(" "*"'")? {
int all_posn = find_dot_all (yytext);
if (all_posn == -1 && yytext[yyleng-1] == '\'')
{
BEGIN BEFORE_QUAL_QUOTE;
yyless (yyleng-1);
}
else if (all_posn >= 0)
yyless (all_posn);
yylval.sval = processId (yytext, yyleng);
return NAME;
}
/* GDB EXPRESSION CONSTRUCTS */
"'"[^']+"'"{WHITE}*:: {
yyless (yyleng - 2);
yylval.sval = processId (yytext, yyleng);
return NAME;
}
"::" { return COLONCOLON; }
[{}@] { return yytext[0]; }
/* REGISTERS AND GDB CONVENIENCE VARIABLES */
"$"({LETTER}|{DIG}|"$")* {
yylval.sval.ptr = yytext;
yylval.sval.length = yyleng;
return DOLLAR_VARIABLE;
}
/* CATCH-ALL ERROR CASE */
. { error (_("Invalid character '%s' in expression."), yytext); }
%%
#include <ctype.h>
/* Initialize the lexer for processing new expression. */
static void
lexer_init (FILE *inp)
{
BEGIN INITIAL;
paren_depth = 0;
yyrestart (inp);
}
/* Copy S2 to S1, removing all underscores, and downcasing all letters. */
static void
canonicalizeNumeral (char *s1, const char *s2)
{
for (; *s2 != '\000'; s2 += 1)
{
if (*s2 != '_')
{
*s1 = tolower(*s2);
s1 += 1;
}
}
s1[0] = '\000';
}
/* Interprets the prefix of NUM that consists of digits of the given BASE
as an integer of that BASE, with the string EXP as an exponent.
Puts value in yylval, and returns INT, if the string is valid. Causes
an error if the number is improperly formated. BASE, if NULL, defaults
to "10", and EXP to "1". The EXP does not contain a leading 'e' or 'E'.
*/
static int
processInt (struct parser_state *par_state, const char *base0,
const char *num0, const char *exp0)
{
long exp;
int base;
/* For the based literal with an "f" prefix, we'll return a
floating-point number. This counts the the number of "l"s seen,
to decide the width of the floating-point number to return. -1
means no "f". */
int floating_point_l_count = -1;
if (base0 == NULL)
base = 10;
else
{
char *end_of_base;
base = strtol (base0, &end_of_base, 10);
if (base < 2 || base > 16)
error (_("Invalid base: %d."), base);
while (*end_of_base == 'l')
{
++floating_point_l_count;
++end_of_base;
}
/* This assertion is ensured by the pattern. */
gdb_assert (floating_point_l_count == -1 || *end_of_base == 'f');
if (*end_of_base == 'f')
{
++end_of_base;
++floating_point_l_count;
}
/* This assertion is ensured by the pattern. */
gdb_assert (*end_of_base == '#');
}
if (exp0 == NULL)
exp = 0;
else
exp = strtol(exp0, (char **) NULL, 10);
gdb_mpz result;
while (isxdigit (*num0))
{
int dig = fromhex (*num0);
if (dig >= base)
error (_("Invalid digit `%c' in based literal"), *num0);
mpz_mul_ui (result.val, result.val, base);
mpz_add_ui (result.val, result.val, dig);
++num0;
}
while (exp > 0)
{
mpz_mul_ui (result.val, result.val, base);
exp -= 1;
}
if (floating_point_l_count > -1)
{
struct type *fp_type;
if (floating_point_l_count == 0)
fp_type = language_lookup_primitive_type (par_state->language (),
par_state->gdbarch (),
"float");
else if (floating_point_l_count == 1)
fp_type = language_lookup_primitive_type (par_state->language (),
par_state->gdbarch (),
"long_float");
else
{
/* This assertion is ensured by the pattern. */
gdb_assert (floating_point_l_count == 2);
fp_type = language_lookup_primitive_type (par_state->language (),
par_state->gdbarch (),
"long_long_float");
}
yylval.typed_val_float.type = fp_type;
result.write (gdb::make_array_view (yylval.typed_val_float.val,
TYPE_LENGTH (fp_type)),
type_byte_order (fp_type),
true);
return FLOAT;
}
gdb_mpz maxval (ULONGEST_MAX / base);
if (mpz_cmp (result.val, maxval.val) > 0)
error (_("Integer literal out of range"));
LONGEST value = result.as_integer<LONGEST> ();
if ((value >> (gdbarch_int_bit (par_state->gdbarch ())-1)) == 0)
yylval.typed_val.type = type_int (par_state);
else if ((value >> (gdbarch_long_bit (par_state->gdbarch ())-1)) == 0)
yylval.typed_val.type = type_long (par_state);
else if (((value >> (gdbarch_long_bit (par_state->gdbarch ())-1)) >> 1) == 0)
{
/* We have a number representable as an unsigned integer quantity.
For consistency with the C treatment, we will treat it as an
anonymous modular (unsigned) quantity. Alas, the types are such
that we need to store .val as a signed quantity. Sorry
for the mess, but C doesn't officially guarantee that a simple
assignment does the trick (no, it doesn't; read the reference manual).
*/
yylval.typed_val.type
= builtin_type (par_state->gdbarch ())->builtin_unsigned_long;
if (value & LONGEST_SIGN)
yylval.typed_val.val =
(LONGEST) (value & ~LONGEST_SIGN)
- (LONGEST_SIGN>>1) - (LONGEST_SIGN>>1);
else
yylval.typed_val.val = (LONGEST) value;
return INT;
}
else
yylval.typed_val.type = type_long_long (par_state);
yylval.typed_val.val = value;
return INT;
}
static int
processReal (struct parser_state *par_state, const char *num0)
{
yylval.typed_val_float.type = type_long_double (par_state);
bool parsed = parse_float (num0, strlen (num0),
yylval.typed_val_float.type,
yylval.typed_val_float.val);
gdb_assert (parsed);
return FLOAT;
}
/* Store a canonicalized version of NAME0[0..LEN-1] in yylval.ssym. The
resulting string is valid until the next call to ada_parse. If
NAME0 contains the substring "___", it is assumed to be already
encoded and the resulting name is equal to it. Similarly, if the name
starts with '<', it is copied verbatim. Otherwise, it differs
from NAME0 in that:
+ Characters between '...' are transfered verbatim to yylval.ssym.
+ Trailing "'" characters in quoted sequences are removed (a leading quote is
preserved to indicate that the name is not to be GNAT-encoded).
+ Unquoted whitespace is removed.
+ Unquoted alphabetic characters are mapped to lower case.
Result is returned as a struct stoken, but for convenience, the string
is also null-terminated. Result string valid until the next call of
ada_parse.
*/
static struct stoken
processId (const char *name0, int len)
{
char *name = (char *) obstack_alloc (&temp_parse_space, len + 11);
int i0, i;
struct stoken result;
result.ptr = name;
while (len > 0 && isspace (name0[len-1]))
len -= 1;
if (name0[0] == '<' || strstr (name0, "___") != NULL)
{
strncpy (name, name0, len);
name[len] = '\000';
result.length = len;
return result;
}
i = i0 = 0;
while (i0 < len)
{
if (isalnum (name0[i0]))
{
name[i] = tolower (name0[i0]);
i += 1; i0 += 1;
}
else switch (name0[i0])
{
default:
name[i] = name0[i0];
i += 1; i0 += 1;
break;
case ' ': case '\t':
i0 += 1;
break;
case '\'':
do
{
name[i] = name0[i0];
i += 1; i0 += 1;
}
while (i0 < len && name0[i0] != '\'');
i0 += 1;
break;
}
}
name[i] = '\000';
result.length = i;
return result;
}
/* Return TEXT[0..LEN-1], a string literal without surrounding quotes,
with special hex character notations replaced with characters.
Result valid until the next call to ada_parse. */
static struct stoken
processString (const char *text, int len)
{
const char *p;
char *q;
const char *lim = text + len;
struct stoken result;
q = (char *) obstack_alloc (&temp_parse_space, len);
result.ptr = q;
p = text;
while (p < lim)
{
if (p[0] == '[' && p[1] == '"' && p+2 < lim)
{
if (p[2] == '"') /* "...["""]... */
{
*q = '"';
p += 4;
}
else
{
const char *end;
ULONGEST chr = strtoulst (p + 2, &end, 16);
if (chr > 0xff)
error (_("wide strings are not yet supported"));
*q = (char) chr;
p = end + 1;
}
}
else
*q = *p;
q += 1;
p += 1;
}
result.length = q - result.ptr;
return result;
}
/* Returns the position within STR of the '.' in a
'.{WHITE}*all' component of a dotted name, or -1 if there is none.
Note: we actually don't need this routine, since 'all' can never be an
Ada identifier. Thus, looking up foo.all or foo.all.x as a name
must fail, and will eventually be interpreted as (foo).all or
(foo).all.x. However, this does avoid an extraneous lookup. */
static int
find_dot_all (const char *str)
{
int i;
for (i = 0; str[i] != '\000'; i++)
if (str[i] == '.')
{
int i0 = i;
do
i += 1;
while (isspace (str[i]));
if (strncasecmp (str + i, "all", 3) == 0
&& !isalnum (str[i + 3]) && str[i + 3] != '_')
return i0;
}
return -1;
}
/* Returns non-zero iff string SUBSEQ matches a subsequence of STR, ignoring
case. */
static int
subseqMatch (const char *subseq, const char *str)
{
if (subseq[0] == '\0')
return 1;
else if (str[0] == '\0')
return 0;
else if (tolower (subseq[0]) == tolower (str[0]))
return subseqMatch (subseq+1, str+1) || subseqMatch (subseq, str+1);
else
return subseqMatch (subseq, str+1);
}
static struct { const char *name; int code; }
attributes[] = {
{ "address", TICK_ADDRESS },
{ "unchecked_access", TICK_ACCESS },
{ "unrestricted_access", TICK_ACCESS },
{ "access", TICK_ACCESS },
{ "first", TICK_FIRST },
{ "last", TICK_LAST },
{ "length", TICK_LENGTH },
{ "max", TICK_MAX },
{ "min", TICK_MIN },
{ "modulus", TICK_MODULUS },
{ "pos", TICK_POS },
{ "range", TICK_RANGE },
{ "size", TICK_SIZE },
{ "tag", TICK_TAG },
{ "val", TICK_VAL },
{ NULL, -1 }
};
/* Return the syntactic code corresponding to the attribute name or
abbreviation STR. */
static int
processAttribute (const char *str)
{
int i, k;
for (i = 0; attributes[i].code != -1; i += 1)
if (strcasecmp (str, attributes[i].name) == 0)
return attributes[i].code;
for (i = 0, k = -1; attributes[i].code != -1; i += 1)
if (subseqMatch (str, attributes[i].name))
{
if (k == -1)
k = i;
else
error (_("ambiguous attribute name: `%s'"), str);
}
if (k == -1)
error (_("unrecognized attribute: `%s'"), str);
return attributes[k].code;
}
/* Back up lexptr by yyleng and then to the rightmost occurrence of
character CH, case-folded (there must be one). WARNING: since
lexptr points to the next input character that Flex has not yet
transferred to its internal buffer, the use of this function
depends on the assumption that Flex calls YY_INPUT only when it is
logically necessary to do so (thus, there is no reading ahead
farther than needed to identify the next token.) */
static void
rewind_to_char (int ch)
{
pstate->lexptr -= yyleng;
while (toupper (*pstate->lexptr) != toupper (ch))
pstate->lexptr -= 1;
yyrestart (NULL);
}
/* Dummy definition to suppress warnings about unused static definitions. */
typedef void (*dummy_function) ();
dummy_function ada_flex_use[] =
{
(dummy_function) yyunput
};
DIAGNOSTIC_POP