ack/lang/cem/cemcom.ansi/ival.g

761 lines
18 KiB
Plaintext
Raw Normal View History

/*
* (c) copyright 1987 by the Vrije Universiteit, Amsterdam, The Netherlands.
* See the copyright notice in the ACK home directory, in the file "Copyright".
*/
/* $Id$ */
/* CODE FOR THE INITIALISATION OF GLOBAL VARIABLES */
{
#include <assert.h>
#include <stdlib.h>
#include "parameters.h"
#ifndef LINT
#include <em.h>
#else
#include "l_em.h"
#include "l_lint.h"
#endif /* LINT */
#include <ack_string.h>
#include <alloc.h>
#include <assert.h>
#include <string.h>
#include <flt_arith.h>
#include "idf.h"
#include "arith.h"
#include "label.h"
#include "expr.h"
#include "type.h"
#include "proto.h"
#include "struct.h"
#include "field.h"
#include "Lpars.h"
#include "sizes.h"
#include "align.h"
#include "level.h"
#include "error.h"
#include "def.h"
#include "LLlex.h"
#include "estack.h"
#include "stack.h"
#include "ch3.h"
#define con_nullbyte() C_con_ucon("0", (arith)1)
#define aggregate_type(tp) ((tp)->tp_fund == ARRAY || (tp)->tp_fund == STRUCT)
extern char options[];
static int gen_error;
static int pack_level;
struct e_stack *p_stack;
void gen_tpcheck(struct type **);
void gen_simple_exp(struct type **, struct expr **);
struct type **arr_elem(struct type **, struct e_stack *);
struct sdef *next_field(register struct sdef *,register struct e_stack *);
struct type **gen_tphead(struct type **, int);
struct type **gen_tpmiddle(void);
struct sdef *gen_align_to_next(register struct e_stack *);
void gen_tpend(void);
void check_and_pad(struct expr **, struct type **);
void pad(struct type *);
void check_ival(struct expr **, register struct type *);
void ch_array(struct type **, /* type tp = array of characters */
struct expr *);
void str_cst(register char *, register int, int);
#ifndef NOBITFIELD
void put_bf(struct type *, arith );
#endif /* NOBITFIELD */
int zero_bytes(register struct sdef *);
int valid_type(struct type *, char *);
void con_int(register struct expr *);
void illegal_init_cst(struct expr *);
void too_many_initialisers(void);
}
/* initial_value recursively guides the initialisation expression.
*/
/* 3.5 */
initial_value(register struct type **tpp; register struct expr **expp;) :
{ if (tpp) gen_tpcheck(tpp); }
[
{ if (pack_level == 0) gen_error = 0; }
assignment_expression(expp)
{
#ifdef LINT
lint_expr(*expp, USED);
#endif /* LINT */
if ((*expp)->ex_type->tp_fund == ARRAY)
array2pointer(*expp);
if (tpp) {
if (level >= L_LOCAL
|| is_ld_cst(*expp)
|| is_fp_cst(*expp)
|| (*expp)->ex_class == String) {
gen_simple_exp(tpp, expp);
free_expression(*expp);
*expp = 0;
} else {
expr_error(*expp,"illegal initialization");
free_expression(*expp);
*expp = 0;
}
}
}
|
initial_value_pack(tpp, expp)
]
;
initial_value_pack(struct type **tpp; struct expr **expp;)
:
'{'
{ if (pack_level == 0) gen_error = 0; pack_level++; }
initial_value_list(tpp, expp)
{ pack_level--;
if (!pack_level) {
while (p_stack) {
struct e_stack *p = p_stack->next;
free_e_stack(p_stack);
p_stack = p;
}
}
if (pack_level < gen_error) gen_error = 0;
}
'}'
;
initial_value_list(register struct type **tpp; struct expr **expp;)
{ struct expr *e1;
register struct type **tpp2 = 0;
int err_flag = gen_error;
}
:
{ if (tpp) tpp2 = gen_tphead(tpp, 0); }
initial_value(tpp2, &e1)
{ if (!tpp) init_expression(&expp, e1); }
[%while (AHEAD != '}') /* >>> conflict on ',' */
','
{ if (tpp) tpp2 = gen_tpmiddle(); }
initial_value(tpp2, &e1)
{ if (!tpp) init_expression(&expp, e1); }
]*
{ if (tpp && ! err_flag) gen_tpend(); }
','? /* optional trailing comma */
;
{
void gen_tpcheck(struct type **tpp)
{
register struct type *tp;
if (gen_error) return;
switch((tp = *tpp)->tp_fund) {
case ARRAY:
if (! valid_type(tp->tp_up, "array element"))
gen_error = pack_level;
break;
case STRUCT:
if (! valid_type(tp, "struct"))
gen_error = pack_level;
break;
case UNION:
if (! valid_type(tp, "union"))
gen_error = pack_level;
break;
case ERRONEOUS:
if (! gen_error) gen_error = pack_level;
break;
}
}
void gen_simple_exp(struct type **tpp, struct expr **expp)
{
register struct type *tp;
if (gen_error) return;
tp = *tpp;
switch(tp->tp_fund) {
case ARRAY:
if ((*expp)->ex_class == String && tp->tp_up->tp_fund == CHAR) {
ch_array(tpp,*expp);
break;
}
/* Fall through */
case UNION:
case STRUCT:
check_and_pad(expp, tpp);
break;
case ERRONEOUS:
case FUNCTION:
case VOID:
gen_error = pack_level;
break;
default:
check_ival(expp, tp);
break;
}
}
struct type **arr_elem(struct type **tpp, struct e_stack *p)
{
register struct type *tp = *tpp;
if (tp->tp_up->tp_fund == CHAR && AHEAD == STRING && p->elem_count == 1) {
p->nelem = 1;
return tpp;
}
if (AHEAD == '{' || (! aggregate_type(tp->tp_up) && tp->tp_up->tp_fund != UNION))
return &(tp->tp_up);
return gen_tphead(&(tp->tp_up), 1);
}
struct sdef *next_field(register struct sdef *sd,
register struct e_stack *p)
{
if (sd->sd_sdef)
p->bytes_upto_here += zero_bytes(sd);
p->bytes_upto_here +=
size_of_type(sd->sd_type, "selector");
p->last_offset = sd->sd_offset;
return sd->sd_sdef;
}
struct type **gen_tphead(struct type **tpp, int nest)
{
register struct type *tp = *tpp;
register struct e_stack *p;
register struct sdef *sd;
if (tpp && *tpp == error_type) {
gen_error = pack_level;
return 0;
}
if (gen_error) return tpp;
if (tp->tp_fund == UNION) {
/* Here, we saw a {, which could be the start of a union
initializer. It could, however, also be the start of the
initializer for the first union field ...
*/
sd = tp->tp_sdef;
if (AHEAD != '{' &&
(aggregate_type(sd->sd_type) ||
sd->sd_type->tp_fund == UNION)) {
/* In this case, assume that it is the start of the
initializer of the union field, so:
*/
return gen_tphead(&(tp->tp_sdef->sd_type), nest);
}
}
p = new_e_stack();
p->next = p_stack;
p_stack = p;
p->s_nested = nest;
p->s_tpp = tpp;
switch(tp->tp_fund) {
case UNION:
p->s_def = sd = tp->tp_sdef;
p->bytes_upto_here = 0;
return &(sd->sd_type);
case ARRAY:
p->nelem = -1;
p->elem_count = 1;
if (tp->tp_size != (arith) -1) {
p->nelem = (tp->tp_size / tp->tp_up->tp_size);
}
return arr_elem(tpp, p);
case STRUCT:
p->s_def = sd = tp->tp_sdef;
p->bytes_upto_here = 0;
p->last_offset = -1;
#ifndef NOBITFIELD
while (sd && is_anon_idf(sd->sd_idf)) {
put_bf(sd->sd_type, (arith) 0);
sd = next_field(sd, p);
}
#endif
if (! sd) {
/* something wrong with this struct */
gen_error = pack_level;
p_stack = p->next;
free_e_stack(p);
return 0;
}
p->s_def = sd;
if (AHEAD != '{' && aggregate_type(sd->sd_type)) {
return gen_tphead(&(sd->sd_type), 1);
}
return &(sd->sd_type);
case ERRONEOUS:
if (! gen_error) gen_error = pack_level;
/* fall through */
default:
p->nelem = 1;
p->elem_count = 1;
return tpp;
}
}
struct type **gen_tpmiddle(void)
{
register struct type *tp;
register struct sdef *sd;
register struct e_stack *p = p_stack;
if (gen_error) {
if (p) return p->s_tpp;
return 0;
}
again:
tp = *(p->s_tpp);
switch(tp->tp_fund) {
case ERRONEOUS:
if (! gen_error) gen_error = pack_level;
return p->s_tpp;
case UNION:
sd = p->s_def;
p->bytes_upto_here +=
size_of_type(sd->sd_type, "selector");
return p->s_tpp;
default:
if (p->elem_count == p->nelem && p->s_nested) {
p = p->next;
free_e_stack(p_stack);
p_stack = p;
goto again;
}
p->elem_count++;
if (p->nelem >= 0 && p->elem_count > p->nelem) {
too_many_initialisers();
return p->s_tpp;
}
if (tp->tp_fund == ARRAY) {
return arr_elem(p->s_tpp, p);
}
return p->s_tpp;
case STRUCT:
sd = gen_align_to_next(p);
if (! sd) {
while (p->bytes_upto_here++ < tp->tp_size)
con_nullbyte();
if (p->s_nested) {
p = p->next;
free_e_stack(p_stack);
p_stack = p;
goto again;
}
too_many_initialisers();
return p->s_tpp;
}
if (AHEAD != '{' && aggregate_type(sd->sd_type)) {
return gen_tphead(&(sd->sd_type), 1);
}
return &(sd->sd_type);
}
}
struct sdef *gen_align_to_next(register struct e_stack *p)
{
register struct sdef *sd = p->s_def;
if (! sd) return sd;
#ifndef NOBITFIELD
do {
if (is_anon_idf(sd->sd_idf)) put_bf(sd->sd_type, (arith) 0);
#endif
sd = next_field(sd, p);
#ifndef NOBITFIELD
} while (sd && is_anon_idf(sd->sd_idf));
#endif
p->s_def = sd;
return sd;
}
void gen_tpend(void)
{
register struct e_stack *p = p_stack;
register struct type *tp;
register struct sdef *sd;
int getout = 0;
while (!getout && p) {
if (!gen_error) {
tp = *(p->s_tpp);
switch(tp->tp_fund) {
case UNION:
sd = p->s_def;
p->bytes_upto_here +=
size_of_type(sd->sd_type, "selector");
while (p->bytes_upto_here++ < tp->tp_size)
con_nullbyte();
break;
case ARRAY:
if (tp->tp_size == -1) {
*(p->s_tpp) = construct_type(ARRAY, tp->tp_up,
0, p->elem_count, NO_PROTO);
}
else {
while (p->nelem-- > p->elem_count) {
pad(tp->tp_up);
}
}
break;
case STRUCT:
sd = gen_align_to_next(p);
while (sd) {
pad(sd->sd_type);
if (sd->sd_sdef)
p->bytes_upto_here += zero_bytes(sd);
p->bytes_upto_here +=
size_of_type(sd->sd_type, "selector");
sd = sd->sd_sdef;
}
while (p->bytes_upto_here++ < tp->tp_size)
con_nullbyte();
break;
}
}
if (! p->s_nested) getout = 1;
p = p->next;
free_e_stack(p_stack);
p_stack = p;
}
}
/* check_and_pad() is given a simple initialisation expression
where the type can be either a simple or an aggregate type.
In the latter case, only the first member is initialised and
the rest is zeroed.
*/
void check_and_pad(struct expr **expp, struct type **tpp)
{
register struct type *tp = *tpp;
if (tp->tp_fund == ARRAY) {
check_and_pad(expp, &(tp->tp_up)); /* first member */
if (tp->tp_size == (arith)-1)
/* no size specified upto here: just
set it to the size of one member.
*/
tp = *tpp = construct_type(ARRAY, tp->tp_up,
0, (arith)1, NO_PROTO);
else {
register int dim = tp->tp_size / tp->tp_up->tp_size;
/* pad remaining members with zeroes */
while (--dim > 0)
pad(tp->tp_up);
}
}
else
if (tp->tp_fund == STRUCT) {
register struct sdef *sd = tp->tp_sdef;
check_and_pad(expp, &(sd->sd_type));
/* next selector is aligned by adding extra zeroes */
if (sd->sd_sdef)
zero_bytes(sd);
2019-02-23 16:46:05 +00:00
while ( (sd = sd->sd_sdef)!=0) { /* pad remaining selectors */
pad(sd->sd_type);
if (sd->sd_sdef)
zero_bytes(sd);
}
}
else if (tp->tp_fund == UNION) {
/* only the first selector can be initialized */
register struct sdef *sd = tp->tp_sdef;
check_and_pad(expp, &(sd->sd_type));
}
else /* simple type */
check_ival(expp, tp);
}
/* pad() fills an element of type tp with zeroes.
If the element is an aggregate, pad() is called recursively.
*/
void pad(struct type *tpx)
{
register struct type *tp = tpx;
register arith sz = tp->tp_size;
gen_tpcheck(&tpx);
if (gen_error) return;
#ifndef NOBITFIELD
if (tp->tp_fund == FIELD) {
put_bf(tp, (arith)0);
return;
}
#endif /* NOBITFIELD */
if (tp->tp_align >= word_align) while (sz >= word_size) {
C_con_cst((arith) 0);
sz -= word_size;
}
while (sz) {
C_con_icon("0", (arith) 1);
sz--;
}
}
/* check_ival() checks whether the initialisation of an element
of a fundamental type is legal and, if so, performs the initialisation
by directly generating the necessary code.
No further comment is needed to explain the internal structure
of this straightforward function.
*/
void check_ival(struct expr **expp, register struct type *tp)
{
/* The philosophy here is that ch3cast puts an explicit
conversion node in front of the expression if the types
are not compatible. In this case, the initialisation
expression is no longer a constant.
*/
register struct expr *expr = *expp;
switch (tp->tp_fund) {
case CHAR:
case SHORT:
case INT:
case LONG:
Begin to add `long long` to C compiler for linux386. Add long long type, but without literals; you can't say '123LL' yet. You can try constant operations, like `(long long)123 + 1`, but the compiler's `arith` type might not be wide enough. Conversions, shifts, and some other operations don't work in i386 ncg; I am using a union instead of conversions: union q { long long ll; unsigned long long ull; int i[2]; }; Hack plat/linux386/descr to enable long long (size 8, alignment 4) only for this platform. The default for other platforms is to disable long long (size -1). In lang/cem/cemcom.ansi, - BigPars, SmallPars: Add default size, alignment of long long. - align.h: Add lnglng_align. - arith.c: Convert arithmetic operands to long long or unsigned long long when necessary; avoid conversion from long long to long. Allow long long as an arithmetic, integral, or logical operand. - ch3.c: Handle long long like int and long when erroneously applying a selector, like `long long ll; ll.member` or `ll->member`. Add long long to integral and arithmetic types. - code.c: Add long long to type stabs for debugging. - conversion.c: Add long long to integral conversions. - cstoper.c: Write masks up to full_mask[8]. Add FIXME comment. - declar.g: Parse `long long` in code. - decspecs.c: Understand long long in type declarations. - eval.c: Add long long to operations, to generate code like `adi 8`. Don't use `ldc` with constant over 4 bytes. - ival.g: Allow long long in initializations. - main.c: Set lnglng_type and related values. - options.c: Add option like `-Vq8.4` to set long long to size 8, alignment 4. I chose 'q', because Perl's pack and Ruby's Array#pack use 'q' for 64-bit or long long values; it might be a reference to BSD's old quad_t alias for long long. - sizes.h: Add lnglng_size. - stab.c: Allow long long when writing the type stab for debugging. Switch from calculating the ranges to hardcoding them in strings; add 8-byte ranges as a special case. This also hardcodes the unsigned 4-byte range as "0;-1". Before it was either "0;-1" or "0;4294967295", depending on sizeof(long) in the compiler. - struct.c: Try long long bitfield, but it will probably give the error, "bit field type long long does not fit in a word". - switch.c: Update comment. - tokenname.c: Define LNGLNG (long long) like LNGDBL (long double). - type.c, type.str: Add lnglng_type and ulnglng_type. Add function no_long_long() to check if long long is disabled.
2019-09-02 15:24:44 +00:00
case LNGLNG:
case ENUM:
case POINTER:
ch3cast(expp, '=', tp);
expr = *expp;
#ifdef DEBUG
print_expr("init-expr after cast", expr);
#endif /* DEBUG */
if (!is_ld_cst(expr))
illegal_init_cst(expr);
else
if (expr->VL_CLASS == Const)
con_int(expr);
else
if (expr->VL_CLASS == Name) {
register struct idf *idf = expr->VL_IDF;
if (idf->id_def->df_level >= L_LOCAL
&& idf->id_def->df_sc != GLOBAL
&& idf->id_def->df_sc != EXTERN) {
illegal_init_cst(expr);
}
else /* e.g., int f(); int p = f; */
if (idf->id_def->df_type->tp_fund == FUNCTION)
C_con_pnam(idf->id_text);
else /* e.g., int a; int *p = &a; */
Add long long literals like 123LL to ACK C. For now, a long long literal must have the 'LL' or 'll' suffix. A literal without 'LL' or 'll' acts as before: it may become unsigned long but not long long. (For targets where int and long have the same size, some literals change from unsigned int to unsigned long.) Type `arith` may be too narrow for long long values. Add a second type `writh` for wide arithmetic, and change some variables from arith to writh. This may cause bugs if I forget to use writh, or if a conversion from writh to arith overflows. I mark some conversions with (arith) or (writh) casts. - BigPars, SmallPars: Remove SPECIAL_ARITHMETICS. This feature would change arith to a different type, but can't work, because it would conflict with definitions of arith in both <em_arith.h> and <flt_arith.h>. - LLlex.c: Understand 'LL' or 'll' suffix. Cut size of constant when it overflows writh, not only when it overflows the target machine's types. (This cut might not be necessary, because we might cut it again later.) When picking signed long or unsigned long, check the target's long type, not the compiler's arith type; the old check for `val >= 0` was broken where sizeof(arith) > 4. - LLlex.h: Change struct token's tok_ival to writh, so it can hold a long long literal. - arith.c: Adjust to VL_VALUE being writh. Don't convert between float and integer at compile-time if the integer might be too wide for <flt_arith.h>. Add writh2str(), because writh might be too wide for long2str(). - arith.h: Remove SPECIAL_ARITHMETICS. Declare full_mask[] here, not in several *.c files. Declare writh2str(). - ch3.c, ch3bin.c, ch3mon.c, declarator.c, statement.g: Remove obsolete casts. Adjust to VL_VALUE being writh. - conversion.c, stab.c: Don't declare full_mask[]. - cstoper.c: Use writh for constant operations on VL_VALUE, and for full_mask[]. - declar., field.c, ival.g: Add casts. - dumpidf.c: Need to #include "parameters.h" before checking DEBUG. Use writh2str, because "%ld" might not work. - eval.c, eval.h: Add casts. Use writh when writing a wide constant in EM. - expr.c: Add and remove casts. In fill_int_expr(), make expression from long long literal. In chk_cst_expr(), allow long long as constant expression, so the compiler may accept `case 123LL:` in a switch statement. - expr.str: Change struct value's vl_value and struct expr's VL_VALUE to writh, so an expression may have a long long value at compile time. - statement.g: Remove obsolete casts. - switch.c, switch.str: Use writh in case entries for switch statements, so `switch (ll) {...}` with long long ll works. - tokenname.c: Add ULNGLNG so LLlex.c can use it for literals.
2019-09-05 02:14:38 +00:00
C_con_dnam(idf->id_text, (arith)expr->VL_VALUE);
}
else {
assert(expr->VL_CLASS == Label);
Add long long literals like 123LL to ACK C. For now, a long long literal must have the 'LL' or 'll' suffix. A literal without 'LL' or 'll' acts as before: it may become unsigned long but not long long. (For targets where int and long have the same size, some literals change from unsigned int to unsigned long.) Type `arith` may be too narrow for long long values. Add a second type `writh` for wide arithmetic, and change some variables from arith to writh. This may cause bugs if I forget to use writh, or if a conversion from writh to arith overflows. I mark some conversions with (arith) or (writh) casts. - BigPars, SmallPars: Remove SPECIAL_ARITHMETICS. This feature would change arith to a different type, but can't work, because it would conflict with definitions of arith in both <em_arith.h> and <flt_arith.h>. - LLlex.c: Understand 'LL' or 'll' suffix. Cut size of constant when it overflows writh, not only when it overflows the target machine's types. (This cut might not be necessary, because we might cut it again later.) When picking signed long or unsigned long, check the target's long type, not the compiler's arith type; the old check for `val >= 0` was broken where sizeof(arith) > 4. - LLlex.h: Change struct token's tok_ival to writh, so it can hold a long long literal. - arith.c: Adjust to VL_VALUE being writh. Don't convert between float and integer at compile-time if the integer might be too wide for <flt_arith.h>. Add writh2str(), because writh might be too wide for long2str(). - arith.h: Remove SPECIAL_ARITHMETICS. Declare full_mask[] here, not in several *.c files. Declare writh2str(). - ch3.c, ch3bin.c, ch3mon.c, declarator.c, statement.g: Remove obsolete casts. Adjust to VL_VALUE being writh. - conversion.c, stab.c: Don't declare full_mask[]. - cstoper.c: Use writh for constant operations on VL_VALUE, and for full_mask[]. - declar., field.c, ival.g: Add casts. - dumpidf.c: Need to #include "parameters.h" before checking DEBUG. Use writh2str, because "%ld" might not work. - eval.c, eval.h: Add casts. Use writh when writing a wide constant in EM. - expr.c: Add and remove casts. In fill_int_expr(), make expression from long long literal. In chk_cst_expr(), allow long long as constant expression, so the compiler may accept `case 123LL:` in a switch statement. - expr.str: Change struct value's vl_value and struct expr's VL_VALUE to writh, so an expression may have a long long value at compile time. - statement.g: Remove obsolete casts. - switch.c, switch.str: Use writh in case entries for switch statements, so `switch (ll) {...}` with long long ll works. - tokenname.c: Add ULNGLNG so LLlex.c can use it for literals.
2019-09-05 02:14:38 +00:00
C_con_dlb(expr->VL_LBL, (arith)expr->VL_VALUE);
}
break;
case FLOAT:
case DOUBLE:
case LNGDBL:
ch3cast(expp, '=', tp);
expr = *expp;
#ifdef DEBUG
print_expr("init-expr after cast", expr);
#endif /* DEBUG */
if (expr->ex_class == Float) {
char buf[FLT_STRLEN];
flt_flt2str(&(expr->FL_ARITH), buf, FLT_STRLEN);
C_con_fcon(buf, expr->ex_type->tp_size);
}
#ifdef NOTDEF
Coercion from int to float is now always done compile time.
This, to accept declarations like
double x = -(double)1;
and also to prevent runtime coercions for compile-time constants.
else
if (expr->ex_class == Oper && expr->OP_OPER == INT2FLOAT) {
/* float f = 1; */
expr = expr->OP_RIGHT;
if (is_cp_cst(expr))
C_con_fcon(long2str((long)expr->VL_VALUE, 10),
tp->tp_size);
else
illegal_init_cst(expr);
}
#endif /* NOTDEF */
else
illegal_init_cst(expr);
break;
#ifndef NOBITFIELD
case FIELD:
ch3cast(expp, '=', tp->tp_up);
expr = *expp;
#ifdef DEBUG
print_expr("init-expr after cast", expr);
#endif /* DEBUG */
if (is_cp_cst(expr))
Add long long literals like 123LL to ACK C. For now, a long long literal must have the 'LL' or 'll' suffix. A literal without 'LL' or 'll' acts as before: it may become unsigned long but not long long. (For targets where int and long have the same size, some literals change from unsigned int to unsigned long.) Type `arith` may be too narrow for long long values. Add a second type `writh` for wide arithmetic, and change some variables from arith to writh. This may cause bugs if I forget to use writh, or if a conversion from writh to arith overflows. I mark some conversions with (arith) or (writh) casts. - BigPars, SmallPars: Remove SPECIAL_ARITHMETICS. This feature would change arith to a different type, but can't work, because it would conflict with definitions of arith in both <em_arith.h> and <flt_arith.h>. - LLlex.c: Understand 'LL' or 'll' suffix. Cut size of constant when it overflows writh, not only when it overflows the target machine's types. (This cut might not be necessary, because we might cut it again later.) When picking signed long or unsigned long, check the target's long type, not the compiler's arith type; the old check for `val >= 0` was broken where sizeof(arith) > 4. - LLlex.h: Change struct token's tok_ival to writh, so it can hold a long long literal. - arith.c: Adjust to VL_VALUE being writh. Don't convert between float and integer at compile-time if the integer might be too wide for <flt_arith.h>. Add writh2str(), because writh might be too wide for long2str(). - arith.h: Remove SPECIAL_ARITHMETICS. Declare full_mask[] here, not in several *.c files. Declare writh2str(). - ch3.c, ch3bin.c, ch3mon.c, declarator.c, statement.g: Remove obsolete casts. Adjust to VL_VALUE being writh. - conversion.c, stab.c: Don't declare full_mask[]. - cstoper.c: Use writh for constant operations on VL_VALUE, and for full_mask[]. - declar., field.c, ival.g: Add casts. - dumpidf.c: Need to #include "parameters.h" before checking DEBUG. Use writh2str, because "%ld" might not work. - eval.c, eval.h: Add casts. Use writh when writing a wide constant in EM. - expr.c: Add and remove casts. In fill_int_expr(), make expression from long long literal. In chk_cst_expr(), allow long long as constant expression, so the compiler may accept `case 123LL:` in a switch statement. - expr.str: Change struct value's vl_value and struct expr's VL_VALUE to writh, so an expression may have a long long value at compile time. - statement.g: Remove obsolete casts. - switch.c, switch.str: Use writh in case entries for switch statements, so `switch (ll) {...}` with long long ll works. - tokenname.c: Add ULNGLNG so LLlex.c can use it for literals.
2019-09-05 02:14:38 +00:00
put_bf(tp, (arith)expr->VL_VALUE);
else
illegal_init_cst(expr);
break;
#endif /* NOBITFIELD */
case ERRONEOUS:
if (! gen_error) gen_error = pack_level;
/* fall through */
case VOID:
break;
default:
crash("check_ival");
/*NOTREACHED*/
}
}
/* ch_array() initialises an array of characters when given
a string constant.
Alignment is taken care of.
*/
void ch_array(struct type **tpp, /* type tp = array of characters */
struct expr *ex)
{
register struct type *tp = *tpp;
register int length = ex->SG_LEN, i;
register char *to, *from, *s;
assert(ex->ex_class == String);
if (tp->tp_size == (arith)-1) {
/* set the dimension */
tp = *tpp = construct_type(ARRAY, tp->tp_up, 0, (arith)length, NO_PROTO);
}
else {
arith dim = tp->tp_size / tp->tp_up->tp_size;
#ifdef LINT
if (length == dim + 1) {
expr_warning(ex, "array is not null-terminated");
} else
#endif
if (length > dim + 1) {
expr_strict(ex, "too many initializers");
}
length = dim;
}
/* throw out the characters of the already prepared string */
s = Malloc((unsigned) (length));
clear(s, (unsigned)length);
i = length <= ex->SG_LEN ? length : ex->SG_LEN;
to = s; from = ex->SG_VALUE;
while(--i >= 0) {
*to++ = *from++;
}
free(ex->SG_VALUE);
str_cst(s, length, 0); /* a string, but not in rom */
free(s);
}
/* As long as some parts of the pipeline cannot handle very long string
constants, string constants are written out in chunks
*/
void str_cst(register char *str, register int len, int inrom)
{
int chunksize = ((127 + (int) word_size) / (int) word_size) * (int) word_size;
while (len > chunksize) {
if (inrom)
C_rom_scon(str, (arith) chunksize);
else C_con_scon(str, (arith) chunksize);
len -= chunksize;
str += chunksize;
}
if (inrom)
C_rom_scon(str, (arith) len);
else C_con_scon(str, (arith) len);
}
#ifndef NOBITFIELD
/* put_bf() takes care of the initialisation of (bit-)field
selectors of a struct: each time such an initialisation takes place,
put_bf() is called instead of the normal code generating routines.
Put_bf() stores the given integral value into "field" and
"throws" the result of "field" out if the current selector
is the last of this number of fields stored at the same address.
*/
void put_bf(struct type *tp, arith val)
{
static long field = (arith)0;
static arith offset = (arith)-1;
register struct field *fd = tp->tp_field;
register struct sdef *sd = fd->fd_sdef;
static struct expr exp;
assert(sd);
if (offset == (arith)-1) {
/* first bitfield in this field */
offset = sd->sd_offset;
exp.ex_type = tp->tp_up;
exp.ex_class = Value;
exp.VL_CLASS = Const;
}
if (val != 0) /* insert the value into "field" */
field |= (val & fd->fd_mask) << fd->fd_shift;
if (sd->sd_sdef == 0 || sd->sd_sdef->sd_offset != offset) {
/* the selector was the last stored at this address */
Add long long literals like 123LL to ACK C. For now, a long long literal must have the 'LL' or 'll' suffix. A literal without 'LL' or 'll' acts as before: it may become unsigned long but not long long. (For targets where int and long have the same size, some literals change from unsigned int to unsigned long.) Type `arith` may be too narrow for long long values. Add a second type `writh` for wide arithmetic, and change some variables from arith to writh. This may cause bugs if I forget to use writh, or if a conversion from writh to arith overflows. I mark some conversions with (arith) or (writh) casts. - BigPars, SmallPars: Remove SPECIAL_ARITHMETICS. This feature would change arith to a different type, but can't work, because it would conflict with definitions of arith in both <em_arith.h> and <flt_arith.h>. - LLlex.c: Understand 'LL' or 'll' suffix. Cut size of constant when it overflows writh, not only when it overflows the target machine's types. (This cut might not be necessary, because we might cut it again later.) When picking signed long or unsigned long, check the target's long type, not the compiler's arith type; the old check for `val >= 0` was broken where sizeof(arith) > 4. - LLlex.h: Change struct token's tok_ival to writh, so it can hold a long long literal. - arith.c: Adjust to VL_VALUE being writh. Don't convert between float and integer at compile-time if the integer might be too wide for <flt_arith.h>. Add writh2str(), because writh might be too wide for long2str(). - arith.h: Remove SPECIAL_ARITHMETICS. Declare full_mask[] here, not in several *.c files. Declare writh2str(). - ch3.c, ch3bin.c, ch3mon.c, declarator.c, statement.g: Remove obsolete casts. Adjust to VL_VALUE being writh. - conversion.c, stab.c: Don't declare full_mask[]. - cstoper.c: Use writh for constant operations on VL_VALUE, and for full_mask[]. - declar., field.c, ival.g: Add casts. - dumpidf.c: Need to #include "parameters.h" before checking DEBUG. Use writh2str, because "%ld" might not work. - eval.c, eval.h: Add casts. Use writh when writing a wide constant in EM. - expr.c: Add and remove casts. In fill_int_expr(), make expression from long long literal. In chk_cst_expr(), allow long long as constant expression, so the compiler may accept `case 123LL:` in a switch statement. - expr.str: Change struct value's vl_value and struct expr's VL_VALUE to writh, so an expression may have a long long value at compile time. - statement.g: Remove obsolete casts. - switch.c, switch.str: Use writh in case entries for switch statements, so `switch (ll) {...}` with long long ll works. - tokenname.c: Add ULNGLNG so LLlex.c can use it for literals.
2019-09-05 02:14:38 +00:00
exp.VL_VALUE = (writh)field;
con_int(&exp);
field = (arith)0;
offset = (arith)-1;
}
}
#endif /* NOBITFIELD */
int zero_bytes(register struct sdef *sd)
{
/* fills the space between a selector of a struct
and the next selector of that struct with zero-bytes.
*/
register int n = sd->sd_sdef->sd_offset - sd->sd_offset -
size_of_type(sd->sd_type, "struct member");
int count = n;
while (n-- > 0)
con_nullbyte();
return count;
}
int valid_type(struct type *tp, char *str)
{
assert(tp!=(struct type *)0);
if (tp->tp_size < 0) {
error("size of %s unknown", str);
return 0;
}
return 1;
}
void con_int(register struct expr *ex)
{
register struct type *tp = ex->ex_type;
assert(is_cp_cst(ex));
if (tp->tp_unsigned)
Add long long literals like 123LL to ACK C. For now, a long long literal must have the 'LL' or 'll' suffix. A literal without 'LL' or 'll' acts as before: it may become unsigned long but not long long. (For targets where int and long have the same size, some literals change from unsigned int to unsigned long.) Type `arith` may be too narrow for long long values. Add a second type `writh` for wide arithmetic, and change some variables from arith to writh. This may cause bugs if I forget to use writh, or if a conversion from writh to arith overflows. I mark some conversions with (arith) or (writh) casts. - BigPars, SmallPars: Remove SPECIAL_ARITHMETICS. This feature would change arith to a different type, but can't work, because it would conflict with definitions of arith in both <em_arith.h> and <flt_arith.h>. - LLlex.c: Understand 'LL' or 'll' suffix. Cut size of constant when it overflows writh, not only when it overflows the target machine's types. (This cut might not be necessary, because we might cut it again later.) When picking signed long or unsigned long, check the target's long type, not the compiler's arith type; the old check for `val >= 0` was broken where sizeof(arith) > 4. - LLlex.h: Change struct token's tok_ival to writh, so it can hold a long long literal. - arith.c: Adjust to VL_VALUE being writh. Don't convert between float and integer at compile-time if the integer might be too wide for <flt_arith.h>. Add writh2str(), because writh might be too wide for long2str(). - arith.h: Remove SPECIAL_ARITHMETICS. Declare full_mask[] here, not in several *.c files. Declare writh2str(). - ch3.c, ch3bin.c, ch3mon.c, declarator.c, statement.g: Remove obsolete casts. Adjust to VL_VALUE being writh. - conversion.c, stab.c: Don't declare full_mask[]. - cstoper.c: Use writh for constant operations on VL_VALUE, and for full_mask[]. - declar., field.c, ival.g: Add casts. - dumpidf.c: Need to #include "parameters.h" before checking DEBUG. Use writh2str, because "%ld" might not work. - eval.c, eval.h: Add casts. Use writh when writing a wide constant in EM. - expr.c: Add and remove casts. In fill_int_expr(), make expression from long long literal. In chk_cst_expr(), allow long long as constant expression, so the compiler may accept `case 123LL:` in a switch statement. - expr.str: Change struct value's vl_value and struct expr's VL_VALUE to writh, so an expression may have a long long value at compile time. - statement.g: Remove obsolete casts. - switch.c, switch.str: Use writh in case entries for switch statements, so `switch (ll) {...}` with long long ll works. - tokenname.c: Add ULNGLNG so LLlex.c can use it for literals.
2019-09-05 02:14:38 +00:00
C_con_ucon(writh2str(ex->VL_VALUE, 1), tp->tp_size);
else if (tp->tp_size == word_size)
Add long long literals like 123LL to ACK C. For now, a long long literal must have the 'LL' or 'll' suffix. A literal without 'LL' or 'll' acts as before: it may become unsigned long but not long long. (For targets where int and long have the same size, some literals change from unsigned int to unsigned long.) Type `arith` may be too narrow for long long values. Add a second type `writh` for wide arithmetic, and change some variables from arith to writh. This may cause bugs if I forget to use writh, or if a conversion from writh to arith overflows. I mark some conversions with (arith) or (writh) casts. - BigPars, SmallPars: Remove SPECIAL_ARITHMETICS. This feature would change arith to a different type, but can't work, because it would conflict with definitions of arith in both <em_arith.h> and <flt_arith.h>. - LLlex.c: Understand 'LL' or 'll' suffix. Cut size of constant when it overflows writh, not only when it overflows the target machine's types. (This cut might not be necessary, because we might cut it again later.) When picking signed long or unsigned long, check the target's long type, not the compiler's arith type; the old check for `val >= 0` was broken where sizeof(arith) > 4. - LLlex.h: Change struct token's tok_ival to writh, so it can hold a long long literal. - arith.c: Adjust to VL_VALUE being writh. Don't convert between float and integer at compile-time if the integer might be too wide for <flt_arith.h>. Add writh2str(), because writh might be too wide for long2str(). - arith.h: Remove SPECIAL_ARITHMETICS. Declare full_mask[] here, not in several *.c files. Declare writh2str(). - ch3.c, ch3bin.c, ch3mon.c, declarator.c, statement.g: Remove obsolete casts. Adjust to VL_VALUE being writh. - conversion.c, stab.c: Don't declare full_mask[]. - cstoper.c: Use writh for constant operations on VL_VALUE, and for full_mask[]. - declar., field.c, ival.g: Add casts. - dumpidf.c: Need to #include "parameters.h" before checking DEBUG. Use writh2str, because "%ld" might not work. - eval.c, eval.h: Add casts. Use writh when writing a wide constant in EM. - expr.c: Add and remove casts. In fill_int_expr(), make expression from long long literal. In chk_cst_expr(), allow long long as constant expression, so the compiler may accept `case 123LL:` in a switch statement. - expr.str: Change struct value's vl_value and struct expr's VL_VALUE to writh, so an expression may have a long long value at compile time. - statement.g: Remove obsolete casts. - switch.c, switch.str: Use writh in case entries for switch statements, so `switch (ll) {...}` with long long ll works. - tokenname.c: Add ULNGLNG so LLlex.c can use it for literals.
2019-09-05 02:14:38 +00:00
C_con_cst((arith)ex->VL_VALUE);
else
Add long long literals like 123LL to ACK C. For now, a long long literal must have the 'LL' or 'll' suffix. A literal without 'LL' or 'll' acts as before: it may become unsigned long but not long long. (For targets where int and long have the same size, some literals change from unsigned int to unsigned long.) Type `arith` may be too narrow for long long values. Add a second type `writh` for wide arithmetic, and change some variables from arith to writh. This may cause bugs if I forget to use writh, or if a conversion from writh to arith overflows. I mark some conversions with (arith) or (writh) casts. - BigPars, SmallPars: Remove SPECIAL_ARITHMETICS. This feature would change arith to a different type, but can't work, because it would conflict with definitions of arith in both <em_arith.h> and <flt_arith.h>. - LLlex.c: Understand 'LL' or 'll' suffix. Cut size of constant when it overflows writh, not only when it overflows the target machine's types. (This cut might not be necessary, because we might cut it again later.) When picking signed long or unsigned long, check the target's long type, not the compiler's arith type; the old check for `val >= 0` was broken where sizeof(arith) > 4. - LLlex.h: Change struct token's tok_ival to writh, so it can hold a long long literal. - arith.c: Adjust to VL_VALUE being writh. Don't convert between float and integer at compile-time if the integer might be too wide for <flt_arith.h>. Add writh2str(), because writh might be too wide for long2str(). - arith.h: Remove SPECIAL_ARITHMETICS. Declare full_mask[] here, not in several *.c files. Declare writh2str(). - ch3.c, ch3bin.c, ch3mon.c, declarator.c, statement.g: Remove obsolete casts. Adjust to VL_VALUE being writh. - conversion.c, stab.c: Don't declare full_mask[]. - cstoper.c: Use writh for constant operations on VL_VALUE, and for full_mask[]. - declar., field.c, ival.g: Add casts. - dumpidf.c: Need to #include "parameters.h" before checking DEBUG. Use writh2str, because "%ld" might not work. - eval.c, eval.h: Add casts. Use writh when writing a wide constant in EM. - expr.c: Add and remove casts. In fill_int_expr(), make expression from long long literal. In chk_cst_expr(), allow long long as constant expression, so the compiler may accept `case 123LL:` in a switch statement. - expr.str: Change struct value's vl_value and struct expr's VL_VALUE to writh, so an expression may have a long long value at compile time. - statement.g: Remove obsolete casts. - switch.c, switch.str: Use writh in case entries for switch statements, so `switch (ll) {...}` with long long ll works. - tokenname.c: Add ULNGLNG so LLlex.c can use it for literals.
2019-09-05 02:14:38 +00:00
C_con_icon(writh2str(ex->VL_VALUE, 0), tp->tp_size);
}
void illegal_init_cst(struct expr *ex)
{
expr_error(ex, "illegal initialization constant");
gen_error = pack_level;
}
void too_many_initialisers(void)
{
error("too many initializers");
gen_error = pack_level;
}
}