mirror of
https://sourceware.org/git/binutils-gdb.git
synced 2025-02-11 13:02:10 +08:00
This commit splits the `set/show print elements' option into two. We retain `set/show print elements' for controlling how many elements of an array we print, but a new `set/show print characters' setting is added which is used for controlling how many characters of a string are printed. The motivation behind this change is to allow users a finer level of control over how data is printed, reflecting that, although strings can be thought of as arrays of characters, users often want to treat these two things differently. For compatibility reasons by default the `set/show print characters' option is set to `elements', which makes the limit for character strings follow the setting of the `set/show print elements' option, as it used to. Using `set print characters' with any other value makes the limit independent from the `set/show print elements' setting, however it can be restored to the default with the `set print characters elements' command at any time. A corresponding `-characters' option for the `print' command is added, with the same semantics, i.e. one can use `elements' to make a given `print' invocation follow the limit of elements, be it set with the `-elements' option also given with the same invocation or taken from the `set/show print elements' setting, for characters as well regardless of the current setting of the `set/show print characters' option. The GDB changes are all pretty straightforward, just changing references to the old 'print_max' to use a new `get_print_max_chars' helper which figures out which of the two of `print_max' and `print_max_chars' values to use. Likewise, the documentation is just updated to reference the new setting where appropriate. To make people's life easier the message shown by `show print elements' now indicates if the setting also applies to character strings: (gdb) set print characters elements (gdb) show print elements Limit on string chars or array elements to print is 200. (gdb) set print characters unlimited (gdb) show print elements Limit on array elements to print is 200. (gdb) and the help text shows the dependency as well: (gdb) help set print elements Set limit on array elements to print. "unlimited" causes there to be no limit. This setting also applies to string chars when "print characters" is set to "elements". (gdb) In the testsuite there are two minor updates, one to add `-characters' to the list of completions now shown for the `print' command, and a bare minimum pair of checks for the right handling of `set print characters' and `show print characters', copied from the corresponding checks for `set print elements' and `show print elements' respectively. Co-Authored-By: Maciej W. Rozycki <macro@embecosm.com> Approved-By: Simon Marchi <simon.marchi@efficios.com>
591 lines
17 KiB
C
591 lines
17 KiB
C
/* Support for printing C values for GDB, the GNU debugger.
|
||
|
||
Copyright (C) 1986-2023 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/>. */
|
||
|
||
#include "defs.h"
|
||
#include "symtab.h"
|
||
#include "gdbtypes.h"
|
||
#include "expression.h"
|
||
#include "value.h"
|
||
#include "valprint.h"
|
||
#include "language.h"
|
||
#include "c-lang.h"
|
||
#include "cp-abi.h"
|
||
#include "target.h"
|
||
#include "objfiles.h"
|
||
|
||
|
||
/* A helper for c_textual_element_type. This checks the name of the
|
||
typedef. This is bogus but it isn't apparent that the compiler
|
||
provides us the help we may need. */
|
||
|
||
static int
|
||
textual_name (const char *name)
|
||
{
|
||
return (!strcmp (name, "wchar_t")
|
||
|| !strcmp (name, "char16_t")
|
||
|| !strcmp (name, "char32_t"));
|
||
}
|
||
|
||
/* Apply a heuristic to decide whether an array of TYPE or a pointer
|
||
to TYPE should be printed as a textual string. Return non-zero if
|
||
it should, or zero if it should be treated as an array of integers
|
||
or pointer to integers. FORMAT is the current format letter, or 0
|
||
if none.
|
||
|
||
We guess that "char" is a character. Explicitly signed and
|
||
unsigned character types are also characters. Integer data from
|
||
vector types is not. The user can override this by using the /s
|
||
format letter. */
|
||
|
||
int
|
||
c_textual_element_type (struct type *type, char format)
|
||
{
|
||
struct type *true_type, *iter_type;
|
||
|
||
if (format != 0 && format != 's')
|
||
return 0;
|
||
|
||
/* We also rely on this for its side effect of setting up all the
|
||
typedef pointers. */
|
||
true_type = check_typedef (type);
|
||
|
||
/* TYPE_CODE_CHAR is always textual. */
|
||
if (true_type->code () == TYPE_CODE_CHAR)
|
||
return 1;
|
||
|
||
/* Any other character-like types must be integral. */
|
||
if (true_type->code () != TYPE_CODE_INT)
|
||
return 0;
|
||
|
||
/* We peel typedefs one by one, looking for a match. */
|
||
iter_type = type;
|
||
while (iter_type)
|
||
{
|
||
/* Check the name of the type. */
|
||
if (iter_type->name () && textual_name (iter_type->name ()))
|
||
return 1;
|
||
|
||
if (iter_type->code () != TYPE_CODE_TYPEDEF)
|
||
break;
|
||
|
||
/* Peel a single typedef. If the typedef doesn't have a target
|
||
type, we use check_typedef and hope the result is ok -- it
|
||
might be for C++, where wchar_t is a built-in type. */
|
||
if (iter_type->target_type ())
|
||
iter_type = iter_type->target_type ();
|
||
else
|
||
iter_type = check_typedef (iter_type);
|
||
}
|
||
|
||
if (format == 's')
|
||
{
|
||
/* Print this as a string if we can manage it. For now, no wide
|
||
character support. */
|
||
if (true_type->code () == TYPE_CODE_INT
|
||
&& true_type->length () == 1)
|
||
return 1;
|
||
}
|
||
else
|
||
{
|
||
/* If a one-byte TYPE_CODE_INT is missing the not-a-character
|
||
flag, then we treat it as text; otherwise, we assume it's
|
||
being used as data. */
|
||
if (true_type->code () == TYPE_CODE_INT
|
||
&& true_type->length () == 1
|
||
&& !TYPE_NOTTEXT (true_type))
|
||
return 1;
|
||
}
|
||
|
||
return 0;
|
||
}
|
||
|
||
/* Decorations for C. */
|
||
|
||
static const struct generic_val_print_decorations c_decorations =
|
||
{
|
||
"",
|
||
" + ",
|
||
"i",
|
||
"true",
|
||
"false",
|
||
"void",
|
||
"{",
|
||
"}"
|
||
};
|
||
|
||
/* Print a pointer based on the type of its target.
|
||
|
||
Arguments to this functions are roughly the same as those in c_val_print.
|
||
A difference is that ADDRESS is the address to print, with embedded_offset
|
||
already added. UNRESOLVED_ELTTYPE and ELTTYPE represent the pointed type,
|
||
respectively before and after check_typedef. */
|
||
|
||
static void
|
||
print_unpacked_pointer (struct type *type, struct type *elttype,
|
||
struct type *unresolved_elttype,
|
||
const gdb_byte *valaddr, int embedded_offset,
|
||
CORE_ADDR address, struct ui_file *stream, int recurse,
|
||
const struct value_print_options *options)
|
||
{
|
||
int want_space = 0;
|
||
struct gdbarch *gdbarch = type->arch ();
|
||
|
||
if (elttype->code () == TYPE_CODE_FUNC)
|
||
{
|
||
/* Try to print what function it points to. */
|
||
print_function_pointer_address (options, gdbarch, address, stream);
|
||
return;
|
||
}
|
||
|
||
if (options->symbol_print)
|
||
want_space = print_address_demangle (options, gdbarch, address, stream,
|
||
demangle);
|
||
else if (options->addressprint)
|
||
{
|
||
gdb_puts (paddress (gdbarch, address), stream);
|
||
want_space = 1;
|
||
}
|
||
|
||
/* For a pointer to a textual type, also print the string
|
||
pointed to, unless pointer is null. */
|
||
|
||
if (c_textual_element_type (unresolved_elttype, options->format)
|
||
&& address != 0)
|
||
{
|
||
if (want_space)
|
||
gdb_puts (" ", stream);
|
||
val_print_string (unresolved_elttype, NULL, address, -1, stream, options);
|
||
}
|
||
else if (cp_is_vtbl_member (type))
|
||
{
|
||
/* Print vtbl's nicely. */
|
||
CORE_ADDR vt_address = unpack_pointer (type, valaddr + embedded_offset);
|
||
struct bound_minimal_symbol msymbol =
|
||
lookup_minimal_symbol_by_pc (vt_address);
|
||
|
||
/* If 'symbol_print' is set, we did the work above. */
|
||
if (!options->symbol_print
|
||
&& (msymbol.minsym != NULL)
|
||
&& (vt_address == msymbol.value_address ()))
|
||
{
|
||
if (want_space)
|
||
gdb_puts (" ", stream);
|
||
gdb_puts (" <", stream);
|
||
gdb_puts (msymbol.minsym->print_name (), stream);
|
||
gdb_puts (">", stream);
|
||
want_space = 1;
|
||
}
|
||
|
||
if (vt_address && options->vtblprint)
|
||
{
|
||
struct value *vt_val;
|
||
struct symbol *wsym = NULL;
|
||
struct type *wtype;
|
||
|
||
if (want_space)
|
||
gdb_puts (" ", stream);
|
||
|
||
if (msymbol.minsym != NULL)
|
||
{
|
||
const char *search_name = msymbol.minsym->search_name ();
|
||
wsym = lookup_symbol_search_name (search_name, NULL,
|
||
VAR_DOMAIN).symbol;
|
||
}
|
||
|
||
if (wsym)
|
||
{
|
||
wtype = wsym->type ();
|
||
}
|
||
else
|
||
{
|
||
wtype = unresolved_elttype;
|
||
}
|
||
vt_val = value_at (wtype, vt_address);
|
||
common_val_print (vt_val, stream, recurse + 1, options,
|
||
current_language);
|
||
if (options->prettyformat)
|
||
{
|
||
gdb_printf (stream, "\n");
|
||
print_spaces (2 + 2 * recurse, stream);
|
||
}
|
||
}
|
||
}
|
||
}
|
||
|
||
/* c_value_print helper for TYPE_CODE_ARRAY. */
|
||
|
||
static void
|
||
c_value_print_array (struct value *val,
|
||
struct ui_file *stream, int recurse,
|
||
const struct value_print_options *options)
|
||
{
|
||
struct type *type = check_typedef (value_type (val));
|
||
CORE_ADDR address = value_address (val);
|
||
const gdb_byte *valaddr = value_contents_for_printing (val).data ();
|
||
struct type *unresolved_elttype = type->target_type ();
|
||
struct type *elttype = check_typedef (unresolved_elttype);
|
||
|
||
if (type->length () > 0 && unresolved_elttype->length () > 0)
|
||
{
|
||
LONGEST low_bound, high_bound;
|
||
int eltlen, len;
|
||
enum bfd_endian byte_order = type_byte_order (type);
|
||
|
||
if (!get_array_bounds (type, &low_bound, &high_bound))
|
||
error (_("Could not determine the array high bound"));
|
||
|
||
eltlen = elttype->length ();
|
||
len = high_bound - low_bound + 1;
|
||
|
||
/* Print arrays of textual chars with a string syntax, as
|
||
long as the entire array is valid. */
|
||
if (c_textual_element_type (unresolved_elttype,
|
||
options->format)
|
||
&& value_bytes_available (val, 0, type->length ())
|
||
&& !value_bits_any_optimized_out (val, 0,
|
||
TARGET_CHAR_BIT * type->length ()))
|
||
{
|
||
int force_ellipses = 0;
|
||
|
||
/* If requested, look for the first null char and only
|
||
print elements up to it. */
|
||
if (options->stop_print_at_null)
|
||
{
|
||
unsigned int print_max_chars = get_print_max_chars (options);
|
||
unsigned int temp_len;
|
||
|
||
for (temp_len = 0;
|
||
(temp_len < len
|
||
&& temp_len < print_max_chars
|
||
&& extract_unsigned_integer (valaddr + temp_len * eltlen,
|
||
eltlen, byte_order) != 0);
|
||
++temp_len)
|
||
;
|
||
|
||
/* Force printstr to print ellipses if
|
||
we've printed the maximum characters and
|
||
the next character is not \000. */
|
||
if (temp_len == print_max_chars && temp_len < len)
|
||
{
|
||
ULONGEST ival
|
||
= extract_unsigned_integer (valaddr + temp_len * eltlen,
|
||
eltlen, byte_order);
|
||
if (ival != 0)
|
||
force_ellipses = 1;
|
||
}
|
||
|
||
len = temp_len;
|
||
}
|
||
|
||
current_language->printstr (stream, unresolved_elttype, valaddr, len,
|
||
NULL, force_ellipses, options);
|
||
}
|
||
else
|
||
{
|
||
unsigned int i = 0;
|
||
gdb_printf (stream, "{");
|
||
/* If this is a virtual function table, print the 0th
|
||
entry specially, and the rest of the members
|
||
normally. */
|
||
if (cp_is_vtbl_ptr_type (elttype))
|
||
{
|
||
i = 1;
|
||
gdb_printf (stream, _("%d vtable entries"),
|
||
len - 1);
|
||
}
|
||
value_print_array_elements (val, stream, recurse, options, i);
|
||
gdb_printf (stream, "}");
|
||
}
|
||
}
|
||
else
|
||
{
|
||
/* Array of unspecified length: treat like pointer to first elt. */
|
||
print_unpacked_pointer (type, elttype, unresolved_elttype, valaddr,
|
||
0, address, stream, recurse, options);
|
||
}
|
||
}
|
||
|
||
/* c_value_print_inner helper for TYPE_CODE_PTR. */
|
||
|
||
static void
|
||
c_value_print_ptr (struct value *val, struct ui_file *stream, int recurse,
|
||
const struct value_print_options *options)
|
||
{
|
||
if (options->format && options->format != 's')
|
||
{
|
||
value_print_scalar_formatted (val, options, 0, stream);
|
||
return;
|
||
}
|
||
|
||
struct type *type = check_typedef (value_type (val));
|
||
const gdb_byte *valaddr = value_contents_for_printing (val).data ();
|
||
|
||
if (options->vtblprint && cp_is_vtbl_ptr_type (type))
|
||
{
|
||
/* Print the unmangled name if desired. */
|
||
/* Print vtable entry - we only get here if we ARE using
|
||
-fvtable_thunks. (Otherwise, look under
|
||
TYPE_CODE_STRUCT.) */
|
||
CORE_ADDR addr = extract_typed_address (valaddr, type);
|
||
|
||
print_function_pointer_address (options, type->arch (), addr, stream);
|
||
}
|
||
else
|
||
{
|
||
struct type *unresolved_elttype = type->target_type ();
|
||
struct type *elttype = check_typedef (unresolved_elttype);
|
||
CORE_ADDR addr = unpack_pointer (type, valaddr);
|
||
|
||
print_unpacked_pointer (type, elttype, unresolved_elttype, valaddr,
|
||
0, addr, stream, recurse, options);
|
||
}
|
||
}
|
||
|
||
/* c_value_print helper for TYPE_CODE_STRUCT and TYPE_CODE_UNION. */
|
||
|
||
static void
|
||
c_value_print_struct (struct value *val, struct ui_file *stream, int recurse,
|
||
const struct value_print_options *options)
|
||
{
|
||
struct type *type = check_typedef (value_type (val));
|
||
|
||
if (type->code () == TYPE_CODE_UNION && recurse && !options->unionprint)
|
||
gdb_printf (stream, "{...}");
|
||
else if (options->vtblprint && cp_is_vtbl_ptr_type (type))
|
||
{
|
||
/* Print the unmangled name if desired. */
|
||
/* Print vtable entry - we only get here if NOT using
|
||
-fvtable_thunks. (Otherwise, look under
|
||
TYPE_CODE_PTR.) */
|
||
int offset = type->field (VTBL_FNADDR_OFFSET).loc_bitpos () / 8;
|
||
struct type *field_type = type->field (VTBL_FNADDR_OFFSET).type ();
|
||
const gdb_byte *valaddr = value_contents_for_printing (val).data ();
|
||
CORE_ADDR addr = extract_typed_address (valaddr + offset, field_type);
|
||
|
||
print_function_pointer_address (options, type->arch (), addr, stream);
|
||
}
|
||
else
|
||
cp_print_value_fields (val, stream, recurse, options, NULL, 0);
|
||
}
|
||
|
||
/* c_value_print helper for TYPE_CODE_INT. */
|
||
|
||
static void
|
||
c_value_print_int (struct value *val, struct ui_file *stream,
|
||
const struct value_print_options *options)
|
||
{
|
||
if (options->format || options->output_format)
|
||
{
|
||
struct value_print_options opts = *options;
|
||
|
||
opts.format = (options->format ? options->format
|
||
: options->output_format);
|
||
value_print_scalar_formatted (val, &opts, 0, stream);
|
||
}
|
||
else
|
||
{
|
||
value_print_scalar_formatted (val, options, 0, stream);
|
||
/* C and C++ has no single byte int type, char is used
|
||
instead. Since we don't know whether the value is really
|
||
intended to be used as an integer or a character, print
|
||
the character equivalent as well. */
|
||
struct type *type = value_type (val);
|
||
const gdb_byte *valaddr = value_contents_for_printing (val).data ();
|
||
if (c_textual_element_type (type, options->format))
|
||
{
|
||
gdb_puts (" ", stream);
|
||
current_language->printchar (unpack_long (type, valaddr), type,
|
||
stream);
|
||
}
|
||
}
|
||
}
|
||
|
||
/* See c-lang.h. */
|
||
|
||
void
|
||
c_value_print_inner (struct value *val, struct ui_file *stream, int recurse,
|
||
const struct value_print_options *options)
|
||
{
|
||
struct type *type = value_type (val);
|
||
|
||
type = check_typedef (type);
|
||
switch (type->code ())
|
||
{
|
||
case TYPE_CODE_ARRAY:
|
||
c_value_print_array (val, stream, recurse, options);
|
||
break;
|
||
|
||
case TYPE_CODE_PTR:
|
||
c_value_print_ptr (val, stream, recurse, options);
|
||
break;
|
||
|
||
case TYPE_CODE_UNION:
|
||
case TYPE_CODE_STRUCT:
|
||
c_value_print_struct (val, stream, recurse, options);
|
||
break;
|
||
|
||
case TYPE_CODE_CHAR:
|
||
case TYPE_CODE_INT:
|
||
c_value_print_int (val, stream, options);
|
||
break;
|
||
|
||
case TYPE_CODE_METHODPTR:
|
||
case TYPE_CODE_MEMBERPTR:
|
||
case TYPE_CODE_REF:
|
||
case TYPE_CODE_RVALUE_REF:
|
||
case TYPE_CODE_ENUM:
|
||
case TYPE_CODE_FLAGS:
|
||
case TYPE_CODE_FUNC:
|
||
case TYPE_CODE_METHOD:
|
||
case TYPE_CODE_BOOL:
|
||
case TYPE_CODE_RANGE:
|
||
case TYPE_CODE_FLT:
|
||
case TYPE_CODE_DECFLOAT:
|
||
case TYPE_CODE_VOID:
|
||
case TYPE_CODE_ERROR:
|
||
case TYPE_CODE_UNDEF:
|
||
case TYPE_CODE_COMPLEX:
|
||
default:
|
||
generic_value_print (val, stream, recurse, options, &c_decorations);
|
||
break;
|
||
}
|
||
}
|
||
|
||
|
||
void
|
||
c_value_print (struct value *val, struct ui_file *stream,
|
||
const struct value_print_options *options)
|
||
{
|
||
struct type *type, *real_type;
|
||
int full, using_enc;
|
||
LONGEST top;
|
||
struct value_print_options opts = *options;
|
||
|
||
opts.deref_ref = true;
|
||
|
||
/* If it is a pointer, indicate what it points to.
|
||
|
||
Print type also if it is a reference.
|
||
|
||
C++: if it is a member pointer, we will take care
|
||
of that when we print it. */
|
||
|
||
type = check_typedef (value_type (val));
|
||
|
||
if (type->is_pointer_or_reference ())
|
||
{
|
||
struct type *original_type = value_type (val);
|
||
|
||
/* Hack: remove (char *) for char strings. Their
|
||
type is indicated by the quoted string anyway.
|
||
(Don't use c_textual_element_type here; quoted strings
|
||
are always exactly (char *), (wchar_t *), or the like. */
|
||
if (original_type->code () == TYPE_CODE_PTR
|
||
&& original_type->name () == NULL
|
||
&& original_type->target_type ()->name () != NULL
|
||
&& (strcmp (original_type->target_type ()->name (),
|
||
"char") == 0
|
||
|| textual_name (original_type->target_type ()->name ())))
|
||
{
|
||
/* Print nothing. */
|
||
}
|
||
else if (options->objectprint
|
||
&& (type->target_type ()->code () == TYPE_CODE_STRUCT))
|
||
{
|
||
int is_ref = TYPE_IS_REFERENCE (type);
|
||
enum type_code refcode = TYPE_CODE_UNDEF;
|
||
|
||
if (is_ref)
|
||
{
|
||
val = value_addr (val);
|
||
refcode = type->code ();
|
||
}
|
||
|
||
/* Pointer to class, check real type of object. */
|
||
gdb_printf (stream, "(");
|
||
|
||
if (value_entirely_available (val))
|
||
{
|
||
real_type = value_rtti_indirect_type (val, &full, &top,
|
||
&using_enc);
|
||
if (real_type)
|
||
{
|
||
/* RTTI entry found. */
|
||
|
||
/* Need to adjust pointer value. */
|
||
val = value_from_pointer (real_type,
|
||
value_as_address (val) - top);
|
||
|
||
/* Note: When we look up RTTI entries, we don't get
|
||
any information on const or volatile
|
||
attributes. */
|
||
}
|
||
}
|
||
|
||
if (is_ref)
|
||
val = value_ref (value_ind (val), refcode);
|
||
|
||
type = value_type (val);
|
||
type_print (type, "", stream, -1);
|
||
gdb_printf (stream, ") ");
|
||
}
|
||
else
|
||
{
|
||
/* normal case */
|
||
gdb_printf (stream, "(");
|
||
type_print (value_type (val), "", stream, -1);
|
||
gdb_printf (stream, ") ");
|
||
}
|
||
}
|
||
|
||
if (!value_initialized (val))
|
||
gdb_printf (stream, " [uninitialized] ");
|
||
|
||
if (options->objectprint && (type->code () == TYPE_CODE_STRUCT))
|
||
{
|
||
/* Attempt to determine real type of object. */
|
||
real_type = value_rtti_type (val, &full, &top, &using_enc);
|
||
if (real_type)
|
||
{
|
||
/* We have RTTI information, so use it. */
|
||
val = value_full_object (val, real_type,
|
||
full, top, using_enc);
|
||
/* In a destructor we might see a real type that is a
|
||
superclass of the object's type. In this case it is
|
||
better to leave the object as-is. */
|
||
if (!(full
|
||
&& (real_type->length ()
|
||
< value_enclosing_type (val)->length ())))
|
||
val = value_cast (real_type, val);
|
||
gdb_printf (stream, "(%s%s) ",
|
||
real_type->name (),
|
||
full ? "" : _(" [incomplete object]"));
|
||
}
|
||
else if (type != check_typedef (value_enclosing_type (val)))
|
||
{
|
||
/* No RTTI information, so let's do our best. */
|
||
gdb_printf (stream, "(%s ?) ",
|
||
value_enclosing_type (val)->name ());
|
||
val = value_cast (value_enclosing_type (val), val);
|
||
}
|
||
}
|
||
|
||
common_val_print (val, stream, 0, &opts, current_language);
|
||
}
|