[PATCH 0/4] Remove some bfd_boolean uses

[PATCH 0/4] Remove some bfd_boolean uses

[Tom Tromey]

I saw some recent commits removing uses of bfd_boolean from the tree.
This series removes the remaining uses from gdb as well, and then
arranges to poison the identifier using the BFD-supplied approach.

Signed-off-by: Tom Tromey <tom@tromey.com>
---
Tom Tromey (4):
      Use bool, not bfd_boolean
      Don't use BFD TRUE/FALSE constants
      Drop use of bfd_boolean from findcmd.c
      Poison bfd_boolean

 gdb/arc-linux-tdep.c              |  6 +++---
 gdb/compile/compile-object-load.c |  4 ++--
 gdb/defs.h                        |  2 ++
 gdb/elfread.c                     |  2 +-
 gdb/findcmd.c                     | 33 +++++++++++++--------------------
 gdb/ft32-tdep.c                   |  6 +++---
 gdb/machoread.c                   |  4 ++--
 7 files changed, 26 insertions(+), 31 deletions(-)
---
base-commit: a00fddb175ff5cb9edef18beedfba38fc6fc6017
change-id: 20260307-no-bfd-boolean-1c8cbdbf2849

Best regards,
-- 
Tom Tromey <tom@tromey.com>

[PATCH 1/4] Use bool, not bfd_boolean

[Tom Tromey]

This changes a couple of spots to use bool rather than bfd_boolean.
bfd_boolean appears to be deprecated.
---
 gdb/compile/compile-object-load.c | 2 +-
 gdb/ft32-tdep.c                   | 6 +++---
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/gdb/compile/compile-object-load.c b/gdb/compile/compile-object-load.c
index 27cd0c7c2f3..e31be539741 100644
--- a/gdb/compile/compile-object-load.c
+++ b/gdb/compile/compile-object-load.c
@@ -220,7 +220,7 @@ link_callbacks_warning (struct bfd_link_info *link_info, const char *xwarning,
 static void
 link_callbacks_undefined_symbol (struct bfd_link_info *link_info,
 				 const char *name, bfd *abfd, asection *section,
-				 bfd_vma address, bfd_boolean is_fatal)
+				 bfd_vma address, bool is_fatal)
 {
   warning (_("Cannot resolve relocation to \"%s\" "
 	     "from compiled module \"%s\" section \"%s\"."),
diff --git a/gdb/ft32-tdep.c b/gdb/ft32-tdep.c
index 84b9ce657cb..5a1f3315c7f 100644
--- a/gdb/ft32-tdep.c
+++ b/gdb/ft32-tdep.c
@@ -61,7 +61,7 @@ struct ft32_frame_cache
   /* Saved SP in this frame */
   CORE_ADDR saved_sp;
   /* Has the new frame been LINKed.  */
-  bfd_boolean established;
+  bool established;
 };
 
 /* Implement the "frame_align" gdbarch method.  */
@@ -206,7 +206,7 @@ ft32_analyze_prologue (CORE_ADDR start_addr, CORE_ADDR end_addr,
   if (start_addr >= end_addr)
     return end_addr;
 
-  cache->established = 0;
+  cache->established = false;
   for (next_addr = start_addr; next_addr < end_addr; next_addr += isize)
     {
       inst = ft32_fetch_instruction (next_addr, &isize, byte_order);
@@ -251,7 +251,7 @@ ft32_analyze_prologue (CORE_ADDR start_addr, CORE_ADDR end_addr,
       inst = ft32_fetch_instruction (next_addr, &isize, byte_order);
       if (FT32_IS_LINK (inst))
 	{
-	  cache->established = 1;
+	  cache->established = true;
 	  for (regnum = FT32_R0_REGNUM; regnum < FT32_PC_REGNUM; regnum++)
 	    {
 	      if (cache->saved_regs[regnum] != REG_UNAVAIL)

-- 
2.49.0

[PATCH 2/4] Don't use BFD TRUE/FALSE constants

[Tom Tromey]

This patch replaces uses of the BFD TRUE/FALSE constants with the
corresponding C++ constants.

Note that some uses of TRUE and FALSE remain:

* Mach code (Darwin and Hurd).  I don't know if this is typical and
  anyway I can't build this code.
* Curses code
* Windows code
---
 gdb/arc-linux-tdep.c              | 6 +++---
 gdb/compile/compile-object-load.c | 2 +-
 gdb/elfread.c                     | 2 +-
 gdb/machoread.c                   | 4 ++--
 4 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/gdb/arc-linux-tdep.c b/gdb/arc-linux-tdep.c
index 207235d3732..ca3dbf5a5fa 100644
--- a/gdb/arc-linux-tdep.c
+++ b/gdb/arc-linux-tdep.c
@@ -212,18 +212,18 @@ arc_linux_is_sigtramp (const frame_info_ptr &this_frame)
   if (!safe_frame_unwind_memory (this_frame, pc, {buf, insns_sz}))
     {
       /* Failed to unwind frame.  */
-      return FALSE;
+      return false;
     }
 
   /* Is that code the sigtramp instruction sequence?  */
   if (memcmp (buf, arc_sigtramp_insns, insns_sz) == 0)
-    return TRUE;
+    return true;
 
   /* No - look one instruction earlier in the code...  */
   if (!safe_frame_unwind_memory (this_frame, pc - 4, {buf, insns_sz}))
     {
       /* Failed to unwind frame.  */
-      return FALSE;
+      return false;
     }
 
   return (memcmp (buf, arc_sigtramp_insns, insns_sz) == 0);
diff --git a/gdb/compile/compile-object-load.c b/gdb/compile/compile-object-load.c
index e31be539741..bbee2f7cc56 100644
--- a/gdb/compile/compile-object-load.c
+++ b/gdb/compile/compile-object-load.c
@@ -372,7 +372,7 @@ copy_section (bfd *abfd, asection *sect,
 
   sect_data_got = bfd_get_relocated_section_contents (abfd, &link_info,
 						      &link_order,
-						      sect_data.get (), FALSE,
+						      sect_data.get (), false,
 						      symbol_table.data ());
 
   if (sect_data_got == NULL)
diff --git a/gdb/elfread.c b/gdb/elfread.c
index 7e38f623d12..7a647396f68 100644
--- a/gdb/elfread.c
+++ b/gdb/elfread.c
@@ -601,7 +601,7 @@ elf_rel_plt_read (minimal_symbol_reader &reader,
   if (relplt == NULL)
     return;
 
-  if (! bed->s->slurp_reloc_table (obfd, relplt, dyn_symbol_table, TRUE))
+  if (! bed->s->slurp_reloc_table (obfd, relplt, dyn_symbol_table, true))
     return;
 
   std::string string_buffer;
diff --git a/gdb/machoread.c b/gdb/machoread.c
index 334e3b41b6a..416eb6766fd 100644
--- a/gdb/machoread.c
+++ b/gdb/machoread.c
@@ -483,7 +483,7 @@ macho_add_oso_symfile (oso_el *oso, const gdb_bfd_ref_ptr &abfd,
 	  struct macho_sym_hash_entry *ent;
 
 	  ent = (struct macho_sym_hash_entry *)
-	    bfd_hash_lookup (&table, sym->name, TRUE, FALSE);
+	    bfd_hash_lookup (&table, sym->name, true, false);
 	  if (ent->sym != NULL)
 	    complaint (_("Duplicated symbol %s in symbol table"), sym->name);
 	  else
@@ -548,7 +548,7 @@ macho_add_oso_symfile (oso_el *oso, const gdb_bfd_ref_ptr &abfd,
 		  struct macho_sym_hash_entry *ent;
 
 		  ent = (struct macho_sym_hash_entry *)
-		    bfd_hash_lookup (&table, sym->name, FALSE, FALSE);
+		    bfd_hash_lookup (&table, sym->name, false, false);
 		  if (ent != NULL)
 		    addr = bfd_asymbol_value (ent->sym);
 		}

-- 
2.49.0

[PATCH 3/4] Drop use of bfd_boolean from findcmd.c

[Tom Tromey]

This changes findcmd.c to remove the use of bfd_boolean.  Instead, the
byte order is used directly.  put_bits is also rewritten in terms of
store_unsigned_integer.
---
 gdb/findcmd.c | 33 +++++++++++++--------------------
 1 file changed, 13 insertions(+), 20 deletions(-)

diff --git a/gdb/findcmd.c b/gdb/findcmd.c
index c4ee0ec97e4..b52c44d48d0 100644
--- a/gdb/findcmd.c
+++ b/gdb/findcmd.c
@@ -24,27 +24,20 @@
 #include "cli/cli-utils.h"
 #include <algorithm>
 #include "gdbsupport/byte-vector.h"
+#include "extract-store-integer.h"
 
-/* Copied from bfd_put_bits.  */
+/* Append DATA to BUF, writing only the specified number of bits,
+   using the given endian-ness.  */
 
 static void
-put_bits (uint64_t data, gdb::byte_vector &buf, int bits, bfd_boolean big_p)
+put_bits (uint64_t data, gdb::byte_vector &buf, int bits,
+	  bfd_endian byte_order)
 {
-  int i;
-  int bytes;
-
   gdb_assert (bits % 8 == 0);
-
-  bytes = bits / 8;
+  int bytes = bits / 8;
   size_t last = buf.size ();
   buf.resize (last + bytes);
-  for (i = 0; i < bytes; i++)
-    {
-      int index = big_p ? bytes - i - 1 : i;
-
-      buf[last + index] = data & 0xff;
-      data >>= 8;
-    }
+  store_unsigned_integer (&buf[last], bytes, byte_order, data);
 }
 
 /* Subroutine of find_command to simplify it.
@@ -53,7 +46,7 @@ put_bits (uint64_t data, gdb::byte_vector &buf, int bits, bfd_boolean big_p)
 static gdb::byte_vector
 parse_find_args (const char *args, ULONGEST *max_countp,
 		 CORE_ADDR *start_addrp, ULONGEST *search_space_lenp,
-		 bfd_boolean big_p)
+		 bfd_endian byte_order)
 {
   /* Default to using the specified type.  */
   char size = '\0';
@@ -171,13 +164,13 @@ parse_find_args (const char *args, ULONGEST *max_countp,
 	      pattern_buf.push_back (x);
 	      break;
 	    case 'h':
-	      put_bits (x, pattern_buf, 16, big_p);
+	      put_bits (x, pattern_buf, 16, byte_order);
 	      break;
 	    case 'w':
-	      put_bits (x, pattern_buf, 32, big_p);
+	      put_bits (x, pattern_buf, 32, byte_order);
 	      break;
 	    case 'g':
-	      put_bits (x, pattern_buf, 64, big_p);
+	      put_bits (x, pattern_buf, 64, byte_order);
 	      break;
 	    }
 	}
@@ -210,7 +203,7 @@ static void
 find_command (const char *args, int from_tty)
 {
   struct gdbarch *gdbarch = get_current_arch ();
-  bfd_boolean big_p = gdbarch_byte_order (gdbarch) == BFD_ENDIAN_BIG;
+  bfd_endian byte_order = gdbarch_byte_order (gdbarch);
   /* Command line parameters.
      These are initialized to avoid uninitialized warnings from -Wall.  */
   ULONGEST max_count = 0;
@@ -223,7 +216,7 @@ find_command (const char *args, int from_tty)
   gdb::byte_vector pattern_buf = parse_find_args (args, &max_count,
 						  &start_addr,
 						  &search_space_len,
-						  big_p);
+						  byte_order);
 
   /* Perform the search.  */
 

-- 
2.49.0

Re: [PATCH 3/4] Drop use of bfd_boolean from findcmd.c

[Guinevere Larsen]

On 3/7/26 3:20 PM, Tom Tromey wrote:
> This changes findcmd.c to remove the use of bfd_boolean.  Instead, the
> byte order is used directly.  put_bits is also rewritten in terms of
> store_unsigned_integer.
> ---
>   gdb/findcmd.c | 33 +++++++++++++--------------------
>   1 file changed, 13 insertions(+), 20 deletions(-)
>
> diff --git a/gdb/findcmd.c b/gdb/findcmd.c
> index c4ee0ec97e4..b52c44d48d0 100644
> --- a/gdb/findcmd.c
> +++ b/gdb/findcmd.c
> @@ -24,27 +24,20 @@
>   #include "cli/cli-utils.h"
>   #include <algorithm>
>   #include "gdbsupport/byte-vector.h"
> +#include "extract-store-integer.h"
>   
> -/* Copied from bfd_put_bits.  */
> +/* Append DATA to BUF, writing only the specified number of bits,
> +   using the given endian-ness.  */
I think that this comment should probably specify that it should be an 
integer number of bytes, because I thought you could have like 9 bits 
and was wondering how the code managed that (I misread the diff thinking 
the assert was removed).
>   
>   static void
> -put_bits (uint64_t data, gdb::byte_vector &buf, int bits, bfd_boolean big_p)
> +put_bits (uint64_t data, gdb::byte_vector &buf, int bits,
> +	  bfd_endian byte_order)
>   {
> -  int i;
> -  int bytes;
> -
>     gdb_assert (bits % 8 == 0);
> -
> -  bytes = bits / 8;
> +  int bytes = bits / 8;
>     size_t last = buf.size ();
>     buf.resize (last + bytes);
> -  for (i = 0; i < bytes; i++)
> -    {
> -      int index = big_p ? bytes - i - 1 : i;
> -
> -      buf[last + index] = data & 0xff;
> -      data >>= 8;
> -    }
> +  store_unsigned_integer (&buf[last], bytes, byte_order, data);
>   }
>   
>   /* Subroutine of find_command to simplify it.
> @@ -53,7 +46,7 @@ put_bits (uint64_t data, gdb::byte_vector &buf, int bits, bfd_boolean big_p)
>   static gdb::byte_vector
>   parse_find_args (const char *args, ULONGEST *max_countp,
>   		 CORE_ADDR *start_addrp, ULONGEST *search_space_lenp,
> -		 bfd_boolean big_p)
> +		 bfd_endian byte_order)
>   {
>     /* Default to using the specified type.  */
>     char size = '\0';
> @@ -171,13 +164,13 @@ parse_find_args (const char *args, ULONGEST *max_countp,
>   	      pattern_buf.push_back (x);
>   	      break;
>   	    case 'h':
> -	      put_bits (x, pattern_buf, 16, big_p);
> +	      put_bits (x, pattern_buf, 16, byte_order);
>   	      break;
>   	    case 'w':
> -	      put_bits (x, pattern_buf, 32, big_p);
> +	      put_bits (x, pattern_buf, 32, byte_order);
>   	      break;
>   	    case 'g':
> -	      put_bits (x, pattern_buf, 64, big_p);
> +	      put_bits (x, pattern_buf, 64, byte_order);
>   	      break;
>   	    }
>   	}
> @@ -210,7 +203,7 @@ static void
>   find_command (const char *args, int from_tty)
>   {
>     struct gdbarch *gdbarch = get_current_arch ();
> -  bfd_boolean big_p = gdbarch_byte_order (gdbarch) == BFD_ENDIAN_BIG;
> +  bfd_endian byte_order = gdbarch_byte_order (gdbarch);
>     /* Command line parameters.
>        These are initialized to avoid uninitialized warnings from -Wall.  */
>     ULONGEST max_count = 0;
> @@ -223,7 +216,7 @@ find_command (const char *args, int from_tty)
>     gdb::byte_vector pattern_buf = parse_find_args (args, &max_count,
>   						  &start_addr,
>   						  &search_space_len,
> -						  big_p);
> +						  byte_order);
>   
>     /* Perform the search.  */
>   
>

-- 
Cheers,
Guinevere Larsen
it/its
she/her (deprecated)

[PATCH 4/4] Poison bfd_boolean

[Tom Tromey]

BFD is moving away from bfd_boolean.  This patch arranges for gdb to
do the same by having BFD poison the identifier so that no new uses
will creep in.
---
 gdb/defs.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/gdb/defs.h b/gdb/defs.h
index b378449343d..de4f44e2458 100644
--- a/gdb/defs.h
+++ b/gdb/defs.h
@@ -32,6 +32,8 @@
 #undef PACKAGE_TARNAME
 
 #include <config.h>
+
+#define POISON_BFD_BOOLEAN
 #include "bfd.h"
 
 #include <sys/types.h>

-- 
2.49.0

Re: [PATCH 0/4] Remove some bfd_boolean uses

[Guinevere Larsen]

On 3/7/26 3:20 PM, Tom Tromey wrote:
> I saw some recent commits removing uses of bfd_boolean from the tree.
> This series removes the remaining uses from gdb as well, and then
> arranges to poison the identifier using the BFD-supplied approach.
>
> Signed-off-by: Tom Tromey <tom@tromey.com>
> ---

Hi Tom!

I looked over this series and everything looks ok to me. I have sent a 
nitpick suggestion on patch 3, but otherwise I think this looks nice

Reviewed-By: Guinevere Larsen <guinevere@redhat.com>

> Tom Tromey (4):
>        Use bool, not bfd_boolean
>        Don't use BFD TRUE/FALSE constants
>        Drop use of bfd_boolean from findcmd.c
>        Poison bfd_boolean
>
>   gdb/arc-linux-tdep.c              |  6 +++---
>   gdb/compile/compile-object-load.c |  4 ++--
>   gdb/defs.h                        |  2 ++
>   gdb/elfread.c                     |  2 +-
>   gdb/findcmd.c                     | 33 +++++++++++++--------------------
>   gdb/ft32-tdep.c                   |  6 +++---
>   gdb/machoread.c                   |  4 ++--
>   7 files changed, 26 insertions(+), 31 deletions(-)
> ---
> base-commit: a00fddb175ff5cb9edef18beedfba38fc6fc6017
> change-id: 20260307-no-bfd-boolean-1c8cbdbf2849
>
> Best regards,


-- 
Cheers,
Guinevere Larsen
it/its
she/her (deprecated)

Re: [PATCH 0/4] Remove some bfd_boolean uses

[Tom Tromey]

>>>>> "Guinevere" == Guinevere Larsen <guinevere@redhat.com> writes:

Guinevere> Hi Tom!

Guinevere> I looked over this series and everything looks ok to me. I have sent a
Guinevere> nitpick suggestion on patch 3, but otherwise I think this looks nice

Guinevere> Reviewed-By: Guinevere Larsen <guinevere@redhat.com>

FWIW I think these were all obsoleted by Alan's patches to remove bfd_boolean.

Tom