Re: [PATCH] Find debug symbols file by buildid for PE file format also

[Joel Brobecker <brobecker@adacore.com>]

> bfd/ChangeLog:
> 
> 2015-04-07  Jon Turney  <jon.turney@dronecode.org.uk>
> 
> 	* elf-bfd.h : Remove struct elf_build_id.
> 	* bfd.c : Add struct bfd_build_id.
> 	* bfd-in2.h: Regenerate.
> 	* elf.c (elfobj_grok_gnu_build_id): Update to use bfd_build_id.
> 	* libpei.h: Add protoype and macros for
> 	bfd_XXi_slurp_codeview_record.
> 	* peXXigen.c (_bfd_XXi_slurp_codeview_record): Make public
> 	* peicode.h (pe_bfd_read_buildid): Add.
> 	(pe_bfd_object_p): Use pe_bfd_read_buildid().

The BFD part of the patch needs to be approved by the binutils
maintainers.

> gdb/ChangeLog:
> 
> 2015-04-07  Jon Turney  <jon.turney@dronecode.org.uk>
> 
> 	* build-id.c (build_id_bfd_get): Use bfd_build_id.
> 	(build_id_verify): Ditto.
> 	* build-id.h: Ditto.
> 	(find_separate_debug_file_by_buildid): Ditto.
> 	* python/py-objfile.c (objfpy_get_build_id)
> 	(objfpy_build_id_matches, objfpy_lookup_objfile_by_build_id): Ditto.
> 	* coffread.c (coff_symfile_read): Try
> 	find_separate_debug_file_by_buildid.

Can you also mention in the ChangeLog the #include changes
(removing and adding)?

> 
> gdb/doc/ChangeLog:
> 
> 2015-04-07  Jon Turney  <jon.turney@dronecode.org.uk>
> 
> 	* gdb.texinfo (Separate Debug Files): Document that PE is also
> 	supported.
> 
> gdb/testsuite/ChangeLog:
> 
> 2015-04-07  Jon Turney  <jon.turney@dronecode.org.uk>
> 
> 	* gdb.base/sepdebug.exp: Add EXEEXT where needed.
> 	* lib/gdb.exp (get_build_id): Teach how to extract build-id from a
> 	PE file.
> 	* lib/future.exp (gdb_find_objdump): Add gdb_find_objdump.

All GDB changes look mostly OK to me. In fact, the only comments are
very minor in nature, so this part is pre-approved pending those tiny
issues being addressed.

> @@ -29,19 +28,19 @@
>  
>  /* See build-id.h.  */
>  
> -const struct elf_build_id *
> +const struct bfd_build_id *
>  build_id_bfd_get (bfd *abfd)
>  {
> -  if (!bfd_check_format (abfd, bfd_object)
> -      || bfd_get_flavour (abfd) != bfd_target_elf_flavour
> -      /* Although this is ELF_specific, it is safe to do in generic
> -	 code because it does not rely on any ELF-specific symbols at
> -	 link time, and if the ELF code is not available in BFD, then
> -	 ABFD will not have the ELF flavour.  */
> -      || elf_tdata (abfd)->build_id == NULL)
> +  if (!bfd_check_format (abfd, bfd_object))
>      return NULL;
>  
> -  return elf_tdata (abfd)->build_id;
> +  if (abfd->build_id != NULL)
> +    {
> +      return abfd->build_id;
> +    }

Small GDB-specific coding style: For single-statement if blocks,
we decided that curly braces should not be used. Therefore:

  if (abfd->build_id != NULL)
    return abfd->build_id;

The only exception is when you have a comment in addition to the
statement. See:
https://sourceware.org/gdb/wiki/Internals%20GDB-C-Coding-Standards

> diff --git a/gdb/testsuite/lib/future.exp b/gdb/testsuite/lib/future.exp
> index 2fb635b..fd4c153 100644
> --- a/gdb/testsuite/lib/future.exp
> +++ b/gdb/testsuite/lib/future.exp
> @@ -104,6 +104,16 @@ proc gdb_find_objcopy {} {
>      return $objcopy
>  }
>  
> +proc gdb_find_objdump {} {

We should document every new function, by adding an introductory
comment.

-- 
Joel