Re: [Patch] Add private dumper to objdump

["H.J. Lu" <hjl.tools@gmail.com>]

On Mon, May 16, 2011 at 6:35 AM, Hans-Peter Nilsson
<hans-peter.nilsson@axis.com> wrote:
>> From: Tristan Gingold <gingold@adacore.com>
>> Date: Mon, 16 May 2011 14:22:32 +0200
>
>> On May 10, 2011, at 4:46 AM, Alan Modra wrote:
>>
>> > On Thu, May 05, 2011 at 04:27:38PM +0200, Tristan Gingold wrote:
>> >>    * od-xcoff.c: New file.
>> >>    * objdump.h: New file.
>> >>    * objdump.c: Include objdump.h
>> >>    (dump_private_options, objdump_private_vectors): New variables.
>> >>    (usage): Mention -P/--private.  Display handled options.
>> >>    (long_options): Add -P/--private.
>> >>    (dump_target_specific): New function.
>> >>    (dump_bfd): Handle dump_private_options.
>> >>    (main): Handle -P.
>> >>    * doc/binutils.texi (objdump): Document -P/--private.
>> >>    * configure.in (OBJDUMP_PRIVATE_VECTORS, OBJDUMP_PRIVATE_OFILES):
>> >>    New variables, compute them.
>> >>    (od_vectors): Add vectors for private dumpers. Make them uniq.
>> >>    (OBJDUMP_DEFS): Add OBJDUMP_PRIVATE_VECTORS.
>> >>    * Makefile.am (HFILES): Add objdump.h
>> >>    (CFILES): Add od-xcoff.c
>> >>    (OBJDUMP_PRIVATE_OFILES): New variable.
>> >>    (objdump_DEPENDENCIES): Append OBJDUMP_PRIVATE_OFILES.
>> >>    (objdump_LDADD): Ditto.
>> >>    (EXTRA_objdump_SOURCES): Define.
>> >>    * Makefile.in: Regenerate.
>> >>    * configure: Regenerate.
>> >
>> > OK.
>>
>> Thanks, committed.

There is no ChangeLog entry.

> Looks like you missed updating the test-suite.  Changing the
> format of objdump or readelf requires that.  My autotester
> complains for cris-axis-elf and cris-axis-linux-gnu:
>
> Running /tmp/hpautotest-binutils/bsrc/src/ld/testsuite/ld-scripts/phdrs.exp ...
> FAIL: PHDRS
> Running /tmp/hpautotest-binutils/bsrc/src/ld/testsuite/ld-scripts/phdrs2.exp ...
> FAIL: PHDRS2
>
> But, I can't trivially figure out why the test-suite fails as it
> does; it looks like the same thing would happen for all
> architectures and the failure mode is weird.  Quoting the ld.log
> for the first failure ("same" for the second):

The problem is we have

  -p, --private-headers    Display object format specific file header contents

and we support short and unambiguous names.

-- 
H.J.