From: Charles Celerier <cceleri@cs.stanford.edu>
To: David Bremner <david@tethera.net>, Notmuch Mail <notmuch@notmuchmail.org>
Subject: Re: [PATCH v2 5/5] T360-symbol-hiding: Use nm instead of objdump.
In-Reply-To: <87lhua3qvv.fsf@maritornes.cs.unb.ca>
References: <1399395748-44920-1-git-send-email-cceleri@cs.stanford.edu>
	<1399434615-28425-1-git-send-email-cceleri@cs.stanford.edu>
	<1399434615-28425-2-git-send-email-cceleri@cs.stanford.edu>
	<87d2fov2qj.fsf@maritornes.cs.unb.ca> <m2zjisnwie.fsf@jane.lan>
	<87a9arvs9b.fsf@maritornes.cs.unb.ca> <m2iopfoihw.fsf@jane.lan>
	<87lhua3qvv.fsf@maritornes.cs.unb.ca>
User-Agent: Notmuch/0.18~rc0+32~gd5092c2 (http://notmuchmail.org) Emacs/24.3.1
	(x86_64-apple-darwin13.1.0)
Date: Sat, 10 May 2014 10:55:43 -0400
Message-ID: <m24n0xy974.fsf@jane.lan>
MIME-Version: 1.0
Content-Type: multipart/signed; boundary="=-=-=";
	micalg=pgp-sha1; protocol="application/pgp-signature"
Precedence: list

--=-=-=
Content-Type: text/plain

David Bremner <david@tethera.net> writes:

> Charles Celerier <cceleri@cs.stanford.edu> writes:
>
>> At this point, I'm not sure how to create a better version of this
>> patch. Are we renaming functions in notmuch-private.h? Should we stick
>> with objdump or switch to using nm?
>
> I'd go with renaming any functions that start with notmuch to start with
> _notmuch.

Ok. Should I pull the patch from
id:1399402716-13714-1-git-send-email-cceleri@cs.stanford.edu into this
patch series?

> Since it seems neither nm nor objdump provides uniform output between
> OS/X and Linux, there isn't an obvious advantage to switching to nm. Any
> idea if objdump can be made to work (in this test) on OS/X ?

I can install GNU objdump on OS/X, but the problem seems to be that
objdump changes its output depending on the object file format (see -t
option documentation in objdump(1)). What does the output of nm look
like for you? Is really that different? Here is what my objdump output
looks like:

    $ objdump -t lib/*.o | sed -n '/\[\.text\] __\?notmuch/p' | tail
    00000000000009a0 g       0f SECT   01 0000 [.text] _notmuch_thread_get_authors
    0000000000000990 g       0f SECT   01 0000 [.text] _notmuch_thread_get_matched_messages
    0000000000000960 g       0f SECT   01 0000 [.text] _notmuch_thread_get_messages
    00000000000009d0 g       0f SECT   01 0000 [.text] _notmuch_thread_get_newest_date
    00000000000009c0 g       0f SECT   01 0000 [.text] _notmuch_thread_get_oldest_date
    00000000000009b0 g       0f SECT   01 0000 [.text] _notmuch_thread_get_subject
    00000000000009e0 g       0f SECT   01 0000 [.text] _notmuch_thread_get_tags
    0000000000000970 g       0f SECT   01 0000 [.text] _notmuch_thread_get_thread_id
    0000000000000950 g       0f SECT   01 0000 [.text] _notmuch_thread_get_toplevel_messages
    0000000000000980 g       0f SECT   01 0000 [.text] _notmuch_thread_get_total_messages

I suppose we could just account for the difference in output of objdump
on OS/X, but I was hoping that the output of nm would be more consistent
since its a simpler program.

chuck

--=-=-=
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG/MacGPG2 v2.0.22 (Darwin)

iQEcBAEBAgAGBQJTbj3wAAoJEOjk698fDjuAYEMH/iNwqlM0QtSt85w0uuFvcVpz
MwU86+EcpnSMk5Qw1ygzn8NKZHNbsR32Lka1Nko3PCzr+Uk0I35MIZkBoN08q6QX
snN0ME0/4VdtlEAaSoHbtK2KgPSGS8wBlcUQmVNoC5FV3qarUnzNK0qORdZ1QftG
GBuoNAY68l9UHp5RbQ+9cpxI2C8CESDc6IBXmYZemZOGlYKrqQiDk7PJOq+fUwPG
661ZV0HTGQ5BHZnSsXDYk3jIjeQTxfqJqWziTqwLAikx3Z7O7L/2laObP+ZrlQ4U
RjdkVdrfZWVk7ofeNwmqYC0eAs7O1V1TU2ijHTWHWFOw09b+DEo2nvZ5dshhpZE=
=Hio3
-----END PGP SIGNATURE-----
--=-=-=--