Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 05 Dec 2012 10:06:43 +0100
From:      Niclas Zeising <zeising@freebsd.org>
To:        Hans Ottevanger <hans@beastielabs.net>
Cc:        toolchain@freebsd.org
Subject:   Re: [CFT] devel/binutils 2.23
Message-ID:  <50BF0EA3.2080600@freebsd.org>
In-Reply-To: <50BEF61C.2050908@beastielabs.net>
References:  <201211141445.qAEEjTXQ047896@mech-cluster241.men.bris.ac.uk> <50A3FCEF.9060204@freebsd.org> <50A4A5A2.2000902@beastielabs.net> <50A4A69B.7030200@freebsd.org> <50B76AC2.4050207@freebsd.org> <50BA27F1.3080002@beastielabs.net> <50BCF220.6040905@beastielabs.net> <50BCFB43.8040906@freebsd.org> <50BEF61C.2050908@beastielabs.net>

next in thread | previous in thread | raw e-mail | index | archive | help
On 12/05/12 08:22, Hans Ottevanger wrote:
> On 12/03/12 20:19, Niclas Zeising wrote:
>> On 12/03/12 19:40, Hans Ottevanger wrote:
>>> On 12/01/12 16:53, Hans Ottevanger wrote:
>>>> On 11/29/12 15:01, Niclas Zeising wrote:
>>>>> On 11/15/12 09:23, Niclas Zeising wrote:
>>>>>> On 2012-11-15 09:19, Hans Ottevanger wrote:
>>>>>>> On 11/14/12 21:19, Niclas Zeising wrote:
>>>>>>>> On 11/14/12 15:45, Anton Shterenlikht wrote:
>>>>>>>>> It installed fine on ia64 and sparc64, both -current.
>>>>>>>>> I don't know how to test. Please advise if there are
>>>>>>>>> simple tests.
>>>>>>>>>
>>>>>>>>> Also, just to check, I manually deleted *orig files
>>>>>>>>> from under files/ after applying the patch:
>>>>>>>>>
>>>>>>>>> # ls -al /usr/ports/devel/binutils/files/
>>>>>>>>> total 20
>>>>>>>>> drwxr-xr-x  2 root  wheel  1024 Nov 14 12:58 .
>>>>>>>>> drwxr-xr-x  4 root  wheel   512 Nov 14 13:00 ..
>>>>>>>>> -rw-r--r--  1 root  wheel   353 Nov 14 12:55 patch-bfd_Makefile.in
>>>>>>>>> -rw-r--r--  1 root  wheel   297 Nov 14 12:55 patch-gold_Makefile.in
>>>>>>>>> -rw-r--r--  1 root  wheel   471 Nov 14 12:55 patch-gold_script.cc
>>>>>>>>> #
>>>>>>>>>
>>>>>>>>> because I think all files in this directory
>>>>>>>>> will be used as patches, no matter the name.
>>>>>>>>> Am I wrong?
>>>>>>>>>
>>>>>>>>> Anton
>>>>>>>>
>>>>>>>> Just compile test some binaries and see that they link and work ok.
>>>>>>>> The .orig files are left over when running patch, and has to be removed.
>>>>>>>>   Sorry if I wasn't clear on that in my previous mail.
>>>>>>>> Thanks for testing!
>>>>>>>> Regards!
>>>>>>>
>>>>>>> Please be aware that apparently something went wrong with the release of
>>>>>>> binutils-2.23 (see the discussion ending in:
>>>>>>>
>>>>>>> http://sourceware.org/ml/binutils/2012-10/msg00339.html
>>>>>>>
>>>>>>> though I doubt the glitches will affect your usage) and it has been
>>>>>>> re-released as binutils-2.23.1. Maybe it is better to base the update if
>>>>>>> the binutils port on that release.
>>>>>>>
>>>>>>
>>>>>> I noticed that late last night, but haven't had time to update the patch
>>>>>> yet.  Thank you for pointing it out.
>>>>>> Regards!
>>>>>>
>>>>>
>>>>> Hi!
>>>>> Apologies for the delay.  Attached is a patch that updates binutils from
>>>>> 2.22 to 2.23.1.  Please test it.  The plan is to commit it once 9.1 is
>>>>> out the door and the feature freeze on the ports tree is lifted.
>>>>> Regards!
>>>>>
>>>>
>>>> I tested your patch on amd64 and i386 systems (all a recent 8.3-STABLE
>>>> r243569).
>>>>
>>>> The patch applied cleanly and the resulting port compiled without
>>>> problems, both by directly using make and by using portmaster. I tested
>>>> the results by recompiling a fairly large application (my gcc based
>>>> cross-build environment for embedded development) using gcc 4.7 from the
>>>> ports and the new binutils-2.23.1 on both i386 and amd64, Everything
>>>> functioned as it should and up to now there were no surprises whatsoever.
>>>>
>>>> I do not have the systems to test the other architectures, but I will
>>>> retest on the 10.0-CURRENT i386 and amd64 systems that I expect to
>>>> install one of these days. I will come back to you to report on that.
>>>>
>>>> Kind regards,
>>>>
>>>> Hans Ottevanger
>>>>
>>>
>>> I have been taking a closer look at the output of make and find the
>>> following:
>>>
>>> =>  SHA256 Checksum OK for binutils-2.23.1.tar.bz2.
>>> ===>   Patching for binutils-2.23.1
>>> ===>   Applying FreeBSD patches for binutils-2.23.1
>>>    I can't seem to find a patch in there anywhere.
>>>    I can't seem to find a patch in there anywhere.
>>>    I can't seem to find a patch in there anywhere.
>>>    I can't seem to find a patch in there anywhere.
>>>    I can't seem to find a patch in there anywhere.
>>>    I can't seem to find a patch in there anywhere.
>>>    I can't seem to find a patch in there anywhere.
>>>    I can't seem to find a patch in there anywhere.
>>>    I can't seem to find a patch in there anywhere.
>>>    I can't seem to find a patch in there anywhere.
>>>    I can't seem to find a patch in there anywhere.
>>> ===>    binutils-2.23.1 depends on file: /usr/local/lib/libgmp.so - found
>>>
>>> This happens on both 8.3-STABLE and 10.0-CURRENT. It implies that 11 of
>>> the 14 patches in the directory "files" are not applied. I wonder how
>>> the binutils get to function at all without them, but the patches are
>>> probably for exceptional situations and other architectures then amd64
>>> and i386.
>>>
>>> Kind regards,
>>>
>>> Hans Ottevanger
>>>
>>
>> Have a look at the files/ directory.  It is probably only the .orig
>> files that are left from applying the patch, or if you didn't apply the
>> patch with -E, the files are probably still there but empty.
>> Regards!
>>
>
> Yep, those files are all empty, I should have looked there before
> grabbing my mailer. Using patch -E indeed solves the problem. BTW, good
> to see that only 3 patches are left.
>
> In other news, I also tested your pathes on an i386 10-CURRENT
> installation (r243831 and everything seems to work fine. I had to change
> my plans a bit concerning a test on amd64 10-CURRENT. That will have to
> wait till a system reshuffle over here, taking place only after 9.1 is
> released.

Thank you very much for your help in testing binutils!
There is no need for you to test amd64, unless you want to.  I have an 
amd64 installation here that I've tested on, and it works fine.
Regards!
-- 
Niclas Zeising



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?50BF0EA3.2080600>