Hi Joel,

that's great info to have, thanks for taking the time to spell it out. I'll 
digest this documentation before continuing.

Regarding i586, this appears to be translated into i686 internally by the make 
system such that the builds which actually take place are i686 builds, which 
puzzled me, but now puzzles me less. 

If I knew up front that I ( and yes, there's a distributed team involved too ) 
would only be touching packages x and y, I'd probably be taking a different 
approach right now, but I take the point that lots of work and bandwidth can 
potentially be saved by focusing on just the 'relevant' pieces of the puzzle. 
When I know what those pieces are, I can shift strategies. In the meantime, 
this is proving educational.  

Thanks,
Mark
________________________________________
From: Clark, Joel [[email protected]]
Sent: 15 January 2014 17:33
To: Mark De Roussier; ronan; [email protected]
Subject: RE: i586 build failing for GMP ?

1) The https://source.tizen.org/documentation/developer-guide is not IVI 
specific and not easily updated by the IVI team.  You should look at 
https://wiki.tizen.org/wiki/IVI#Tizen_IVI_OS_development for hints on building 
IVI.
2) Tizen IVI for IA currently utilizes the i686 build, which we run on 
SandyBridge, IvyBridge and Baytrail systems. Since the i586 builds are not 
being done in tizen.org, they are more likely to have issues.
3) gbs build environment is designed for building a few packages in sync with 
the git and download repos on tizen.org.  Typical development only keeps a 
small part of tizen in the local developer build system, only the things you 
are really looking at or modifying, and only rebuild the packages you modifiy 
on your local build system.  Many developers start off wanting to build all of 
tizen on their local system using gbs, but all of them find this difficult to 
maintain in the long run. On the other hand investing in a full replica of the 
tizen build system for local builds is also resource intensive and typically 
only done when a vendors product development plans requirements and resources 
make this investment worthwhile.  Small teams and individual developers doing 
small incremental local builds using GBS has been shown to be very viable.  I 
can assure you there are Tizen IVI teams with several dozen members distributed 
over 3-4 continents that constantly use gbs for local builds by only keeping a 
few packages local and depending on download.tizen.org for the majority of the 
packages.

Regards
Joel


-----Original Message-----
From: [email protected] [mailto:[email protected]] On 
Behalf Of Mark De Roussier
Sent: Wednesday, January 15, 2014 8:34 AM
To: ronan; [email protected]
Subject: RE: i586 build failing for GMP ?

Hi Ronan,

I am working my way through the instructions given here :

https://source.tizen.org/documentation/developer-guide

for setting up a dev environment, cloning the Tizen IVI project ( master 
branch, latest ), building all the packages, and building an image. I'm running 
on Kubuntu 12.04, fully patched.

So I have git, repo and gbs as per the instructions. I've cloned the project 
using this command given in the docs :

repo init -u ssh://[email protected]:29418/scm/manifest -b tizen -m 
ivi.xml

I'm building with :

gbs build -A i586 --threads=4 --exclude=<whatever needs to be excluded>

Contrary to what is stated in the docs, I'm finding circular dependencies being 
reported, and --exclude does seem to be required for Tizen 3. I exclude 
whatever packages I need to based on the circular dependency complaints from 
the command run without the exclude. At the moment, it's texinfo that's 
reporting a circular dependency.

For ppl, git log shows

commit 45d013e7f198300c47d2db36b212d8e0f5bbd438
Author: Alexandru Cornea <[email protected]>
Date:   Mon Jul 1 22:47:35 2013 +0300

    resetting manifest requested domain to floor

The problem seems to be :

[  139s] libtool: compile:  g++ -DHAVE_CONFIG_H -I. -I.. -I.. -I../src -g -O2 
-frounding-math -O2 -g -m32 -march=i686 -mtune=i686 -fmessage-length=0 
-D_FORTIFY_SOURCE=2 -fstack-protector -funwind-tables 
-fasynchronous-unwind-tables -W -Wall -MT Congruence_System.lo -MD -MP -MF 
.deps/Congruence_System.Tpo -c Congruence_System.cc  -fPIC -DPIC -o 
.libs/Congruence_System.o [  142s] In file included from checked.defs.hh:29:0,
[  142s]                  from Checked_Number.defs.hh:28,
[  142s]                  from Coefficient.types.hh:16,
[  142s]                  from Coefficient.defs.hh:27,
[  142s]                  from Row.defs.hh:29,
[  142s]                  from Linear_Row.defs.hh:29,
[  142s]                  from Constraint.defs.hh:29,
[  142s]                  from Constraint.cc:26:
[  142s] mp_std_bits.defs.hh:48:7: error: redefinition of 'class 
std::numeric_limits<__gmp_expr<__mpz_struct [1], __mpz_struct [1]> >'
[  142s]  class numeric_limits<mpz_class> {
[  142s]        ^

...and many similar complaints.

For libffi, git log shows

commit 9afe851c146c40a137b5cf099d1e16658c3a38a3
Author: Anas Nashif <[email protected]>
Date:   Mon Jul 8 08:11:25 2013 -0400

    fixed include path

Not so clear what the problem is here, but maybe :

[    0s] processing specfile 
/home/mderouss/GBS-ROOT/local/sources/tizen3.0_ivi/libffi48-4.8.2-1/libffi48.spec
 ...
[    0s] init_buildsystem --configdir /usr/lib/build/configs --cachedir 
/home/mderouss/GBS-ROOT/local/cache --repository 
/home/mderouss/GBS-ROOT/local/repos/tizen3.0_ivi/i586/RPMS --repository 
/home/mderouss/TizenIVI3v0/pre-built/toolchain-x86/ 
/home/mderouss/GBS-ROOT/local/sources/tizen3.0_ivi/libffi48-4.8.2-1/libffi48.spec
 ...
[    0s] initializing 
/home/mderouss/GBS-ROOT/local/BUILD-ROOTS/scratch.i586.2/.srcfiles.cache ...
[    0s] /usr/lib/build/createrpmdeps 
/home/mderouss/GBS-ROOT/local/repos/tizen3.0_ivi/i586/RPMS
[    1s] /usr/lib/build/createrepomddeps 
/home/mderouss/TizenIVI3v0/pre-built/toolchain-x86/
[    2s] expanding package dependencies...
[    3s] Warning: spec file parser line 384: can't expand %(...)
[    3s] Warning: spec file parser line 398: can't expand %(...)

Cheers,
Mark


MARK DE ROUSSIER
Team Lead

Symphony Teleca
Sunley House, 46 Jewry Street, Winchester, Hampshire, SO23 8RY
Phone: +441962891219, Fax: +441962868867 
mailto:[email protected]
http://www.symphonyteleca.com

Teleca Limited, a company registered in England & Wales, registration number 
2773878, registered office at Sunley House, 46 Jewry Street, Winchester, 
Hampshire SO23 8RY. VAT registration number GB 674 6583 90

Follow what's going on at Symphony Teleca's blog on 
http://www.symphonyteleca.com/blog. Please consider the environment before you 
print.

Notice to recipient: This e-mail (including any attachments) is meant for the 
intended recipient only, may contain confidential and proprietary information, 
and is protected by law. If you received this e-mail in error, please 
immediately notify the sender of the error by return e-mail, delete this 
communication and any attachments, and shred any printouts. Unauthorized 
review, use, dissemination, distribution, copying or taking of any action based 
on this communication is strictly prohibited.

________________________________________
From: ronan [[email protected]]
Sent: 15 January 2014 14:59
To: Mark De Roussier; [email protected]
Subject: Re: i586 build failing for GMP ?

Currently I'm build libffi and ppl in Tizen:Generic project, and they build 
fine.

(1) Can you explan what do you try to do?
(2) From what Tizen Project ?
(3) Your sources packages git src+commit(sha).
(4) What is your tools?

Thanks


Le 15/01/2014 15:18, Mark De Roussier a écrit :
> Hi Ronan,
>
> looks like we crossed over here, this particular issue has gone now with a 
> fresh sync, as I've now noted in a followup to my original post.
>
> But I do now have two further fails, with libffi and ppl. The ppl one appears 
> to be GMP related, the libffi one seems to be a make problem. I was going to 
> make a separate post for those, but would you prefer them on this thread ?
>
> Cheers,
> Mark
_______________________________________________
IVI mailing list
[email protected]
https://lists.tizen.org/listinfo/ivi
_______________________________________________
IVI mailing list
[email protected]
https://lists.tizen.org/listinfo/ivi

Reply via email to