Patchwork boost: bump version to 1.53.0

login
register
mail settings
Submitter Victor Hiairrassary
Date Feb. 6, 2013, 9:30 a.m.
Message ID <1360143004-29603-2-git-send-email-victor.hiairrassary.ml@gmail.com>
Download mbox | patch
Permalink /patch/218496/
State Superseded
Headers show

Comments

Victor Hiairrassary - Feb. 6, 2013, 9:30 a.m.
From: Victor Hiairrassary <victor.hiairrassary.ml@gmail.com>


Signed-off-by: Victor Hiairrassary <victor.hiairrassary.ml@gmail.com>
---
 package/boost/Config.in | 3 +++
 package/boost/boost.mk  | 5 +++--
 2 files changed, 6 insertions(+), 2 deletions(-)
Thomas Petazzoni - Feb. 6, 2013, 9:32 a.m.
Dear victor.hiairrassary.ml@gmail.com,

On Wed,  6 Feb 2013 10:30:04 +0100, victor.hiairrassary.ml@gmail.com
wrote:

>  BOOST_FLAGS =
> -BOOST_WITHOUT_FLAGS = python context
> +BOOST_WITHOUT_FLAGS = python atomic

Maybe there was a reason why "context" was unconditionally excluded
from the build, before your patch. Did you check why?

Also, it would be good to have a comment here that explains why python
and atomic are unconditionally excluded for the build.

Thanks,

Thomas
Victor Hiairrassary - Feb. 6, 2013, 11:15 a.m.
On 02/06/2013 10:32 AM, Thomas Petazzoni wrote:

Dear victor.hiairrassary.ml@gmail.com,

On Wed,  6 Feb 2013 10:30:04 +0100, victor.hiairrassary.ml@gmail.com
wrote:


  BOOST_FLAGS =
-BOOST_WITHOUT_FLAGS = python context
+BOOST_WITHOUT_FLAGS = python atomic

 Maybe there was a reason why "context" was unconditionally excluded
from the build, before your patch. Did you check why?

 context was excluded because compilation is impossible with boost
1.52.0 and uClibc, there is a problem with uintptr_t and intptr_t.
But this is fixed in latest version, 1.53.0.

Also, it would be good to have a comment here that explains why python
and atomic are unconditionally excluded for the build.

 atomic is a new boost library since boost 1.53.0. It is excluded because
it can not compile with uClibc (uintptr_t missing too), but this is now
fixed in
trunk version, see ticket https://svn.boost.org/trac/boost/ticket/7979

In this patch, atomic is simply excluded. So either we wait for boost 1.54.0
releasing orI can add the patch to fix atomic compilation in buildroot and
add boost-atomic in menuconfig?

I do not nothing about pyhton exclusion, sorry.

 Thanks,

Thomas
Thomas Petazzoni - Feb. 6, 2013, 11:41 a.m.
Dear Victor Hiairrassary,

Would it be possible to use an e-mail client that complies with the
open-source community standards, by sending e-mails as plain text? The
plain text version of your e-mail makes no discussion between what you
said and what I said, which is very confusing.

On Wed, 6 Feb 2013 12:15:43 +0100, Victor Hiairrassary wrote:

>  Maybe there was a reason why "context" was unconditionally excluded
> from the build, before your patch. Did you check why?
> 
>  context was excluded because compilation is impossible with boost
> 1.52.0 and uClibc, there is a problem with uintptr_t and intptr_t.
> But this is fixed in latest version, 1.53.0.

Ok. This should be mentioned in the commit log, then.

> Also, it would be good to have a comment here that explains why python
> and atomic are unconditionally excluded for the build.
> 
>  atomic is a new boost library since boost 1.53.0. It is excluded because
> it can not compile with uClibc (uintptr_t missing too), but this is now
> fixed in
> trunk version, see ticket https://svn.boost.org/trac/boost/ticket/7979

Ok. This should be mentioned in a comment above the line that disables
the atomic module, so that we know in the future why it is disabled.

> In this patch, atomic is simply excluded. So either we wait for boost 1.54.0
> releasing orI can add the patch to fix atomic compilation in buildroot and
> add boost-atomic in menuconfig?

No, it's perfectly fine to have the atomic thing disabled for now,
since it wasn't available in Buildroot until now. I just want a comment
explaining why it's disabled.

> I do not nothing about pyhton exclusion, sorry.

Ok. But since you're not changing this, it's not necessarily your
responsibility to justify this, so we can leave this part as is.

Could you resubmit a new version of the patch, with more details in the
commit log, and a comment before the line disabling the atomic module?

Thanks a lot!

Thomas
Victor Hiairrassary - Feb. 6, 2013, 4:27 p.m.
On 02/06/2013 12:41 PM, Thomas Petazzoni wrote:
> Dear Victor Hiairrassary,
>
> Would it be possible to use an e-mail client that complies with the
> open-source community standards, by sending e-mails as plain text? The
> plain text version of your e-mail makes no discussion between what you
> said and what I said, which is very confusing.
Oh, sorry for the mistake!
> Could you resubmit a new version of the patch, with more details in the
> commit log, and a comment before the line disabling the atomic module?
Done in another mail! I hope that everything is good!

Victor Hiairrassary
> Thanks a lot!
>
> Thomas
Victor Hiairrassary - Feb. 7, 2013, 3:42 p.m.
On 02/06/2013 12:41 PM, Thomas Petazzoni wrote:
> Dear Victor Hiairrassary,
>
> Could you resubmit a new version of the patch, with more details in the
> commit log, and a comment before the line disabling the atomic module?
>
> Thanks a lot!
>
> Thomas
Yesterday I've sent the new patch through (git send-email) on
buildroot@busybox.net. But it is only available on buildroot
patchwork at http://patchwork.ozlabs.org/patch/218673/ and not
on mailing list (I don't know why).

Regards,

Victor Hiairrassary
Baruch Siach - Feb. 7, 2013, 7:04 p.m.
Hi Victor,

On Thu, Feb 07, 2013 at 04:42:39PM +0100, Victor Hiairrassary wrote:
> On 02/06/2013 12:41 PM, Thomas Petazzoni wrote:
> >Could you resubmit a new version of the patch, with more details in the
> >commit log, and a comment before the line disabling the atomic module?
> >
> >Thanks a lot!
> >
> >Thomas
> Yesterday I've sent the new patch through (git send-email) on
> buildroot@busybox.net. But it is only available on buildroot
> patchwork at http://patchwork.ozlabs.org/patch/218673/ and not
> on mailing list (I don't know why).

I got it. It is also archived at 
http://lists.busybox.net/pipermail/buildroot/2013-February/066841.html.

baruch

Patch

diff --git a/package/boost/Config.in b/package/boost/Config.in
index 19dcc60..e36e7d3 100644
--- a/package/boost/Config.in
+++ b/package/boost/Config.in
@@ -21,6 +21,9 @@  if BR2_PACKAGE_BOOST
 config BR2_PACKAGE_BOOST_CHRONO
 	bool "boost-chrono"
 
+config BR2_PACKAGE_BOOST_CONTEXT
+	bool "boost-context"
+
 config BR2_PACKAGE_BOOST_DATE_TIME
 	bool "boost-date_time"
 
diff --git a/package/boost/boost.mk b/package/boost/boost.mk
index 5e1bb4e..a056662 100644
--- a/package/boost/boost.mk
+++ b/package/boost/boost.mk
@@ -4,7 +4,7 @@ 
 #
 #############################################################
 
-BOOST_VERSION = 1.52.0
+BOOST_VERSION = 1.53.0
 BOOST_FILE_VERSION = $(subst .,_,$(BOOST_VERSION))
 BOOST_SOURCE = boost_$(BOOST_FILE_VERSION).tar.bz2
 BOOST_SITE = http://downloads.sourceforge.net/project/boost/boost/$(BOOST_VERSION)
@@ -15,9 +15,10 @@  TARGET_CC_VERSION = $(shell $(TARGET_CC) -dumpversion)
 BOOST_DEPENDENCIES = bzip2 zlib
 
 BOOST_FLAGS =
-BOOST_WITHOUT_FLAGS = python context
+BOOST_WITHOUT_FLAGS = python atomic
 
 BOOST_WITHOUT_FLAGS += $(if $(BR2_PACKAGE_BOOST_CHRONO),,chrono)
+BOOST_WITHOUT_FLAGS += $(if $(BR2_PACKAGE_BOOST_CHRONO),,context)
 BOOST_WITHOUT_FLAGS += $(if $(BR2_PACKAGE_BOOST_DATE_TIME),,date_time)
 BOOST_WITHOUT_FLAGS += $(if $(BR2_PACKAGE_BOOST_EXCEPTION),,exception)
 BOOST_WITHOUT_FLAGS += $(if $(BR2_PACKAGE_BOOST_FILESYSTEM),,filesystem)