Buildroot Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [Buildroot] [autobuild.buildroot.net] Build results for 2014-04-07
@ 2014-04-08  6:30 Thomas Petazzoni
  2014-04-08 21:00 ` Max Filippov
  0 siblings, 1 reply; 8+ messages in thread
From: Thomas Petazzoni @ 2014-04-08  6:30 UTC (permalink / raw)
  To: buildroot

Build statistics for 2014-04-07
===============================

        success : 51 
       failures : 34 
       timeouts : 1  
          TOTAL : 86 

Classification of failures by reason
====================================

  toolchain-external-undefined | 4 
             lttng-tools-2.4.0 | 4 
                    gpm-1.20.7 | 2 
             util-linux-2.24.1 | 2 
                 cairo-1.12.10 | 2 
czmq-b5730c5f8290a611fd3b92... | 2 
                    fltk-1.3.2 | 1 
          libpthread-stubs-0.3 | 1 
                latencytop-0.5 | 1 
                libsoup-2.43.1 | 1 
        gst1-plugins-bad-1.2.3 | 1 
make: *** [core-dependencie... | 1 
              fetchmail-6.3.26 | 1 
                     exim-4.82 | 1 
                    fbgrab-1.2 | 1 
    xserver_xorg-server-1.15.0 | 1 
                  luajit-2.0.3 | 1 
                    php-5.5.11 | 1 
                   duma-2.5.15 | 1 
      compiling ../../corelib/ | 1 
                  taglib-1.9.1 | 1 
            lttng-libust-2.4.0 | 1 
                  ffmpeg-1.2.5 | 1 
               qt5script-5.2.1 | 1 
libubox-bbd846ec2d72b262975... | 1 

Detail of failures
===================

    xtensa |                  cairo-1.12.10 | NOK | http://autobuild.buildroot.net/results/26991c8736853ee55989ac647eccd653b7bf234b/
    x86_64 |                  cairo-1.12.10 | NOK | http://autobuild.buildroot.net/results/6e78bb4bc501bda36ee6842637ab2a1098f17c2e/
    x86_64 |       compiling ../../corelib/ | TIM | http://autobuild.buildroot.net/results/a25b95e378514628de09f6aa595499f1d0f70c56/
      bfin | czmq-b5730c5f8290a611fd3b92... | NOK | http://autobuild.buildroot.net/results/dc61e328929fed951249c7638ed1e914877a43ce/
      bfin | czmq-b5730c5f8290a611fd3b92... | NOK | http://autobuild.buildroot.net/results/c51bc4db85b3558ac720637dd0c599839157c51f/
      bfin |                    duma-2.5.15 | NOK | http://autobuild.buildroot.net/results/22a49ba5a5019d0f67844ec26b44f784d9abaa66/
      bfin |                      exim-4.82 | NOK | http://autobuild.buildroot.net/results/95ddd8e9be96f3808612db02642d72cfe77b3c26/
      bfin |                     fbgrab-1.2 | NOK | http://autobuild.buildroot.net/results/f94ad329426b12dedf2b351040d132e7caf2dab8/
      i686 |               fetchmail-6.3.26 | NOK | http://autobuild.buildroot.net/results/cd386bd441ae5f16c99b2750d2fba0a8e34f7cac/
       arc |                   ffmpeg-1.2.5 | NOK | http://autobuild.buildroot.net/results/e5ec894a18806671133360a1bb0b88a8a9654a34/
    x86_64 |                     fltk-1.3.2 | NOK | http://autobuild.buildroot.net/results/53cfd61c9d04b693adb9c7ba2171ef061a7944d1/
microblazeel |                     gpm-1.20.7 | NOK | http://autobuild.buildroot.net/results/5c92df5ef01c461fd7ed0ab1c02c8674c7e8a5f9/
microblazeel |                     gpm-1.20.7 | NOK | http://autobuild.buildroot.net/results/cfc539adda2b836e5b38d852381d21f2a3e77f37/
       arm |         gst1-plugins-bad-1.2.3 | NOK | http://autobuild.buildroot.net/results/080ce592260728baef1cf9d1d08952f6faf29d17/
   powerpc |                 latencytop-0.5 | NOK | http://autobuild.buildroot.net/results/e62d35bcf7d6c98d0a0348532c2b159affb20e83/
       arm |           libpthread-stubs-0.3 | NOK | http://autobuild.buildroot.net/results/adde6ae82b9e66ee0bb206c2f334ea7a856cb0f5/
       arm |                 libsoup-2.43.1 | NOK | http://autobuild.buildroot.net/results/d22edc19001037f9593f45444ed953f74ee6b806/
      bfin | libubox-bbd846ec2d72b262975... | NOK | http://autobuild.buildroot.net/results/1a5c50f8ca0ae70c1e0d3733765f521790b97a30/
   powerpc |             lttng-libust-2.4.0 | NOK | http://autobuild.buildroot.net/results/a94c9f0c03bd968b5db3fde198e7ed033a709219/
       arm |              lttng-tools-2.4.0 | NOK | http://autobuild.buildroot.net/results/6b49c71fa183194bebde88398cb336d6e4f92a25/
       arm |              lttng-tools-2.4.0 | NOK | http://autobuild.buildroot.net/results/c17bdffb7c65312e8dc03acbc24b1f9b3171b00f/
       arm |              lttng-tools-2.4.0 | NOK | http://autobuild.buildroot.net/results/b96d30d896d087eda7b0e0d3b87b6f8ebcf1f689/
       arm |              lttng-tools-2.4.0 | NOK | http://autobuild.buildroot.net/results/8cd431f0f4f34683006648b3b2f8e8f3c4ef4358/
    mipsel |                   luajit-2.0.3 | NOK | http://autobuild.buildroot.net/results/d5232a215e9e9aeb40b2e4e353329a27603c79ab/
       arm | make: *** [core-dependencie... | NOK | http://autobuild.buildroot.net/results/f4d26e7e91f3ad134d6215ef0d969eda4deac240/
    xtensa |                     php-5.5.11 | NOK | http://autobuild.buildroot.net/results/959b77fa2c1f13b1958b234803437e09734e882e/
       arm |                qt5script-5.2.1 | NOK | http://autobuild.buildroot.net/results/4daee63d6e538f8dc77ed196487be1342ccac122/
   powerpc |                   taglib-1.9.1 | NOK | http://autobuild.buildroot.net/results/e92241a3f8cb6c25de2071eea90bbc38d4229dda/
     avr32 |   toolchain-external-undefined | NOK | http://autobuild.buildroot.net/results/996927e6fda49a09c9bb83b134a798ff908aca5a/
     avr32 |   toolchain-external-undefined | NOK | http://autobuild.buildroot.net/results/a5797c025bec50c10efdcff74945aab4021d05e4/
     avr32 |   toolchain-external-undefined | NOK | http://autobuild.buildroot.net/results/c5f4a68ac2789296c057010c7647ef7b694799ac/
     avr32 |   toolchain-external-undefined | NOK | http://autobuild.buildroot.net/results/f331a6eff0b0b93c73af52db3a6b43e4e598577e/
       sh4 |              util-linux-2.24.1 | NOK | http://autobuild.buildroot.net/results/b22131fee532bae01cdb896f0347aebd5505fdc6/
       sh4 |              util-linux-2.24.1 | NOK | http://autobuild.buildroot.net/results/a379afb8e09c78895a1876b8f1f99e9c4fccb5b5/
    x86_64 |     xserver_xorg-server-1.15.0 | NOK | http://autobuild.buildroot.net/results/f21c63cf5530e51586806ecce845e85d3e0a25c1/


-- 
http://autobuild.buildroot.net

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] [autobuild.buildroot.net] Build results for 2014-04-07
  2014-04-08  6:30 [Buildroot] [autobuild.buildroot.net] Build results for 2014-04-07 Thomas Petazzoni
@ 2014-04-08 21:00 ` Max Filippov
  2014-04-08 21:41   ` Thomas Petazzoni
  0 siblings, 1 reply; 8+ messages in thread
From: Max Filippov @ 2014-04-08 21:00 UTC (permalink / raw)
  To: buildroot

On Tue, Apr 8, 2014 at 10:30 AM, Thomas Petazzoni
<thomas.petazzoni@free-electrons.com> wrote:
> Build statistics for 2014-04-07
[...]
>     xtensa |                     php-5.5.11 | NOK | http://autobuild.buildroot.net/results/959b77fa2c1f13b1958b234803437e09734e882e/

Couldn't reproduce. But I had to disable BR2_PACKAGE_HOST_SAM_BA,
otherwise it refused to build with the following message:

Your Buildroot configuration uses pre-built tools for the x86 architecture,
but your build machine uses the x86-64 architecture without the 32
bits compatibility
library.

-- 
Thanks.
-- Max

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] [autobuild.buildroot.net] Build results for 2014-04-07
  2014-04-08 21:00 ` Max Filippov
@ 2014-04-08 21:41   ` Thomas Petazzoni
  2014-04-08 21:51     ` Max Filippov
  0 siblings, 1 reply; 8+ messages in thread
From: Thomas Petazzoni @ 2014-04-08 21:41 UTC (permalink / raw)
  To: buildroot

Dear Max Filippov,

On Wed, 9 Apr 2014 01:00:51 +0400, Max Filippov wrote:
> On Tue, Apr 8, 2014 at 10:30 AM, Thomas Petazzoni
> <thomas.petazzoni@free-electrons.com> wrote:
> > Build statistics for 2014-04-07
> [...]
> >     xtensa |                     php-5.5.11 | NOK | http://autobuild.buildroot.net/results/959b77fa2c1f13b1958b234803437e09734e882e/
> 
> Couldn't reproduce. But I had to disable BR2_PACKAGE_HOST_SAM_BA,
> otherwise it refused to build with the following message:
> 
> Your Buildroot configuration uses pre-built tools for the x86 architecture,
> but your build machine uses the x86-64 architecture without the 32
> bits compatibility library.

I think we can safely assume that the problem is not due to
BR2_PACKAGE_HOST_SAM_BA. I'll try to reproduce the same build failure
on the autobuilder, and gather the config.log file of php. I'll let you
know.

Thanks a lot for your investigation!

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] [autobuild.buildroot.net] Build results for 2014-04-07
  2014-04-08 21:41   ` Thomas Petazzoni
@ 2014-04-08 21:51     ` Max Filippov
  2014-04-10 21:41       ` Thomas Petazzoni
  0 siblings, 1 reply; 8+ messages in thread
From: Max Filippov @ 2014-04-08 21:51 UTC (permalink / raw)
  To: buildroot

On Wed, Apr 9, 2014 at 1:41 AM, Thomas Petazzoni
<thomas.petazzoni@free-electrons.com> wrote:
> Dear Max Filippov,
>
> On Wed, 9 Apr 2014 01:00:51 +0400, Max Filippov wrote:
>> On Tue, Apr 8, 2014 at 10:30 AM, Thomas Petazzoni
>> <thomas.petazzoni@free-electrons.com> wrote:
>> > Build statistics for 2014-04-07
>> [...]
>> >     xtensa |                     php-5.5.11 | NOK | http://autobuild.buildroot.net/results/959b77fa2c1f13b1958b234803437e09734e882e/
>>
>> Couldn't reproduce. But I had to disable BR2_PACKAGE_HOST_SAM_BA,
>> otherwise it refused to build with the following message:
>>
>> Your Buildroot configuration uses pre-built tools for the x86 architecture,
>> but your build machine uses the x86-64 architecture without the 32
>> bits compatibility library.
>
> I think we can safely assume that the problem is not due to
> BR2_PACKAGE_HOST_SAM_BA. I'll try to reproduce the same build failure
> on the autobuilder, and gather the config.log file of php. I'll let you
> know.

The tail of the failing build log:

...
checking for MySQL support... yes
checking for specified location of the MySQL UNIX socket... no
checking for mysql_close in -lmysqlclient... no
checking for mysql_error in -lmysqlclient... no
configure: error: mysql configure failed. Please check config.log for
more information.
make: *** [/home/test/test/2/output/build/php-5.5.11/.stamp_configured] Error 1

in my build the same place looks like this:

...
checking for MySQL support... yes
checking for specified location of the MySQL UNIX socket... no
checking for mysql_close in -lmysqlclient... yes
checking for MySQL UNIX socket location... no
checking for MySQLi support... no
checking whether to enable embedded MySQLi support... no
checking for Oracle Database OCI8 support... no
checking whether to enable Zend OPcache support... no
checking whether to enable pcntl support... yes
checking for fork... yes
...

so something must have happened with MySQL.

-- 
Thanks.
-- Max

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] [autobuild.buildroot.net] Build results for 2014-04-07
  2014-04-08 21:51     ` Max Filippov
@ 2014-04-10 21:41       ` Thomas Petazzoni
  2014-04-11  6:01         ` Max Filippov
  0 siblings, 1 reply; 8+ messages in thread
From: Thomas Petazzoni @ 2014-04-10 21:41 UTC (permalink / raw)
  To: buildroot

Dear Max Filippov,

On Wed, 9 Apr 2014 01:51:08 +0400, Max Filippov wrote:

> >> >     xtensa |                     php-5.5.11 | NOK | http://autobuild.buildroot.net/results/959b77fa2c1f13b1958b234803437e09734e882e/
> >>
> >> Couldn't reproduce. But I had to disable BR2_PACKAGE_HOST_SAM_BA,
> >> otherwise it refused to build with the following message:
> >>
> >> Your Buildroot configuration uses pre-built tools for the x86 architecture,
> >> but your build machine uses the x86-64 architecture without the 32
> >> bits compatibility library.
> >
> > I think we can safely assume that the problem is not due to
> > BR2_PACKAGE_HOST_SAM_BA. I'll try to reproduce the same build failure
> > on the autobuilder, and gather the config.log file of php. I'll let you
> > know.
> 
> The tail of the failing build log:
> 
> ...
> checking for MySQL support... yes
> checking for specified location of the MySQL UNIX socket... no
> checking for mysql_close in -lmysqlclient... no
> checking for mysql_error in -lmysqlclient... no
> configure: error: mysql configure failed. Please check config.log for
> more information.
> make: *** [/home/test/test/2/output/build/php-5.5.11/.stamp_configured] Error 1

Ok, I reproduced the problem. The config.log of php contains the
following problem:

configure:58794: result: no
configure:59004: checking for mysql_error in -lmysqlclient
configure:59029: /home/test/outputs/959b77fa2c1f13b1958b234803437e09734e882e/output/host/usr/bin/xtensa-buildroot-linux-uclibc-gcc -o conftest -I/include -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURC
E -D_FILE_OFFSET_BITS=64 -mlongcalls -mtext-section-literals -pipe -Os  -fvisibility=hidden -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -L/home/test/outputs/959b77fa2c1f13b1
958b234803437e09734e882e/output/host/usr/xtensa-buildroot-linux-uclibc/sysroot/usr/lib/mysql -L/home/test/outputs/959b77fa2c1f13b1958b234803437e09734e882e/output/host/usr/xtensa-buildroot-linux-
uclibc/sysroot/usr/lib -L/lib  -L/home/test/outputs/959b77fa2c1f13b1958b234803437e09734e882e/output/host/usr/xtensa-buildroot-linux-uclibc/sysroot/usr/lib -L/home/test/outputs/959b77fa2c1f13b195
8b234803437e09734e882e/output/host/usr/xtensa-buildroot-linux-uclibc/sysroot/usr conftest.c -lmysqlclient  -lz -lintl -lcrypto -lssl -lcrypto -lcurl -lz -lcrypto -lssl -lcrypto -lm  -lxml2 -lz -
lm -ldl -lcurl >&5
/lib/libgcc_s.so.1: file not recognized: File format not recognized
collect2: error: ld returned 1 exit status

The problem is due to the compilation command line containing -L/lib,
which is wrong when cross-compiling.

See also
https://gist.github.com/anonymous/10425853#file-gistfile1-txt-L8083 for
the full config.log file. I'm keeping the build output directory
around, in case additional informations are needed.

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] [autobuild.buildroot.net] Build results for 2014-04-07
  2014-04-10 21:41       ` Thomas Petazzoni
@ 2014-04-11  6:01         ` Max Filippov
  2014-04-11  7:51           ` Thomas Petazzoni
  0 siblings, 1 reply; 8+ messages in thread
From: Max Filippov @ 2014-04-11  6:01 UTC (permalink / raw)
  To: buildroot

Hi Thomas,

On Fri, Apr 11, 2014 at 1:41 AM, Thomas Petazzoni
<thomas.petazzoni@free-electrons.com> wrote:
> Dear Max Filippov,
>
> On Wed, 9 Apr 2014 01:51:08 +0400, Max Filippov wrote:
>
>> >> >     xtensa |                     php-5.5.11 | NOK | http://autobuild.buildroot.net/results/959b77fa2c1f13b1958b234803437e09734e882e/
>> >>
>> >> Couldn't reproduce. But I had to disable BR2_PACKAGE_HOST_SAM_BA,
>> >> otherwise it refused to build with the following message:
>> >>
>> >> Your Buildroot configuration uses pre-built tools for the x86 architecture,
>> >> but your build machine uses the x86-64 architecture without the 32
>> >> bits compatibility library.
>> >
>> > I think we can safely assume that the problem is not due to
>> > BR2_PACKAGE_HOST_SAM_BA. I'll try to reproduce the same build failure
>> > on the autobuilder, and gather the config.log file of php. I'll let you
>> > know.
>>
>> The tail of the failing build log:
>>
>> ...
>> checking for MySQL support... yes
>> checking for specified location of the MySQL UNIX socket... no
>> checking for mysql_close in -lmysqlclient... no
>> checking for mysql_error in -lmysqlclient... no
>> configure: error: mysql configure failed. Please check config.log for
>> more information.
>> make: *** [/home/test/test/2/output/build/php-5.5.11/.stamp_configured] Error 1
>
> Ok, I reproduced the problem. The config.log of php contains the
> following problem:
>
> configure:58794: result: no
> configure:59004: checking for mysql_error in -lmysqlclient
> configure:59029: /home/test/outputs/959b77fa2c1f13b1958b234803437e09734e882e/output/host/usr/bin/xtensa-buildroot-linux-uclibc-gcc -o conftest -I/include -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURC
> E -D_FILE_OFFSET_BITS=64 -mlongcalls -mtext-section-literals -pipe -Os  -fvisibility=hidden -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -L/home/test/outputs/959b77fa2c1f13b1
> 958b234803437e09734e882e/output/host/usr/xtensa-buildroot-linux-uclibc/sysroot/usr/lib/mysql -L/home/test/outputs/959b77fa2c1f13b1958b234803437e09734e882e/output/host/usr/xtensa-buildroot-linux-
> uclibc/sysroot/usr/lib -L/lib  -L/home/test/outputs/959b77fa2c1f13b1958b234803437e09734e882e/output/host/usr/xtensa-buildroot-linux-uclibc/sysroot/usr/lib -L/home/test/outputs/959b77fa2c1f13b195
> 8b234803437e09734e882e/output/host/usr/xtensa-buildroot-linux-uclibc/sysroot/usr conftest.c -lmysqlclient  -lz -lintl -lcrypto -lssl -lcrypto -lcurl -lz -lcrypto -lssl -lcrypto -lm  -lxml2 -lz -
> lm -ldl -lcurl >&5
> /lib/libgcc_s.so.1: file not recognized: File format not recognized
> collect2: error: ld returned 1 exit status
>
> The problem is due to the compilation command line containing -L/lib,
> which is wrong when cross-compiling.

Thanks for the information. I don't have /lib/libgcc_s.so.1 on my build system,
so it doesn't break here.
Apparently this happens when

BR2_PACKAGE_PHP_EXT_ICONV = y and
BR2_PACKAGE_LIBICONV is unset.

In this case buildroot calls

configure --with-iconv

for php, omitting LIBICONV dir, and php adds system dirs to -I and -L
(though I'm not sure it does that correctly).

Looks like we need to always select BR2_PACKAGE_LIBICONV when
BR2_PACKAGE_PHP_EXT_ICONV is selected, i.e.:

diff --git a/package/php/Config.ext b/package/php/Config.ext
index abf4993..1a46061 100644
--- a/package/php/Config.ext
+++ b/package/php/Config.ext
@@ -152,7 +152,7 @@ comment "Gettext support needs a toolchain w/ wchar"

 config BR2_PACKAGE_PHP_EXT_ICONV
        bool "iconv"
-       select BR2_PACKAGE_LIBICONV if !BR2_ENABLE_LOCALE
+       select BR2_PACKAGE_LIBICONV
        help
          iconv character set conversion support

diff --git a/package/php/php.mk b/package/php/php.mk
index ba1200b..8a147c5 100644
--- a/package/php/php.mk
+++ b/package/php/php.mk
@@ -108,12 +108,8 @@ ifeq ($(BR2_PACKAGE_PHP_EXT_GETTEXT),y)
 endif

 ifeq ($(BR2_PACKAGE_PHP_EXT_ICONV),y)
-ifeq ($(BR2_PACKAGE_LIBICONV),y)
        PHP_CONF_OPT += --with-iconv=$(STAGING_DIR)/usr
        PHP_DEPENDENCIES += libiconv
-else
-       PHP_CONF_OPT += --with-iconv
-endif
 endif

 ifeq ($(BR2_PACKAGE_PHP_EXT_INTL),y)

-- 
Thanks.
-- Max

^ permalink raw reply related	[flat|nested] 8+ messages in thread

* [Buildroot] [autobuild.buildroot.net] Build results for 2014-04-07
  2014-04-11  6:01         ` Max Filippov
@ 2014-04-11  7:51           ` Thomas Petazzoni
  2014-04-11  8:08             ` Max Filippov
  0 siblings, 1 reply; 8+ messages in thread
From: Thomas Petazzoni @ 2014-04-11  7:51 UTC (permalink / raw)
  To: buildroot

Dear Max Filippov,

On Fri, 11 Apr 2014 10:01:07 +0400, Max Filippov wrote:

> > The problem is due to the compilation command line containing -L/lib,
> > which is wrong when cross-compiling.
> 
> Thanks for the information. I don't have /lib/libgcc_s.so.1 on my build system,
> so it doesn't break here.

Ok.

> Apparently this happens when
> 
> BR2_PACKAGE_PHP_EXT_ICONV = y and
> BR2_PACKAGE_LIBICONV is unset.
> 
> In this case buildroot calls
> 
> configure --with-iconv
> 
> for php, omitting LIBICONV dir, and php adds system dirs to -I and -L
> (though I'm not sure it does that correctly).
> 
> Looks like we need to always select BR2_PACKAGE_LIBICONV when
> BR2_PACKAGE_PHP_EXT_ICONV is selected, i.e.:
> 
> diff --git a/package/php/Config.ext b/package/php/Config.ext
> index abf4993..1a46061 100644
> --- a/package/php/Config.ext
> +++ b/package/php/Config.ext
> @@ -152,7 +152,7 @@ comment "Gettext support needs a toolchain w/ wchar"
> 
>  config BR2_PACKAGE_PHP_EXT_ICONV
>         bool "iconv"
> -       select BR2_PACKAGE_LIBICONV if !BR2_ENABLE_LOCALE
> +       select BR2_PACKAGE_LIBICONV
>         help
>           iconv character set conversion support
> 
> diff --git a/package/php/php.mk b/package/php/php.mk
> index ba1200b..8a147c5 100644
> --- a/package/php/php.mk
> +++ b/package/php/php.mk
> @@ -108,12 +108,8 @@ ifeq ($(BR2_PACKAGE_PHP_EXT_GETTEXT),y)
>  endif
> 
>  ifeq ($(BR2_PACKAGE_PHP_EXT_ICONV),y)
> -ifeq ($(BR2_PACKAGE_LIBICONV),y)
>         PHP_CONF_OPT += --with-iconv=$(STAGING_DIR)/usr
>         PHP_DEPENDENCIES += libiconv
> -else
> -       PHP_CONF_OPT += --with-iconv
> -endif
>  endif
> 
>  ifeq ($(BR2_PACKAGE_PHP_EXT_INTL),y)

I don't think this is correct, because libiconv is only needed for
toolchains built without locale support. When the toolchain has locale
support (either uClibc with locale support, or glibc/eglibc), the iconv
functionality is provided directly by the C library.

And in fact, your change is going to break the build, because at the
end of libiconv.mk, we have:

# Configurations where the toolchain supports locales and the libiconv
# package is enabled are incorrect, because the toolchain already
# provides libiconv functionality, and having both confuses packages.
ifeq ($(BR2_PACKAGE_LIBICONV)$(BR2_ENABLE_LOCALE),yy)
$(error Libiconv should never be enabled when the toolchain supports locales. Report this failure to Buildroot developers)
endif

So with your change, if the toolchain has locale support, the libiconv
package will be selected, but it should not. I believe what should be
investigated instead is how to ensure that the C library iconv
functionality is properly detected by the PHP.

I'm also puzzled as to why this doesn't seem to happen for other
architecture. I would have to see if I can reproduce it on ARM or x86
for example. Maybe it's a weird uClibc bug, or something?

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] [autobuild.buildroot.net] Build results for 2014-04-07
  2014-04-11  7:51           ` Thomas Petazzoni
@ 2014-04-11  8:08             ` Max Filippov
  0 siblings, 0 replies; 8+ messages in thread
From: Max Filippov @ 2014-04-11  8:08 UTC (permalink / raw)
  To: buildroot

On Fri, Apr 11, 2014 at 11:51 AM, Thomas Petazzoni
<thomas.petazzoni@free-electrons.com> wrote:
>> Apparently this happens when
>>
>> BR2_PACKAGE_PHP_EXT_ICONV = y and
>> BR2_PACKAGE_LIBICONV is unset.
>>
>> In this case buildroot calls
>>
>> configure --with-iconv
>>
>> for php, omitting LIBICONV dir, and php adds system dirs to -I and -L
>> (though I'm not sure it does that correctly).
>>
>> Looks like we need to always select BR2_PACKAGE_LIBICONV when
>> BR2_PACKAGE_PHP_EXT_ICONV is selected, i.e.:

[...]

> I don't think this is correct, because libiconv is only needed for
> toolchains built without locale support. When the toolchain has locale
> support (either uClibc with locale support, or glibc/eglibc), the iconv
> functionality is provided directly by the C library.
>
> And in fact, your change is going to break the build

I see. Have run the test build which failed.

> I believe what should be
> investigated instead is how to ensure that the C library iconv
> functionality is properly detected by the PHP.

Ok, I'll look deeper at that ext/iconv/config.m4 and stuff around...

> I'm also puzzled as to why this doesn't seem to happen for other
> architecture. I would have to see if I can reproduce it on ARM or x86
> for example. Maybe it's a weird uClibc bug, or something?

-- 
Thanks.
-- Max

^ permalink raw reply	[flat|nested] 8+ messages in thread

end of thread, other threads:[~2014-04-11  8:08 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-04-08  6:30 [Buildroot] [autobuild.buildroot.net] Build results for 2014-04-07 Thomas Petazzoni
2014-04-08 21:00 ` Max Filippov
2014-04-08 21:41   ` Thomas Petazzoni
2014-04-08 21:51     ` Max Filippov
2014-04-10 21:41       ` Thomas Petazzoni
2014-04-11  6:01         ` Max Filippov
2014-04-11  7:51           ` Thomas Petazzoni
2014-04-11  8:08             ` Max Filippov

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox