[Buildroot] [PATCH 1/3] package/libevas-generic-loaders: needs zlib

[Buildroot] [PATCH 1/3] package/libevas-generic-loaders: needs zlib

[Romain Naour]

Fixes:
http://autobuild.buildroot.net/results/5d4/5d41cec63a15dc5f31460c8cecd6054119ddb1fc/build-end.log

Signed-off-by: Romain Naour <romain.naour@openwide.fr>
---
 package/efl/libevas-generic-loaders/Config.in                  | 1 +
 package/efl/libevas-generic-loaders/libevas-generic-loaders.mk | 2 +-
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/package/efl/libevas-generic-loaders/Config.in b/package/efl/libevas-generic-loaders/Config.in
index 30fb88b..3f7581f 100644
--- a/package/efl/libevas-generic-loaders/Config.in
+++ b/package/efl/libevas-generic-loaders/Config.in
@@ -1,6 +1,7 @@
 config BR2_PACKAGE_LIBEVAS_GENERIC_LOADERS
 	bool "libevas generic loaders"
 	select BR2_PACKAGE_LIBEINA
+	select BR2_PACKAGE_ZLIB
 	help
 	  These are additional "generic" loaders for Evas that are
 	  stand-alone executables that evas may run from its generic
diff --git a/package/efl/libevas-generic-loaders/libevas-generic-loaders.mk b/package/efl/libevas-generic-loaders/libevas-generic-loaders.mk
index 152a75b..db0ca32 100644
--- a/package/efl/libevas-generic-loaders/libevas-generic-loaders.mk
+++ b/package/efl/libevas-generic-loaders/libevas-generic-loaders.mk
@@ -12,7 +12,7 @@ LIBEVAS_GENERIC_LOADERS_LICENSE_FILES = COPYING
 
 LIBEVAS_GENERIC_LOADERS_INSTALL_STAGING = YES
 
-LIBEVAS_GENERIC_LOADERS_DEPENDENCIES = libeina
+LIBEVAS_GENERIC_LOADERS_DEPENDENCIES = libeina zlib
 
 # For now, we only support the SVG loader
 LIBEVAS_GENERIC_LOADERS_CONF_OPT += \
-- 
1.9.3

[Buildroot] [PATCH 2/3] package/libinput: fix build with libevdev

[Romain Naour]

libevdev and libinput use the same function name (log_msg).
An easy fix is to rename one of the two functions.

Fixes:
http://autobuild.buildroot.net/results/b1e/b1ea94f09d6f7459e3d7794c4a62bbdbb53d5da6/build-end.log

Signed-off-by: Romain Naour <romain.naour@openwide.fr>
---
 ...put-01-rename-log_msg-to-libinput_log_msg.patch | 58 ++++++++++++++++++++++
 1 file changed, 58 insertions(+)
 create mode 100644 package/libinput/libinput-01-rename-log_msg-to-libinput_log_msg.patch

diff --git a/package/libinput/libinput-01-rename-log_msg-to-libinput_log_msg.patch b/package/libinput/libinput-01-rename-log_msg-to-libinput_log_msg.patch
new file mode 100644
index 0000000..55604b5
--- /dev/null
+++ b/package/libinput/libinput-01-rename-log_msg-to-libinput_log_msg.patch
@@ -0,0 +1,58 @@
+From 89a3ae2de21e528b8feb80dbea3928d43982d3be Mon Sep 17 00:00:00 2001
+From: Romain Naour <romain.naour@openwide.fr>
+Date: Thu, 21 Aug 2014 18:18:16 +0200
+Subject: [PATCH 1/1] rename log_msg to libinput_log_msg
+
+This fix conflicts between libevdev and libinput
+
+http://autobuild.buildroot.net/results/c13/c133b7c706ee31302125df8ca94f4d0f0152c6c6/build-end.log
+
+Signed-off-by: Romain Naour <romain.naour@openwide.fr>
+---
+ src/libinput-private.h | 14 +++++++-------
+ src/libinput.c         |  2 +-
+ 2 files changed, 8 insertions(+), 8 deletions(-)
+
+diff --git a/src/libinput-private.h b/src/libinput-private.h
+index f9d2f87..ac1e1c4 100644
+--- a/src/libinput-private.h
++++ b/src/libinput-private.h
+@@ -81,15 +81,15 @@ typedef void (*libinput_source_dispatch_t)(void *data);
+ 
+ struct libinput_source;
+ 
+-#define log_debug(...) log_msg(LIBINPUT_LOG_PRIORITY_DEBUG, __VA_ARGS__)
+-#define log_info(...) log_msg(LIBINPUT_LOG_PRIORITY_INFO, __VA_ARGS__)
+-#define log_error(...) log_msg(LIBINPUT_LOG_PRIORITY_ERROR, __VA_ARGS__)
+-#define log_bug_kernel(...) log_msg(LIBINPUT_LOG_PRIORITY_ERROR, "kernel bug: " __VA_ARGS__)
+-#define log_bug_libinput(...) log_msg(LIBINPUT_LOG_PRIORITY_ERROR, "libinput bug: " __VA_ARGS__);
+-#define log_bug_client(...) log_msg(LIBINPUT_LOG_PRIORITY_ERROR, "client bug: " __VA_ARGS__);
++#define log_debug(...) libinput_log_msg(LIBINPUT_LOG_PRIORITY_DEBUG, __VA_ARGS__)
++#define log_info(...) libinput_log_msg(LIBINPUT_LOG_PRIORITY_INFO, __VA_ARGS__)
++#define log_error(...) libinput_log_msg(LIBINPUT_LOG_PRIORITY_ERROR, __VA_ARGS__)
++#define log_bug_kernel(...) libinput_log_msg(LIBINPUT_LOG_PRIORITY_ERROR, "kernel bug: " __VA_ARGS__)
++#define log_bug_libinput(...) libinput_log_msg(LIBINPUT_LOG_PRIORITY_ERROR, "libinput bug: " __VA_ARGS__);
++#define log_bug_client(...) libinput_log_msg(LIBINPUT_LOG_PRIORITY_ERROR, "client bug: " __VA_ARGS__);
+ 
+ void
+-log_msg(enum libinput_log_priority priority, const char *format, ...);
++libinput_log_msg(enum libinput_log_priority priority, const char *format, ...);
+ 
+ int
+ libinput_init(struct libinput *libinput,
+diff --git a/src/libinput.c b/src/libinput.c
+index eec9efb..4e3f6fa 100644
+--- a/src/libinput.c
++++ b/src/libinput.c
+@@ -110,7 +110,7 @@ static struct log_data log_data = {
+ };
+ 
+ void
+-log_msg(enum libinput_log_priority priority, const char *format, ...)
++libinput_log_msg(enum libinput_log_priority priority, const char *format, ...)
+ {
+ 	va_list args;
+ 
+-- 
+1.9.3
+
-- 
1.9.3

[Buildroot] [PATCH 2/3] package/libinput: fix build with libevdev

[Thomas Petazzoni]

Dear Romain Naour,

On Sun, 24 Aug 2014 18:53:05 +0200, Romain Naour wrote:
> libevdev and libinput use the same function name (log_msg).
> An easy fix is to rename one of the two functions.
> 
> Fixes:
> http://autobuild.buildroot.net/results/b1e/b1ea94f09d6f7459e3d7794c4a62bbdbb53d5da6/build-end.log
> 
> Signed-off-by: Romain Naour <romain.naour@openwide.fr>
> ---
>  ...put-01-rename-log_msg-to-libinput_log_msg.patch | 58 ++++++++++++++++++++++
>  1 file changed, 58 insertions(+)
>  create mode 100644 package/libinput/libinput-01-rename-log_msg-to-libinput_log_msg.patch

Committed, but please submit upstream.

Thanks!

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

[Buildroot] [PATCH 2/3] package/libinput: fix build with libevdev

[Romain Naour]

Hi Thomas,

Le 25/08/2014 20:12, Thomas Petazzoni a ?crit :
> Dear Romain Naour,
> 
> On Sun, 24 Aug 2014 18:53:05 +0200, Romain Naour wrote:
>> libevdev and libinput use the same function name (log_msg).
>> An easy fix is to rename one of the two functions.
>>
>> Fixes:
>> http://autobuild.buildroot.net/results/b1e/b1ea94f09d6f7459e3d7794c4a62bbdbb53d5da6/build-end.log
>>
>> Signed-off-by: Romain Naour <romain.naour@openwide.fr>
>> ---
>>  ...put-01-rename-log_msg-to-libinput_log_msg.patch | 58 ++++++++++++++++++++++
>>  1 file changed, 58 insertions(+)
>>  create mode 100644 package/libinput/libinput-01-rename-log_msg-to-libinput_log_msg.patch
> 
> Committed, but please submit upstream.
> 

I'll do.

Best regards,
Romain

[Buildroot] [PATCH 2/3] package/libinput: fix build with libevdev

[Vicente Olivert Riera]

On 08/25/2014 10:01 PM, Romain Naour wrote:
> Hi Thomas,
>
> Le 25/08/2014 20:12, Thomas Petazzoni a ?crit :
>> Dear Romain Naour,
>>
>> On Sun, 24 Aug 2014 18:53:05 +0200, Romain Naour wrote:
>>> libevdev and libinput use the same function name (log_msg).
>>> An easy fix is to rename one of the two functions.
>>>
>>> Fixes:
>>> http://autobuild.buildroot.net/results/b1e/b1ea94f09d6f7459e3d7794c4a62bbdbb53d5da6/build-end.log
>>>
>>> Signed-off-by: Romain Naour <romain.naour@openwide.fr>
>>> ---
>>>   ...put-01-rename-log_msg-to-libinput_log_msg.patch | 58 ++++++++++++++++++++++
>>>   1 file changed, 58 insertions(+)
>>>   create mode 100644 package/libinput/libinput-01-rename-log_msg-to-libinput_log_msg.patch
>>
>> Committed, but please submit upstream.
>>
>
> I'll do.
>
> Best regards,
> Romain
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot
>

Some changes to fix this kind of issues have been applied upstream. 
Please have a look to:

https://bugs.freedesktop.org/show_bug.cgi?id=82292

Maybe it's worth to do a version bump and check if that new patch is no 
longer necessary.

-- 
Vincent

[Buildroot] [PATCH 2/3] package/libinput: fix build with libevdev

[Romain Naour]

Hi Vicente,

Le 26/08/2014 11:13, Vicente Olivert Riera a ?crit :
> On 08/25/2014 10:01 PM, Romain Naour wrote:
>> Hi Thomas,
>>
>> Le 25/08/2014 20:12, Thomas Petazzoni a ?crit :
>>> Dear Romain Naour,
>>>
>>> On Sun, 24 Aug 2014 18:53:05 +0200, Romain Naour wrote:
>>>> libevdev and libinput use the same function name (log_msg).
>>>> An easy fix is to rename one of the two functions.
>>>>
>>>> Fixes:
>>>> http://autobuild.buildroot.net/results/b1e/b1ea94f09d6f7459e3d7794c4a62bbdbb53d5da6/build-end.log
>>>>
>>>>
>>>> Signed-off-by: Romain Naour <romain.naour@openwide.fr>
>>>> ---
>>>>   ...put-01-rename-log_msg-to-libinput_log_msg.patch | 58
>>>> ++++++++++++++++++++++
>>>>   1 file changed, 58 insertions(+)
>>>>   create mode 100644
>>>> package/libinput/libinput-01-rename-log_msg-to-libinput_log_msg.patch
>>>
>>> Committed, but please submit upstream.
>>>
>>
>> I'll do.
>>
>> Best regards,
>> Romain
>> _______________________________________________
>> buildroot mailing list
>> buildroot at busybox.net
>> http://lists.busybox.net/mailman/listinfo/buildroot
>>
> 
> Some changes to fix this kind of issues have been applied upstream. Please have
> a look to:
> 
> https://bugs.freedesktop.org/show_bug.cgi?id=82292

Thanks for the link !

I should have looked at the latest commits...

> 
> Maybe it's worth to do a version bump and check if that new patch is no longer
> necessary.
> 

Yes, I reverted my commit and I built libinput-0.3.0 using a backported version
of this patch:
http://lists.freedesktop.org/archives/wayland-devel/2014-August/016599.html

Also, libevdev has renamed the functions log_msg to _libevdev_log_msg.
So, the issue will be fixed without my patch in the next version bump of
libevdev or libinput.

Best regards,
Romain

[Buildroot] [PATCH 3/3] package/libarchive: fix linking issue with uClibc

[Romain Naour]

With uClibc the dynamic linking loader functions needs
to be linked with -ldl

Fixes:
http://autobuild.buildroot.net/results/6af/6af140618b274536238b062635d493709905c8f6/build-end.log

Signed-off-by: Romain Naour <romain.naour@openwide.fr>
---
 package/libarchive/libarchive.mk | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/package/libarchive/libarchive.mk b/package/libarchive/libarchive.mk
index 8c6951d..ce3355c 100644
--- a/package/libarchive/libarchive.mk
+++ b/package/libarchive/libarchive.mk
@@ -68,4 +68,8 @@ else
 LIBARCHIVE_CONF_OPT += --without-zlib
 endif
 
+ifeq ($(BR2_TOOLCHAIN_USES_UCLIBC),y)
+LIBARCHIVE_CONF_ENV += LIBS="-ldl"
+endif
+
 $(eval $(autotools-package))
-- 
1.9.3

[Buildroot] [PATCH 3/3] package/libarchive: fix linking issue with uClibc

[Thomas Petazzoni]

Dear Romain Naour,

On Sun, 24 Aug 2014 18:53:06 +0200, Romain Naour wrote:
> With uClibc the dynamic linking loader functions needs
> to be linked with -ldl
> 
> Fixes:
> http://autobuild.buildroot.net/results/6af/6af140618b274536238b062635d493709905c8f6/build-end.log
> 
> Signed-off-by: Romain Naour <romain.naour@openwide.fr>
> ---
>  package/libarchive/libarchive.mk | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/package/libarchive/libarchive.mk b/package/libarchive/libarchive.mk
> index 8c6951d..ce3355c 100644
> --- a/package/libarchive/libarchive.mk
> +++ b/package/libarchive/libarchive.mk
> @@ -68,4 +68,8 @@ else
>  LIBARCHIVE_CONF_OPT += --without-zlib
>  endif
>  
> +ifeq ($(BR2_TOOLCHAIN_USES_UCLIBC),y)
> +LIBARCHIVE_CONF_ENV += LIBS="-ldl"
> +endif
> +
>  $(eval $(autotools-package))

I don't think this is the right fix, for several reasons:

 *) The problem only occurs due to the fact that bsdcpio and bsdtar are
    by default linked statically. When BR2_PREFER_STATIC_LIB is
    disabled, we should pass --enable-bsdtar=shared and
    --enable-bsdcpio=shared to ensure they are dynamically linked. This
    fixes the problem in the dynamic linking case.

 *) The problem is caused by the dependency of OpenSSL on libdl. So
    there is no need to always link against libdl: it should only be
    done when BR2_PREFER_STATIC_LIB=y and OpenSSL is enabled.

 *) I don't see why the problem would be uClibc specific.

Generally speaking, it's annoying to have to teach all packages
using OpenSSL that they should link against -ldl when doing static
linking. It would be so much better if everyone was using pkg-config...

Best regards,

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

[Buildroot] [PATCH 3/3] package/libarchive: fix linking issue with uClibc

[Romain Naour]

Hi Thomas,

Le 25/08/2014 23:03, Thomas Petazzoni a ?crit :
> Dear Romain Naour,
> 
> On Sun, 24 Aug 2014 18:53:06 +0200, Romain Naour wrote:
>> With uClibc the dynamic linking loader functions needs
>> to be linked with -ldl
>>
>> Fixes:
>> http://autobuild.buildroot.net/results/6af/6af140618b274536238b062635d493709905c8f6/build-end.log
>>
>> Signed-off-by: Romain Naour <romain.naour@openwide.fr>
>> ---
>>  package/libarchive/libarchive.mk | 4 ++++
>>  1 file changed, 4 insertions(+)
>>
>> diff --git a/package/libarchive/libarchive.mk b/package/libarchive/libarchive.mk
>> index 8c6951d..ce3355c 100644
>> --- a/package/libarchive/libarchive.mk
>> +++ b/package/libarchive/libarchive.mk
>> @@ -68,4 +68,8 @@ else
>>  LIBARCHIVE_CONF_OPT += --without-zlib
>>  endif
>>  
>> +ifeq ($(BR2_TOOLCHAIN_USES_UCLIBC),y)
>> +LIBARCHIVE_CONF_ENV += LIBS="-ldl"
>> +endif
>> +
>>  $(eval $(autotools-package))
> 
> I don't think this is the right fix, for several reasons:
> 
>  *) The problem only occurs due to the fact that bsdcpio and bsdtar are
>     by default linked statically. When BR2_PREFER_STATIC_LIB is
>     disabled, we should pass --enable-bsdtar=shared and
>     --enable-bsdcpio=shared to ensure they are dynamically linked. This
>     fixes the problem in the dynamic linking case.
> 
>  *) The problem is caused by the dependency of OpenSSL on libdl. So
>     there is no need to always link against libdl: it should only be
>     done when BR2_PREFER_STATIC_LIB=y and OpenSSL is enabled.
> 
>  *) I don't see why the problem would be uClibc specific.
> 
> Generally speaking, it's annoying to have to teach all packages
> using OpenSSL that they should link against -ldl when doing static
> linking. It would be so much better if everyone was using pkg-config...
> 

Ok, I misunderstood the problem, thanks for your explanations.
I'll try to fix this issue before the release.

Best regards,
Romain

[Buildroot] [PATCH 1/3] package/libevas-generic-loaders: needs zlib

[Thomas Petazzoni]

Dear Romain Naour,

On Sun, 24 Aug 2014 18:53:04 +0200, Romain Naour wrote:
> Fixes:
> http://autobuild.buildroot.net/results/5d4/5d41cec63a15dc5f31460c8cecd6054119ddb1fc/build-end.log
> 
> Signed-off-by: Romain Naour <romain.naour@openwide.fr>
> ---
>  package/efl/libevas-generic-loaders/Config.in                  | 1 +
>  package/efl/libevas-generic-loaders/libevas-generic-loaders.mk | 2 +-
>  2 files changed, 2 insertions(+), 1 deletion(-)

Applied, thanks.

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