[Buildroot] [PATCH 2/2] b2g: new package
François Perrad
francois.perrad at gadz.org
Sat Jul 16 19:48:42 UTC 2016
2016-07-16 15:01 GMT+02:00 Thomas Petazzoni
<thomas.petazzoni at free-electrons.com>:
> Hello,
>
> On Sat, 16 Jul 2016 12:00:17 +0200, Francois Perrad wrote:
>> Signed-off-by: Francois Perrad <francois.perrad at gadz.org>
>
> Thanks for this patch. It looks mostly good, but I have some questions.
>
> First of all, the obvious annoying aspect is the need of autoconf2.13.
> Do they have some plans to update to a newer decent version of
> autoconf, or would we have to keep autoconf2.13 forever?
>
Mozilla will stay with autoconf2.13.
FYI, Debian have a package autoconf2.13
>> package/Config.in | 1 +
>> package/b2g/0001-fix-linking.patch | 28 +++++++++
>> package/b2g/Config.in | 29 +++++++++
>> package/b2g/b2g.mk | 125 +++++++++++++++++++++++++++++++++++++
>
> Should the package be named boot2gecko, in order to be more explicit
> than b2g ? I'm not sure, since indeed from upstream's point of view the
> name is b2g. Maybe we should keep it.
>
I just follow the upstream.
>> diff --git a/package/b2g/0001-fix-linking.patch b/package/b2g/0001-fix-linking.patch
>> new file mode 100644
>> index 0000000..5433160
>> --- /dev/null
>> +++ b/package/b2g/0001-fix-linking.patch
>> @@ -0,0 +1,28 @@
>> +fix linking
>> +
>> +b2g-2_5_20160125/dom/audiochannel/AudioChannelService.cpp:580: error:
>> + undefined reference to 'mozilla::dom::TabParent::AudioChannelChangeNotification(nsPIDOMWindow*, mozilla::dom::AudioChannel, float, bool)'
>> +collect2: error: ld returned 1 exit status
>> +
>> +Signed-off-by: Francois Perrad <francois.perrad at gadz.org>
>> +
>> +diff --git a/dom/audiochannel/AudioChannelService.cpp b/dom/audiochannel/AudioChannelService.cpp
>> +index 7c9593e..cd9e1a5 100644
>> +--- a/dom/audiochannel/AudioChannelService.cpp
>> ++++ b/dom/audiochannel/AudioChannelService.cpp
>> +@@ -574,11 +574,13 @@ AudioChannelService::RefreshAgentsVolumeAndPropagate(AudioChannel aAudioChannel,
>> + return;
>> + }
>> +
>> ++#if 0
>> + for (uint32_t i = 0; i < mTabParents.Length(); ++i) {
>> + mTabParents[i]->AudioChannelChangeNotification(aWindow, aAudioChannel,
>> + winData->mChannels[(uint32_t)aAudioChannel].mVolume,
>> + winData->mChannels[(uint32_t)aAudioChannel].mMuted);
>> + }
>> ++#endif
>
> You're just removing the problematic code. Is this the right fix?
Just a quick & dirty fix, as first step.
>
>> diff --git a/package/b2g/Config.in b/package/b2g/Config.in
>> new file mode 100644
>> index 0000000..6c932de
>> --- /dev/null
>> +++ b/package/b2g/Config.in
>> @@ -0,0 +1,29 @@
>> +config BR2_PACKAGE_B2G
>> + bool "b2g"
>> + select BR2_PACKAGE_BZIP2
>> + select BR2_PACKAGE_CAIRO
>> + select BR2_PACKAGE_CAIRO_PDF
>> + select BR2_PACKAGE_CAIRO_TEE
>> + select BR2_PACKAGE_JPEG
>> + select BR2_PACKAGE_LIBCURL
>> + select BR2_PACKAGE_LIBEVENT
>> + select BR2_PACKAGE_LIBGTK2
>> + select BR2_PACKAGE_LIBVPX
>> + select BR2_PACKAGE_OPENSSL # libcurl4-openssl
>> + select BR2_PACKAGE_XLIB_LIBXT
>> + depends on BR2_PACKAGE_XORG7
>> + depends on BR2_INSTALL_LIBSTDCPP # libgtk2
>> + depends on BR2_TOOLCHAIN_HAS_SYNC_4 # libgtk2
>> + depends on BR2_TOOLCHAIN_HAS_THREADS # libgtk2
>
> Also needed because of the libvpx select.
OK
>
>> + depends on BR2_USE_MMU # libgtk2
>> + depends on BR2_USE_WCHAR # libgtk2
>> + help
>> + B2G (Boot 2 Gecko), the Firefox OS desktop client.
>> +
>> + https://developer.mozilla.org/en-US/Firefox_OS
>> +
>> +comment "b2g needs a toolchain w/wchar, threads, C++"
>
> Space between w/ and wchar.
OK
>
>> + depends on BR2_PACKAGE_XORG7
>> + depends on BR2_TOOLCHAIN_HAS_SYNC_4
>> + depends on BR2_USE_MMU
>> + depends on !BR2_USE_WCHAR || !BR2_TOOLCHAIN_HAS_THREADS || !BR2_INSTALL_LIBSTDCPP
>> diff --git a/package/b2g/b2g.mk b/package/b2g/b2g.mk
>> new file mode 100644
>> index 0000000..172b062
>> --- /dev/null
>> +++ b/package/b2g/b2g.mk
>> @@ -0,0 +1,125 @@
>> +################################################################################
>> +#
>> +# b2g
>> +#
>> +################################################################################
>> +
>> +B2G_VERSION = 2_5_20160125
>> +B2G_SITE = https://hg.mozilla.org/releases/mozilla-b2g44_v2_5/archive
>> +B2G_SOURCE = B2G_$(B2G_VERSION)_MERGEDAY.tar.gz
>> +B2G_LICENSE = MPLv2.0
>> +B2G_LICENSE_FILES = toolkit/content/license.html
>> +
>> +B2G_DEPENDENCIES = host-autoconf2-13 host-python jpeg libcurl libevent libgtk2 libvpx xlib_libXt
>> +
>> +define B2G_MOZCONFIG_COMMON
>> + echo "ac_add_options --prefix=/usr" > $(@D)/mozconfig
>> + echo "ac_add_options --target=$(GNU_TARGET_NAME)" >> $(@D)/mozconfig
>> + echo "ac_add_options --host=$(GNU_HOST_NAME)" >> $(@D)/mozconfig
>> + echo "ac_add_options --build=$(GNU_HOST_NAME)" >> $(@D)/mozconfig
>> + echo "ac_add_options --enable-application=b2g" >> $(@D)/mozconfig
>> + echo "ac_add_options --enable-xterm-updates" >> $(@D)/mozconfig
>> + echo "ac_add_options --disable-debug" >> $(@D)/mozconfig
>> + echo "ac_add_options --disable-tests" >> $(@D)/mozconfig
>> + echo "ac_add_options --enable-optimize" >> $(@D)/mozconfig
>> + echo "ac_add_options --enable-mobile-optimize" >> $(@D)/mozconfig
>> + echo "ac_add_options --disable-strip" >> $(@D)/mozconfig
>> + echo "ac_add_options --disable-install-strip" >> $(@D)/mozconfig
>> + echo "ac_add_options --disable-gconf" >> $(@D)/mozconfig
>> + echo "ac_add_options --with-system-bz2" >> $(@D)/mozconfig
>> + echo "ac_add_options --with-system-jpeg" >> $(@D)/mozconfig
>> + echo "ac_add_options --with-system-libevent" >> $(@D)/mozconfig
>> + echo "ac_add_options --with-system-libvpx" >> $(@D)/mozconfig
>> + echo "ac_add_options --with-system-zlib" >> $(@D)/mozconfig
>> + echo "ac_add_options --enable-system-cairo" >> $(@D)/mozconfig
>> + echo "ac_add_options --enable-system-ffi" >> $(@D)/mozconfig
>
> Since you specify --enable-system-ffi, I would have expected a
> dependency on libffi. Did I miss something?
OK
>
>> +ifeq ($(BR2_PACKAGE_GST1_PLUGINS_BASE),y)
>> +B2G_DEPENDENCIES += gst1-plugins-base
>> +define B2G_MOZCONFIG_GSTREAMER
>> + echo "ac_add_options --enable-gstreamer=1.0" >> $(@D)/mozconfig
>> +endef
>> +else
>> +ifeq ($(BR2_PACKAGE_GST_PLUGINS_BASE),y)
>
> Put the
>
> else ifeq ...
>
> on the same line...
OK
>
>> +B2G_DEPENDENCIES += gst-plugins-base
>> +define B2G_MOZCONFIG_GSTREAMER
>> + echo "ac_add_options --enable-gstreamer=0.10" >> $(@D)/mozconfig
>> +endef
>> +else
>> +define B2G_MOZCONFIG_GSTREAMER
>> + echo "ac_add_options --disable-gstreamer" >> $(@D)/mozconfig
>> +endef
>> +endif
>> +endif
>
> ... as it allows to remove one endif:
>
> ifeq ...
> ...
> else ifeq ...
> ...
> endif
>
>> +ifeq ($(BR2_PACKAGE_ICU),y)
>> +B2G_DEPENDENCIES += icu
>> +define B2G_MOZCONFIG_ICU
>> + echo "ac_add_options --with-intl-api" >> $(@D)/mozconfig
>> + echo "ac_add_options --with-system-icu" >> $(@D)/mozconfig
>> +endef
>> +else
>> +define B2G_MOZCONFIG_ICU
>> + echo "ac_add_options --without-intl-api" >> $(@D)/mozconfig
>> +endef
>> +endif
>
> This whole thing seems a bit redundant. What about instead:
>
> ifeq ...
> B2G_CONF_OPTS += --with-intl-api --with-system-icu
> else
> B2G_CONF_OPTS += --without-intl-api
> endif
>
> And then later on:
>
> $(foreach opt,$(B2G_CONF_OPTS),\
> echo "ac_add_options $(opt)" >> $(@D)/mozconfig
> )
>
Great improvement
> It avoids having to duplicate the echo "ac_add_options ..." >>
> $(@D)/mozconfig numerous times.
>
>> +
>> +B2G_CONF_ENV = \
>> + MOZ_OBJDIR=$(@D)/obj-b2g \
>> + CROSS_COMPILE=1 \
>> + AUTOCONF=$(HOST_DIR)/usr/bin/autoconf2.13 \
>> + PERL=$(shell which perl) \
>
> A priori this should not be needed, since package/Makefile.in exports
> PERL with the exact same value. Can you double check if it's really
> needed?
OK
>
>> + PYTHON=$(HOST_DIR)/usr/bin/python
>> +
>> +define B2G_CONFIGURE_CMDS
>> + $(B2G_MOZCONFIG_COMMON)
>> + $(B2G_MOZCONFIG_ALSA)
>> + $(B2G_MOZCONFIG_DBUS)
>> + $(B2G_MOZCONFIG_GSTREAMER)
>> + $(B2G_MOZCONFIG_ICU)
>> + $(TARGET_CONFIGURE_OPTS) \
>> + $(B2G_CONF_ENV) \
>
> Continuation lines should be indented with one more tab than the first
> line:
OK
>
>> + $(MAKE1) -C $(@D) -f client.mk configure
>
> $(TARGET_CONFIGURE_OPTS) $(BG2_CONF_ENV)
> $(MAKE1) -C $(@D) -f client.mk configure
>
>> +endef
>> +
>> +define B2G_BUILD_CMDS
>> + $(MAKE1) -C $(@D) -f client.mk build_all \
>> + MOZ_OBJDIR=$(@D)/obj-b2g
>> + $(MAKE1) -C $(@D)/obj-b2g package
>> +endef
>> +
>> +define B2G_INSTALL_TARGET_CMDS
>> + rm -rf $(TARGET_DIR)/usr/lib/b2g
>
> We normally don't remove stuff from the target prior to the
> installation.
OK
François
>
>> + tar xfj $(@D)/obj-b2g/dist/b2g-*.tar.bz2 -C $(TARGET_DIR)/usr/lib
>> + ln -sf ../lib/b2g/b2g $(TARGET_DIR)/usr/bin
>> +endef
>> +
>> +$(eval $(generic-package))
>
> Thanks,
>
> Thomas
> --
> Thomas Petazzoni, CTO, Free Electrons
> Embedded Linux, Kernel and Android engineering
> http://free-electrons.com
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot
More information about the buildroot
mailing list