Re: [PATCH 01/17] gettext.bbclass: Use _append instead of =+

From: Saul Wold <sgw@linux.intel.com>
To: Patches and discussions about the oe-core layer
	<openembedded-core@lists.openembedded.org>
Subject: Re: [PATCH 01/17] gettext.bbclass: Use _append instead of	=+
Date: Wed, 27 Apr 2011 17:53:49 -0700	[thread overview]
Message-ID: <4DB8BA9D.3010806@linux.intel.com> (raw)
In-Reply-To: <BANLkTim7z6b+k_x3wU8=8nHiKVm-FyoX6g@mail.gmail.com>

On 04/27/2011 04:59 PM, Khem Raj wrote:
> On Wed, Apr 27, 2011 at 3:51 PM, Saul Wold<sgw@linux.intel.com>  wrote:
>> On 04/26/2011 03:17 PM, Khem Raj wrote:
>>>
>>> On Tue, Apr 26, 2011 at 1:02 PM, Khem Raj<raj.khem@gmail.com>    wrote:
>>>>
>>>> On Tue, Apr 26, 2011 at 12:48 PM, Saul Wold<sgw@linux.intel.com>    wrote:
>>>>>
>>>>> On 04/25/2011 11:54 AM, Khem Raj wrote:
>>>>>>
>>>>>> Ensure gettext and gettext-native are removed from DEPENDS when
>>>>>> not using NLS
>>>>>>
>>>>>> Use append instead of += to get gettext dependecies processed
>>>>>> correctly in all cases
>>>>>>
>>>>>> Dont remove gettext-native for native recipes as ENABLE_NLS is
>>>>>> only for target and not for native recipes
>>>>>>
>>>>>> Replace using 1 for a boolean type with True
>>>>>>
>>>>>> Honor INHIBIT_DEFAULT_DEPS
>>>>>>
>>>>>> Remove the added dependencies for gettext if INHIBIT_DEFAULT_DEPS is
>>>>>> non
>>>>>> null
>>>>>>
>>>>> Khem,
>>>>>
>>>>> This patch seems to be causing a circular dependency when you try to
>>>>> build
>>>>> meta-toolchain-sdk, not sure if you tested that target.
>>>>>
>>>>
>>>> No I did not. Let me try it out.
>>
>> Khem,
>>
>> There still seems to be a problem with binutil-cross-canadian and
>> gcc-cross-canadian not finding the gettext-native, it suggests to have these
>> recipe inherit gettext, which binutils does via it's .inc file, but
>> gcc-cross-canadian did not seem to do.  I tried to add inherit gettext to
>> the cross-canadian.bbclass but that did not solve the issue.
>>
>
> Does it break the build ?
> this may be a new issue unearthed. We can fix it with subsequent patch
>
Not breaking the mainline build, but it is breaking the toolchain builds.

Sau!

>> Currently the meta-toolchain gmae builds show this problem.
>>
>
> I will try to see whats going on here.
>
>> Thanks
>>         Sau!
>>
>>>>
>>>>> % bitbake meta-toolchain-sdk -n
>>>>> NOTE: Out of date cache found, rebuilding...
>>>>> WARNING: No bb files matched BBFILE_PATTERN_yocto
>>>>> '^/intel/poky/distro/meta-yocto/'
>>>>> Parsing recipes: 100% |#########################################| Time:
>>>>> 00:00:15
>>>>> Parsing of 796 .bb files complete (0 cached, 796 parsed). 1008 targets,
>>>>> 6
>>>>> skipped, 0 masked, 0 errors.
>>>>>
>>>>> OE Build Configuration:
>>>>> BB_VERSION        = "1.11.0"
>>>>> METADATA_BRANCH   = "stage"
>>>>> METADATA_REVISION = "86f12953dbf57264cfcc06feac446f3edf0c6d8f"
>>>>> TARGET_ARCH       = "i586"
>>>>> TARGET_OS         = "linux"
>>>>> MACHINE           = "qemux86"
>>>>> DISTRO            = "poky-lsb"
>>>>> DISTRO_VERSION    = "1.0+snapshot-20110426"
>>>>> TARGET_FPU        = ""
>>>>>
>>>>> NOTE: Resolving any missing task queue dependencies
>>>>> NOTE: Preparing runqueue
>>>>> ERROR: Task
>>>>>
>>>>> virtual:nativesdk:/intel/poky/distro/meta/recipes-core/gettext/gettext_0.17.bb
>>>>> (do_package) has circular dependency on
>>>>>
>>>>> virtual:nativesdk:/intel/poky/distro/meta/recipes-devtools/gcc/gcc-runtime_4.5.1.bb
>>>>> (do_package)
>>>>>
>>>>> Is there dependencies on this patch with the other patches? (ie can I
>>>>> take
>>>>> part of your other patches or just hold off)?
>>>>>
>>>>
>>>> This patch is kind of needed for few of them for parsing sake.
>>>> some of them dont depend on this e.g.
>>>>
>>>> [PATCH 05/17] sstate.bbclass: Reduce SSTATE_MANIFESTS assignment to be
>>>> weak
>>>> [PATCH 09/17] libiconv: update from 1.9.2 ->    1.11.1
>>>> [PATCH 12/17] elfutils_0.148.bb: Fix compilation issues on uclibc
>>>> [PATCH 13/17] perl_5.12.2.bb: Undefine features not found in uclibc
>>>> [PATCH 14/17] xorg-lib: inherit gettext class instead of adding
>>>> gettext to DEPENDS directly
>>>> [PATCH 15/17] attr: Fix compilation on uclibc
>>>> [PATCH 16/17] linux-tools.inc: Dummify do_compile_perf and
>>>> do_install_perf for uclibc
>>>
>>>
>>> Saul
>>>
>>> I have fixed this patch and pushed it to the pull tree. Can you pull fresh
>>> from the pull request tree and give it a shot ? Let me know how it goes
>>>
>>>>> Sau!
>>>>>
>>>>>> Signed-off-by: Khem Raj<raj.khem@gmail.com>
>>>>>> ---
>>>>>>   meta/classes/gettext.bbclass |   31 +++++++++++++++++++------------
>>>>>>   1 files changed, 19 insertions(+), 12 deletions(-)
>>>>>>
>>>>>> diff --git a/meta/classes/gettext.bbclass
>>>>>> b/meta/classes/gettext.bbclass
>>>>>> index a40e74f..a6f80f2 100644
>>>>>> --- a/meta/classes/gettext.bbclass
>>>>>> +++ b/meta/classes/gettext.bbclass
>>>>>> @@ -1,17 +1,24 @@
>>>>>>   def gettext_after_parse(d):
>>>>>> -    # Remove the NLS bits if USE_NLS is no.
>>>>>> -    if bb.data.getVar('USE_NLS', d, 1) == 'no':
>>>>>> -        cfg = oe_filter_out('^--(dis|en)able-nls$',
>>>>>> bb.data.getVar('EXTRA_OECONF', d, 1) or "", d)
>>>>>> -        cfg += " --disable-nls"
>>>>>> -        depends = bb.data.getVar('DEPENDS', d, 1) or ""
>>>>>> -        bb.data.setVar('DEPENDS',
>>>>>> oe_filter_out('^(virtual/libiconv|virtual/libintl)$', depends, d), d)
>>>>>> -        bb.data.setVar('EXTRA_OECONF', cfg, d)
>>>>>> -
>>>>>> +   # Remove the NLS bits if USE_NLS is no.
>>>>>> +   if bb.data.getVar('USE_NLS', d, True) == 'no':
>>>>>> +       cfg = oe_filter_out('^--(dis|en)able-nls$',
>>>>>> bb.data.getVar('EXTRA_OECONF', d, 1) or "", d)
>>>>>> +       cfg += " --disable-nls"
>>>>>> +       depends = bb.data.getVar('DEPENDS', d, True) or ""
>>>>>> +       depends =
>>>>>>
>>>>>> oe_filter_out('^(virtual/libiconv|virtual/libintl|virtual/gettext|gettext)$',
>>>>>> depends, d)
>>>>>> +       if not oe.utils.inherits(d, 'native', 'nativesdk', 'cross',
>>>>>> 'crosssdk'):
>>>>>> +           depends = oe_filter_out('^(gettext-native)$', depends, d)
>>>>>> +       bb.data.setVar('DEPENDS', depends, d)
>>>>>> +       bb.data.setVar('EXTRA_OECONF', cfg, d)
>>>>>> +   # check if INHIBIT_DEFAULT_DEPS is 1 then we forcibly remove
>>>>>> dependencies
>>>>>> +   # added by this class through DEPENDS_GETTEXT
>>>>>> +   if bb.data.getVar('INHIBIT_DEFAULT_DEPS', d, True):
>>>>>> +       depends = bb.data.getVar('DEPENDS', d, True) or ""
>>>>>> +       gettext_deps = '^(' + bb.data.getVar('DEPENDS_GETTEXT', d,
>>>>>> True) +
>>>>>> ')$'
>>>>>> +       depends = oe_filter_out(gettext_deps, depends, d)
>>>>>> +       bb.data.setVar('DEPENDS', depends, d)
>>>>>>   python () {
>>>>>>       gettext_after_parse(d)
>>>>>>   }
>>>>>> -
>>>>>> -DEPENDS_GETTEXT = "gettext gettext-native"
>>>>>> -
>>>>>> -DEPENDS =+ "${DEPENDS_GETTEXT}"
>>>>>>   EXTRA_OECONF += "--enable-nls"
>>>>>> +DEPENDS_GETTEXT ?= "virtual/gettext"
>>>>>> +DEPENDS_append = " ${DEPENDS_GETTEXT} "
>>>>>
>>>>> _______________________________________________
>>>>> Openembedded-core mailing list
>>>>> Openembedded-core@lists.openembedded.org
>>>>> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
>>>>>
>>>>
>>>
>>> _______________________________________________
>>> Openembedded-core mailing list
>>> Openembedded-core@lists.openembedded.org
>>> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
>>
>> _______________________________________________
>> Openembedded-core mailing list
>> Openembedded-core@lists.openembedded.org
>> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
>>
>
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core