Patchwork [4/8] Revert "Add ptest for glib."

login
register
mail settings
Submitter Bernhard Reutner-Fischer
Date Feb. 5, 2013, 3:55 p.m.
Message ID <3666e39a359b03ba1e054813f4353ddeed11d506.1360079373.git.rep.dot.nop@gmail.com>
Download mbox | patch
Permalink /patch/44113/
State New
Headers show

Comments

Bernhard Reutner-Fischer - Feb. 5, 2013, 3:55 p.m.
This reverts commit dec6f912d68f52748f645b6af411f189cf464f85.

This breaks core-image-sato-sdk for TCLIBC="uclibc".

Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
---
 .../glib-2.0/glib-2.0/Makefile-ptest.patch         |   61 --------------------
 meta/recipes-core/glib-2.0/glib-2.0/run-ptest      |    3 -
 meta/recipes-core/glib-2.0/glib-2.0_2.34.3.bb      |   22 -------
 3 files changed, 86 deletions(-)
 delete mode 100644 meta/recipes-core/glib-2.0/glib-2.0/Makefile-ptest.patch
 delete mode 100644 meta/recipes-core/glib-2.0/glib-2.0/run-ptest
Richard Purdie - Feb. 5, 2013, 4:44 p.m.
On Tue, 2013-02-05 at 16:55 +0100, Bernhard Reutner-Fischer wrote:
> This reverts commit dec6f912d68f52748f645b6af411f189cf464f85.
> 
> This breaks core-image-sato-sdk for TCLIBC="uclibc".
> 
> Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
> ---
>  .../glib-2.0/glib-2.0/Makefile-ptest.patch         |   61 --------------------
>  meta/recipes-core/glib-2.0/glib-2.0/run-ptest      |    3 -
>  meta/recipes-core/glib-2.0/glib-2.0_2.34.3.bb      |   22 -------
>  3 files changed, 86 deletions(-)
>  delete mode 100644 meta/recipes-core/glib-2.0/glib-2.0/Makefile-ptest.patch
>  delete mode 100644 meta/recipes-core/glib-2.0/glib-2.0/run-ptest

How about we fix ptest for glib? The revert isn't really the direction
we want to take the project. At least discussing this with Björn (cc'd)
and/or filing a bug would be a good idea before we move to reverting it.

Cheers,

Richard
Ross Burton - Feb. 5, 2013, 4:44 p.m.
On 5 February 2013 15:55, Bernhard Reutner-Fischer
<rep.dot.nop@gmail.com> wrote:
> This reverts commit dec6f912d68f52748f645b6af411f189cf464f85.
>
> This breaks core-image-sato-sdk for TCLIBC="uclibc".

Ouch.  How about we fix it instead of reverting this?

Ross
Saul Wold - Feb. 5, 2013, 4:46 p.m.
On 02/05/2013 07:55 AM, Bernhard Reutner-Fischer wrote:
> This reverts commit dec6f912d68f52748f645b6af411f189cf464f85.
>
> This breaks core-image-sato-sdk for TCLIBC="uclibc".
>
What's the issue?  Is it fixable instead of reverting the ptest patch? I 
think this would be the wrong approach to this kind of issue as we want 
to expand the usage of ptest.

Thanks

    Sau!

> Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
> ---
>   .../glib-2.0/glib-2.0/Makefile-ptest.patch         |   61 --------------------
>   meta/recipes-core/glib-2.0/glib-2.0/run-ptest      |    3 -
>   meta/recipes-core/glib-2.0/glib-2.0_2.34.3.bb      |   22 -------
>   3 files changed, 86 deletions(-)
>   delete mode 100644 meta/recipes-core/glib-2.0/glib-2.0/Makefile-ptest.patch
>   delete mode 100644 meta/recipes-core/glib-2.0/glib-2.0/run-ptest
>
> diff --git a/meta/recipes-core/glib-2.0/glib-2.0/Makefile-ptest.patch b/meta/recipes-core/glib-2.0/glib-2.0/Makefile-ptest.patch
> deleted file mode 100644
> index cb36c09..0000000
> --- a/meta/recipes-core/glib-2.0/glib-2.0/Makefile-ptest.patch
> +++ /dev/null
> @@ -1,61 +0,0 @@
> -Add 'ptest' target to Makefile, to run tests without checking dependencies.
> -Add 'install-ptest' target, to copy needed files to ptest dir.
> -
> -Signed-off-by: Björn Stenberg <bjst@enea.com>
> -Upstream-status: Pending
> ----
> - Makefile.decl |   43 +++++++++++++++++++++++++++++++++++++++++++
> - 1 file changed, 43 insertions(+)
> -
> ---- a/Makefile.decl	2012-03-12 01:42:39.000000000 +0100
> -+++ b/Makefile.decl	2012-11-26 13:42:08.436865192 +0100
> -@@ -25,6 +25,49 @@
> - test-nonrecursive:
> - endif
> -
> -+ptest: ptest-nonrecursive
> -+	@ for subdir in $(SUBDIRS) .; do \
> -+	    test -d "$$subdir" -a "$$subdir" != "." -a "$$subdir" != "po" && \
> -+	    ( test -d $$subdir && cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) $@ ) \
> -+	  done
> -+
> -+ptest-nonrecursive:
> -+	@test -z "${TEST_PROGS}" || ${GTESTER} --keep-going --verbose ${TEST_PROGS}
> -+
> -+install-ptest: install-ptest-nonrecursive
> -+	@test -z "$(SUBDIRS)" || for subdir in $(SUBDIRS); do \
> -+	    test "$$subdir" == "." || \
> -+	    (cd $$subdir && $(MAKE) DESTDIR=$(DESTDIR)/$$subdir $@ ) \
> -+	done
> -+
> -+# install files used for ptest:
> -+# - copy EXTRA_DIST (for test data in glib/tests)
> -+# - run dist-hook (for glib/tests/bookmarks/ and glib/tests/markups)
> -+# - copy TEST_PROGS (the test programs)
> -+# - copy INSTALL_PROGS (for glib/gtester)
> -+# - avoid libtool wrappers by copying binaries from .libs
> -+# - strip Makefile: target from all Makefiles
> -+install-ptest-nonrecursive:
> -+	@mkdir -p $(DESTDIR)
> -+	@if [ -n "${TEST_PROGS}" ]; then \
> -+		if [ -n "${EXTRA_DIST}" ]; then \
> -+			cp ${EXTRA_DIST} $(DESTDIR); \
> -+		fi; \
> -+		grep -q dist-hook: Makefile && \
> -+			$(MAKE) distdir=$(DESTDIR) dist-hook; \
> -+		if [ -d .libs ] ; then \
> -+			cd .libs; \
> -+		fi; \
> -+		cp ${TEST_PROGS} $(DESTDIR); \
> -+	fi;
> -+	@if [ -n "${INSTALL_PROGS}" ]; then \
> -+		if [ -d .libs ]  ; then \
> -+			cd .libs; \
> -+		fi; \
> -+		cp ${INSTALL_PROGS} $(DESTDIR); \
> -+	fi
> -+	@test -f Makefile && sed 's/^Makefile:/_Makefile:/' <Makefile >$(DESTDIR)/Makefile
> -+
> - # test-report: run tests in subdirs and generate report
> - # perf-report: run tests in subdirs with -m perf and generate report
> - # full-report: like test-report: with -m perf and -m slow
> diff --git a/meta/recipes-core/glib-2.0/glib-2.0/run-ptest b/meta/recipes-core/glib-2.0/glib-2.0/run-ptest
> deleted file mode 100644
> index 3deb586..0000000
> --- a/meta/recipes-core/glib-2.0/glib-2.0/run-ptest
> +++ /dev/null
> @@ -1,3 +0,0 @@
> -#!/bin/sh
> -
> -make -k ptest
> diff --git a/meta/recipes-core/glib-2.0/glib-2.0_2.34.3.bb b/meta/recipes-core/glib-2.0/glib-2.0_2.34.3.bb
> index e1d44e2..e4293c8 100644
> --- a/meta/recipes-core/glib-2.0/glib-2.0_2.34.3.bb
> +++ b/meta/recipes-core/glib-2.0/glib-2.0_2.34.3.bb
> @@ -13,8 +13,6 @@ SRC_URI = "${GNOME_MIRROR}/glib/${SHRT_VER}/glib-${PV}.tar.xz \
>              file://configure-libtool.patch \
>              file://glib-2.0_fix_for_x32.patch \
>              file://obsolete_automake_macros.patch \
> -           file://Makefile-ptest.patch \
> -           file://run-ptest \
>             "
>   SRC_URI[md5sum] = "a4ca31e258273c3761e3de2edd607661"
>   SRC_URI[sha256sum] = "855fcbf87cb93065b488358e351774d8a39177281023bae58c286f41612658a7"
> @@ -22,24 +20,6 @@ SRC_URI[sha256sum] = "855fcbf87cb93065b488358e351774d8a39177281023bae58c286f4161
>   SRC_URI_append_class-native = " file://glib-gettextize-dir.patch"
>   BBCLASSEXTEND = "native nativesdk"
>
> -RDEPENDS_${PN}-ptest += "\
> -            eglibc-gconv-utf-16 \
> -            eglibc-charmap-utf-8 \
> -            eglibc-gconv-cp1255 \
> -            eglibc-charmap-cp1255 \
> -            eglibc-gconv-utf-32 \
> -            eglibc-gconv-utf-7 \
> -            eglibc-charmap-invariant \
> -            eglibc-localedata-translit-cjk-variants \
> -            tzdata \
> -            tzdata-americas \
> -            tzdata-asia \
> -            tzdata-europe \
> -            tzdata-posix \
> -            python-pygobject \
> -            python-dbus \
> -           "
> -
>   do_configure_prepend() {
>   	sed -i -e '1s,#!.*,#!${USRBINPATH}/env python,' ${S}/gio/gdbus-2.0/codegen/gdbus-codegen.in
>   }
> @@ -56,6 +36,4 @@ do_install_append() {
>     if [ -f ${D}${bindir}/glib-mkenums ]; then
>       sed -i -e '1s,#!.*perl,#! ${USRBINPATH}/env perl,' ${D}${bindir}/glib-mkenums
>     fi
> -
> -  ptest_do_install
>   }
>
Bernhard Reutner-Fischer - Feb. 5, 2013, 4:47 p.m.
On 5 February 2013 17:44, Burton, Ross <ross.burton@intel.com> wrote:
> On 5 February 2013 15:55, Bernhard Reutner-Fischer
> <rep.dot.nop@gmail.com> wrote:
>> This reverts commit dec6f912d68f52748f645b6af411f189cf464f85.
>>
>> This breaks core-image-sato-sdk for TCLIBC="uclibc".
>
> Ouch.  How about we fix it instead of reverting this?

I do not have time to look into this, sorry. But yes, of course fixing
it for real would be better..

thanks,
Ross Burton - Feb. 5, 2013, 4:59 p.m.
On 5 February 2013 16:47, Bernhard Reutner-Fischer
<rep.dot.nop@gmail.com> wrote:
> I do not have time to look into this, sorry. But yes, of course fixing
> it for real would be better..

Sharing the log of the failure would help immensely.

Ross
Bernhard Reutner-Fischer - Feb. 5, 2013, 5:13 p.m.
On 5 February 2013 17:46, Saul Wold <sgw@linux.intel.com> wrote:
> On 02/05/2013 07:55 AM, Bernhard Reutner-Fischer wrote:
>>
>> This reverts commit dec6f912d68f52748f645b6af411f189cf464f85.
>>
>> This breaks core-image-sato-sdk for TCLIBC="uclibc".
>>
> What's the issue?  Is it fixable instead of reverting the ptest patch? I

The issue is that it incorrectly and unconditionally pulls in eglibc stuff:

-RDEPENDS_${PN}-ptest += "\
-            eglibc-gconv-utf-16 \
-            eglibc-charmap-utf-8 \
-            eglibc-gconv-cp1255 \
-            eglibc-charmap-cp1255 \
-            eglibc-gconv-utf-32 \
-            eglibc-gconv-utf-7 \
-            eglibc-charmap-invariant \
-            eglibc-localedata-translit-cjk-variants \
-            tzdata \
-            tzdata-americas \
-            tzdata-asia \
-            tzdata-europe \
-            tzdata-posix \
-            python-pygobject \
-            python-dbus \
-           "

> think this would be the wrong approach to this kind of issue as we want to
> expand the usage of ptest.
Bernhard Reutner-Fischer - Feb. 5, 2013, 5:17 p.m.
On 5 February 2013 17:59, Burton, Ross <ross.burton@intel.com> wrote:
> On 5 February 2013 16:47, Bernhard Reutner-Fischer
> <rep.dot.nop@gmail.com> wrote:
>> I do not have time to look into this, sorry. But yes, of course fixing
>> it for real would be better..
>
> Sharing the log of the failure would help immensely.

The uncompressed log is about 50MB, i put it here (still 1.2MB):
http://busybox.net/~aldot/oe/mk.all.20130118-1442.53.log.xz
Saul Wold - Feb. 5, 2013, 6:39 p.m.
On 02/05/2013 09:13 AM, Bernhard Reutner-Fischer wrote:
> On 5 February 2013 17:46, Saul Wold <sgw@linux.intel.com> wrote:
>> On 02/05/2013 07:55 AM, Bernhard Reutner-Fischer wrote:
>>>
>>> This reverts commit dec6f912d68f52748f645b6af411f189cf464f85.
>>>
>>> This breaks core-image-sato-sdk for TCLIBC="uclibc".
>>>
>> What's the issue?  Is it fixable instead of reverting the ptest patch? I
>
> The issue is that it incorrectly and unconditionally pulls in eglibc stuff:
>
> -RDEPENDS_${PN}-ptest += "\
> -            eglibc-gconv-utf-16 \
> -            eglibc-charmap-utf-8 \
> -            eglibc-gconv-cp1255 \
> -            eglibc-charmap-cp1255 \
> -            eglibc-gconv-utf-32 \
> -            eglibc-gconv-utf-7 \
> -            eglibc-charmap-invariant \
> -            eglibc-localedata-translit-cjk-variants \
> -            tzdata \
> -            tzdata-americas \
> -            tzdata-asia \
> -            tzdata-europe \
> -            tzdata-posix \
> -            python-pygobject \
> -            python-dbus \
> -           "
>
Bjorn,

Do you have some time to look at this and see if you can make use of 
TCLIBC to use the correct library and locale package info based on which 
libc is being used?

I guess you did not originally think about the uclibc case.

Thanks
	Sau!

>> think this would be the wrong approach to this kind of issue as we want to
>> expand the usage of ptest.
>
>
Bernhard Reutner-Fischer - Feb. 5, 2013, 7:16 p.m.
On 5 February 2013 19:39, Saul Wold <sgw@linux.intel.com> wrote:
> On 02/05/2013 09:13 AM, Bernhard Reutner-Fischer wrote:
>>
>> On 5 February 2013 17:46, Saul Wold <sgw@linux.intel.com> wrote:
>>>
>>> On 02/05/2013 07:55 AM, Bernhard Reutner-Fischer wrote:
>>>>
>>>>
>>>> This reverts commit dec6f912d68f52748f645b6af411f189cf464f85.
>>>>
>>>> This breaks core-image-sato-sdk for TCLIBC="uclibc".
>>>>
>>> What's the issue?  Is it fixable instead of reverting the ptest patch? I
>>
>>
>> The issue is that it incorrectly and unconditionally pulls in eglibc
>> stuff:
>>
>> -RDEPENDS_${PN}-ptest += "\
>> -            eglibc-gconv-utf-16 \
>> -            eglibc-charmap-utf-8 \
>> -            eglibc-gconv-cp1255 \
>> -            eglibc-charmap-cp1255 \
>> -            eglibc-gconv-utf-32 \
>> -            eglibc-gconv-utf-7 \
>> -            eglibc-charmap-invariant \
>> -            eglibc-localedata-translit-cjk-variants \
>> -            tzdata \
>> -            tzdata-americas \
>> -            tzdata-asia \
>> -            tzdata-europe \
>> -            tzdata-posix \
>> -            python-pygobject \
>> -            python-dbus \
>> -           "
>>
> Bjorn,
>
> Do you have some time to look at this and see if you can make use of TCLIBC
> to use the correct library and locale package info based on which libc is
> being used?

would most likely DTRT, yep.
I'm curious what you would need from uClibc to satisfy that ptest thing?
What does it do, what is it good for? :)

RDEPENDS_${PN}-ptest += "\
            tzdata \
            tzdata-americas \
            tzdata-asia \
            tzdata-europe \
            tzdata-posix \
            python-pygobject \
            python-dbus \
           "
RDEPENDS_${PN}-ptest_append_libc-eglibc = "\
            eglibc-gconv-utf-16 \
            eglibc-charmap-utf-8 \
            eglibc-gconv-cp1255 \
            eglibc-charmap-cp1255 \
            eglibc-gconv-utf-32 \
            eglibc-gconv-utf-7 \
            eglibc-charmap-invariant \
            eglibc-localedata-translit-cjk-variants \
           "
RDEPENDS_${PN}-ptest_append_libc-eglibc = "did not look"

TIA && cheers,
Bernhard Reutner-Fischer - Feb. 5, 2013, 7:36 p.m.
On 5 February 2013 20:16, Bernhard Reutner-Fischer
<rep.dot.nop@gmail.com> wrote:
> On 5 February 2013 19:39, Saul Wold <sgw@linux.intel.com> wrote:
>> On 02/05/2013 09:13 AM, Bernhard Reutner-Fischer wrote:
>>>
>>> On 5 February 2013 17:46, Saul Wold <sgw@linux.intel.com> wrote:
>>>>
>>>> On 02/05/2013 07:55 AM, Bernhard Reutner-Fischer wrote:
>>>>>
>>>>>
>>>>> This reverts commit dec6f912d68f52748f645b6af411f189cf464f85.
>>>>>
>>>>> This breaks core-image-sato-sdk for TCLIBC="uclibc".
>>>>>
>>>> What's the issue?  Is it fixable instead of reverting the ptest patch? I
>>>
>>>
>>> The issue is that it incorrectly and unconditionally pulls in eglibc
>>> stuff:
>>>
>>> -RDEPENDS_${PN}-ptest += "\
>>> -            eglibc-gconv-utf-16 \
>>> -            eglibc-charmap-utf-8 \
>>> -            eglibc-gconv-cp1255 \
>>> -            eglibc-charmap-cp1255 \
>>> -            eglibc-gconv-utf-32 \
>>> -            eglibc-gconv-utf-7 \
>>> -            eglibc-charmap-invariant \
>>> -            eglibc-localedata-translit-cjk-variants \
>>> -            tzdata \
>>> -            tzdata-americas \
>>> -            tzdata-asia \
>>> -            tzdata-europe \
>>> -            tzdata-posix \
>>> -            python-pygobject \
>>> -            python-dbus \
>>> -           "
>>>
>> Bjorn,
>>
>> Do you have some time to look at this and see if you can make use of TCLIBC
>> to use the correct library and locale package info based on which libc is
>> being used?
>
> would most likely DTRT, yep.

Well, given that i see e.g. glibc-gconv-utf-16 nowhere in
openembedded-core i suspect they are misplaced and should live in the
layer that provides them.

> I'm curious what you would need from uClibc to satisfy that ptest thing?
> What does it do, what is it good for? :)

Is e.g. glibc-gconv-utf-16 locale support or something like that?

thanks,
>
> RDEPENDS_${PN}-ptest += "\
>             tzdata \
>             tzdata-americas \
>             tzdata-asia \
>             tzdata-europe \
>             tzdata-posix \
>             python-pygobject \
>             python-dbus \
>            "
> RDEPENDS_${PN}-ptest_append_libc-eglibc = "\
>             eglibc-gconv-utf-16 \
>             eglibc-charmap-utf-8 \
>             eglibc-gconv-cp1255 \
>             eglibc-charmap-cp1255 \
>             eglibc-gconv-utf-32 \
>             eglibc-gconv-utf-7 \
>             eglibc-charmap-invariant \
>             eglibc-localedata-translit-cjk-variants \
>            "
> RDEPENDS_${PN}-ptest_append_libc-eglibc = "did not look"
>
> TIA && cheers,
Saul Wold - Feb. 5, 2013, 7:59 p.m.
On 02/05/2013 11:36 AM, Bernhard Reutner-Fischer wrote:
> On 5 February 2013 20:16, Bernhard Reutner-Fischer
> <rep.dot.nop@gmail.com> wrote:
>> On 5 February 2013 19:39, Saul Wold <sgw@linux.intel.com> wrote:
>>> On 02/05/2013 09:13 AM, Bernhard Reutner-Fischer wrote:
>>>>
>>>> On 5 February 2013 17:46, Saul Wold <sgw@linux.intel.com> wrote:
>>>>>
>>>>> On 02/05/2013 07:55 AM, Bernhard Reutner-Fischer wrote:
>>>>>>
>>>>>>
>>>>>> This reverts commit dec6f912d68f52748f645b6af411f189cf464f85.
>>>>>>
>>>>>> This breaks core-image-sato-sdk for TCLIBC="uclibc".
>>>>>>
>>>>> What's the issue?  Is it fixable instead of reverting the ptest patch? I
>>>>
>>>>
>>>> The issue is that it incorrectly and unconditionally pulls in eglibc
>>>> stuff:
>>>>
>>>> -RDEPENDS_${PN}-ptest += "\
>>>> -            eglibc-gconv-utf-16 \
>>>> -            eglibc-charmap-utf-8 \
>>>> -            eglibc-gconv-cp1255 \
>>>> -            eglibc-charmap-cp1255 \
>>>> -            eglibc-gconv-utf-32 \
>>>> -            eglibc-gconv-utf-7 \
>>>> -            eglibc-charmap-invariant \
>>>> -            eglibc-localedata-translit-cjk-variants \
>>>> -            tzdata \
>>>> -            tzdata-americas \
>>>> -            tzdata-asia \
>>>> -            tzdata-europe \
>>>> -            tzdata-posix \
>>>> -            python-pygobject \
>>>> -            python-dbus \
>>>> -           "
>>>>
>>> Bjorn,
>>>
>>> Do you have some time to look at this and see if you can make use of TCLIBC
>>> to use the correct library and locale package info based on which libc is
>>> being used?
>>
>> would most likely DTRT, yep.
>
> Well, given that i see e.g. glibc-gconv-utf-16 nowhere in
> openembedded-core i suspect they are misplaced and should live in the
> layer that provides them.
>
It's a DYNAMIC package built in eglibc-locale.inc.

>> I'm curious what you would need from uClibc to satisfy that ptest thing?
>> What does it do, what is it good for? :)
>
Bjorn question at this point!

> Is e.g. glibc-gconv-utf-16 locale support or something like that?
>
probably, could the uclibc-gconv provide equivalent?

Sau!


> thanks,
>>
>> RDEPENDS_${PN}-ptest += "\
>>              tzdata \
>>              tzdata-americas \
>>              tzdata-asia \
>>              tzdata-europe \
>>              tzdata-posix \
>>              python-pygobject \
>>              python-dbus \
>>             "
>> RDEPENDS_${PN}-ptest_append_libc-eglibc = "\
>>              eglibc-gconv-utf-16 \
>>              eglibc-charmap-utf-8 \
>>              eglibc-gconv-cp1255 \
>>              eglibc-charmap-cp1255 \
>>              eglibc-gconv-utf-32 \
>>              eglibc-gconv-utf-7 \
>>              eglibc-charmap-invariant \
>>              eglibc-localedata-translit-cjk-variants \
>>             "
>> RDEPENDS_${PN}-ptest_append_libc-eglibc = "did not look"
>>
>> TIA && cheers,
>
>
Khem Raj - Feb. 6, 2013, 12:01 a.m.
On Tue, Feb 5, 2013 at 11:59 AM, Saul Wold <sgw@linux.intel.com> wrote:
> On 02/05/2013 11:36 AM, Bernhard Reutner-Fischer wrote:
>>
>> On 5 February 2013 20:16, Bernhard Reutner-Fischer
>> <rep.dot.nop@gmail.com> wrote:
>>>
>>> On 5 February 2013 19:39, Saul Wold <sgw@linux.intel.com> wrote:
>>>>
>>>> On 02/05/2013 09:13 AM, Bernhard Reutner-Fischer wrote:
>>>>>
>>>>>
>>>>> On 5 February 2013 17:46, Saul Wold <sgw@linux.intel.com> wrote:
>>>>>>
>>>>>>
>>>>>> On 02/05/2013 07:55 AM, Bernhard Reutner-Fischer wrote:
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> This reverts commit dec6f912d68f52748f645b6af411f189cf464f85.
>>>>>>>
>>>>>>> This breaks core-image-sato-sdk for TCLIBC="uclibc".
>>>>>>>
>>>>>> What's the issue?  Is it fixable instead of reverting the ptest patch?
>>>>>> I
>>>>>
>>>>>
>>>>>
>>>>> The issue is that it incorrectly and unconditionally pulls in eglibc
>>>>> stuff:
>>>>>
>>>>> -RDEPENDS_${PN}-ptest += "\
>>>>> -            eglibc-gconv-utf-16 \
>>>>> -            eglibc-charmap-utf-8 \
>>>>> -            eglibc-gconv-cp1255 \
>>>>> -            eglibc-charmap-cp1255 \
>>>>> -            eglibc-gconv-utf-32 \
>>>>> -            eglibc-gconv-utf-7 \
>>>>> -            eglibc-charmap-invariant \
>>>>> -            eglibc-localedata-translit-cjk-variants \
>>>>> -            tzdata \
>>>>> -            tzdata-americas \
>>>>> -            tzdata-asia \
>>>>> -            tzdata-europe \
>>>>> -            tzdata-posix \
>>>>> -            python-pygobject \
>>>>> -            python-dbus \
>>>>> -           "
>>>>>
>>>> Bjorn,
>>>>
>>>> Do you have some time to look at this and see if you can make use of
>>>> TCLIBC
>>>> to use the correct library and locale package info based on which libc
>>>> is
>>>> being used?
>>>
>>>
>>> would most likely DTRT, yep.
>>
>>
>> Well, given that i see e.g. glibc-gconv-utf-16 nowhere in
>> openembedded-core i suspect they are misplaced and should live in the
>> layer that provides them.
>>
> It's a DYNAMIC package built in eglibc-locale.inc.
>
>
>>> I'm curious what you would need from uClibc to satisfy that ptest thing?
>>> What does it do, what is it good for? :)
>>
>>
> Bjorn question at this point!
>
>
>> Is e.g. glibc-gconv-utf-16 locale support or something like that?
>>
> probably, could the uclibc-gconv provide equivalent?
>


split the ptest recipe out into its own and make it eglibc only recipe
locales and gconv stuff wont fly with uclibc


> Sau!
>
>
>
>> thanks,
>>>
>>>
>>> RDEPENDS_${PN}-ptest += "\
>>>              tzdata \
>>>              tzdata-americas \
>>>              tzdata-asia \
>>>              tzdata-europe \
>>>              tzdata-posix \
>>>              python-pygobject \
>>>              python-dbus \
>>>             "
>>> RDEPENDS_${PN}-ptest_append_libc-eglibc = "\
>>>              eglibc-gconv-utf-16 \
>>>              eglibc-charmap-utf-8 \
>>>              eglibc-gconv-cp1255 \
>>>              eglibc-charmap-cp1255 \
>>>              eglibc-gconv-utf-32 \
>>>              eglibc-gconv-utf-7 \
>>>              eglibc-charmap-invariant \
>>>              eglibc-localedata-translit-cjk-variants \
>>>             "
>>> RDEPENDS_${PN}-ptest_append_libc-eglibc = "did not look"
>>>
>>> TIA && cheers,
>>
>>
>>
>
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Björn Stenberg - Feb. 6, 2013, 9:02 a.m.
Saul Wold wrote:
> >-RDEPENDS_${PN}-ptest += "\
> >-            eglibc-gconv-utf-16 \
> >-            eglibc-charmap-utf-8 \
> >-            eglibc-gconv-cp1255 \
> >-            eglibc-charmap-cp1255 \
> >-            eglibc-gconv-utf-32 \
> >-            eglibc-gconv-utf-7 \
> >-            eglibc-charmap-invariant \
> >-            eglibc-localedata-translit-cjk-variants \
> >-            tzdata \
> >-            tzdata-americas \
> >-            tzdata-asia \
> >-            tzdata-europe \
> >-            tzdata-posix \
> >-            python-pygobject \
> >-            python-dbus \
> >-           "
> Do you have some time to look at this and see if you can make use of
> TCLIBC to use the correct library and locale package info based on
> which libc is being used?
> 
> I guess you did not originally think about the uclibc case.

Correct, I did not.

The list of packages added by ptest are things that are used by the glib test suite. Without this data, some tests will fail.

I'll look into making it more uclibc friendly.
Björn Stenberg - Feb. 8, 2013, 8:21 a.m.
Bernhard Reutner-Fischer wrote:
> The uncompressed log is about 50MB, i put it here (still 1.2MB):
> http://busybox.net/~aldot/oe/mk.all.20130118-1442.53.log.xz

Can I ask how you worked around the iconv RPATH issue?

https://bugzilla.yoctoproject.org/show_bug.cgi?id=2669

Patch

diff --git a/meta/recipes-core/glib-2.0/glib-2.0/Makefile-ptest.patch b/meta/recipes-core/glib-2.0/glib-2.0/Makefile-ptest.patch
deleted file mode 100644
index cb36c09..0000000
--- a/meta/recipes-core/glib-2.0/glib-2.0/Makefile-ptest.patch
+++ /dev/null
@@ -1,61 +0,0 @@ 
-Add 'ptest' target to Makefile, to run tests without checking dependencies.
-Add 'install-ptest' target, to copy needed files to ptest dir.
-
-Signed-off-by: Björn Stenberg <bjst@enea.com>
-Upstream-status: Pending
----
- Makefile.decl |   43 +++++++++++++++++++++++++++++++++++++++++++
- 1 file changed, 43 insertions(+)
-
---- a/Makefile.decl	2012-03-12 01:42:39.000000000 +0100
-+++ b/Makefile.decl	2012-11-26 13:42:08.436865192 +0100
-@@ -25,6 +25,49 @@
- test-nonrecursive:
- endif
- 
-+ptest: ptest-nonrecursive
-+	@ for subdir in $(SUBDIRS) .; do \
-+	    test -d "$$subdir" -a "$$subdir" != "." -a "$$subdir" != "po" && \
-+	    ( test -d $$subdir && cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) $@ ) \
-+	  done
-+
-+ptest-nonrecursive:
-+	@test -z "${TEST_PROGS}" || ${GTESTER} --keep-going --verbose ${TEST_PROGS}
-+
-+install-ptest: install-ptest-nonrecursive
-+	@test -z "$(SUBDIRS)" || for subdir in $(SUBDIRS); do \
-+	    test "$$subdir" == "." || \
-+	    (cd $$subdir && $(MAKE) DESTDIR=$(DESTDIR)/$$subdir $@ ) \
-+	done
-+
-+# install files used for ptest:
-+# - copy EXTRA_DIST (for test data in glib/tests)
-+# - run dist-hook (for glib/tests/bookmarks/ and glib/tests/markups)
-+# - copy TEST_PROGS (the test programs)
-+# - copy INSTALL_PROGS (for glib/gtester)
-+# - avoid libtool wrappers by copying binaries from .libs
-+# - strip Makefile: target from all Makefiles
-+install-ptest-nonrecursive:
-+	@mkdir -p $(DESTDIR)
-+	@if [ -n "${TEST_PROGS}" ]; then \
-+		if [ -n "${EXTRA_DIST}" ]; then \
-+			cp ${EXTRA_DIST} $(DESTDIR); \
-+		fi; \
-+		grep -q dist-hook: Makefile && \
-+			$(MAKE) distdir=$(DESTDIR) dist-hook; \
-+		if [ -d .libs ] ; then \
-+			cd .libs; \
-+		fi; \
-+		cp ${TEST_PROGS} $(DESTDIR); \
-+	fi;
-+	@if [ -n "${INSTALL_PROGS}" ]; then \
-+		if [ -d .libs ]  ; then \
-+			cd .libs; \
-+		fi; \
-+		cp ${INSTALL_PROGS} $(DESTDIR); \
-+	fi
-+	@test -f Makefile && sed 's/^Makefile:/_Makefile:/' <Makefile >$(DESTDIR)/Makefile
-+
- # test-report: run tests in subdirs and generate report
- # perf-report: run tests in subdirs with -m perf and generate report
- # full-report: like test-report: with -m perf and -m slow
diff --git a/meta/recipes-core/glib-2.0/glib-2.0/run-ptest b/meta/recipes-core/glib-2.0/glib-2.0/run-ptest
deleted file mode 100644
index 3deb586..0000000
--- a/meta/recipes-core/glib-2.0/glib-2.0/run-ptest
+++ /dev/null
@@ -1,3 +0,0 @@ 
-#!/bin/sh
-
-make -k ptest
diff --git a/meta/recipes-core/glib-2.0/glib-2.0_2.34.3.bb b/meta/recipes-core/glib-2.0/glib-2.0_2.34.3.bb
index e1d44e2..e4293c8 100644
--- a/meta/recipes-core/glib-2.0/glib-2.0_2.34.3.bb
+++ b/meta/recipes-core/glib-2.0/glib-2.0_2.34.3.bb
@@ -13,8 +13,6 @@  SRC_URI = "${GNOME_MIRROR}/glib/${SHRT_VER}/glib-${PV}.tar.xz \
            file://configure-libtool.patch \
            file://glib-2.0_fix_for_x32.patch \
            file://obsolete_automake_macros.patch \
-           file://Makefile-ptest.patch \
-           file://run-ptest \
           "
 SRC_URI[md5sum] = "a4ca31e258273c3761e3de2edd607661"
 SRC_URI[sha256sum] = "855fcbf87cb93065b488358e351774d8a39177281023bae58c286f41612658a7"
@@ -22,24 +20,6 @@  SRC_URI[sha256sum] = "855fcbf87cb93065b488358e351774d8a39177281023bae58c286f4161
 SRC_URI_append_class-native = " file://glib-gettextize-dir.patch"
 BBCLASSEXTEND = "native nativesdk"
 
-RDEPENDS_${PN}-ptest += "\
-            eglibc-gconv-utf-16 \
-            eglibc-charmap-utf-8 \
-            eglibc-gconv-cp1255 \
-            eglibc-charmap-cp1255 \
-            eglibc-gconv-utf-32 \
-            eglibc-gconv-utf-7 \
-            eglibc-charmap-invariant \
-            eglibc-localedata-translit-cjk-variants \
-            tzdata \
-            tzdata-americas \
-            tzdata-asia \
-            tzdata-europe \
-            tzdata-posix \
-            python-pygobject \
-            python-dbus \
-           "
-
 do_configure_prepend() {
 	sed -i -e '1s,#!.*,#!${USRBINPATH}/env python,' ${S}/gio/gdbus-2.0/codegen/gdbus-codegen.in
 }
@@ -56,6 +36,4 @@  do_install_append() {
   if [ -f ${D}${bindir}/glib-mkenums ]; then
     sed -i -e '1s,#!.*perl,#! ${USRBINPATH}/env perl,' ${D}${bindir}/glib-mkenums
   fi
-
-  ptest_do_install
 }