[oe] [PATCH 1/2] mypaint: new recipe for version 0.9.1

Christian Charreyre christian.charreyre at cioinfoindus.fr
Thu Apr 28 08:41:32 UTC 2011


Dear Paul,

sorry for all these mistakes, it is my first patch submission and I'm 
not quite operational ....

A few questions for some of your comments :

Le 27/04/2011 22:40, Paul Menzel a écrit :
> Dear Christian,
>
>
> Am Mittwoch, den 27.04.2011, 16:55 +0200 schrieb Christian Charreyre:
>
>> this new recipe will build the painting soft mypaint.
> thank you for your patches.
>
>> In order to correctly run on the target, it is necessary to add new
>> recipes dependencies in python-numpy recipe.
>>
>> Here are the two patches.
>>
>> Hoping it will be valuable.
> Please send each patch separately using `git send-email` or using your
> mail program and removing the first From line and past the subject into
> the according field of your mailer.
>
When you says "send each patch separetely", do you lean that I must 
split "recipes/mypaint/files/scons-adapt.patch" and 
"recipes/mypaint/mypaint_0.9.1.bb" in 2 separate patches, or that I must 
split those related to mypaint and those related to python-numpy ? The 
second alternative has been done.

I was thinking that when 2 files changes are correlated, they must be in 
the same patch, no ?

>>    From 0cc6d007fecb3ad7f5d09a9d4897d506a5c41d39 Mon Sep 17 00:00:00 2001
>> From: C. Charreyre<christian.charreyre at cioinfoindus.fr>
> A space is missing. You can set it up correctly using `git config`.
I suppose the missing space is before the 2nd From with my email 
address. Is is ok ?
>> Date: Thu, 21 Apr 2011 16:40:54 +0200
>> Subject: [PATCH] mypaint: new recipe for version 0.9.1
>>
>> Signed-off-by: C. Charreyre<christian.charreyre at cioinfoindus.fr>
>> ---
>>     recipes/mypaint/files/scons-adapt.patch |   20 +++++++++++++++++
>>     recipes/mypaint/mypaint_0.9.1.bb        |   36
>> +++++++++++++++++++++++++++++++
> Your mail program mangled the patch. Please resend as `[PATCH v2]` and
> remember to turn off line wrapping.
OK, noted.
>>     2 files changed, 56 insertions(+), 0 deletions(-)
>>     create mode 100644 recipes/mypaint/files/scons-adapt.patch
>>     create mode 100644 recipes/mypaint/mypaint_0.9.1.bb
>>
>> diff --git a/recipes/mypaint/files/scons-adapt.patch
>> b/recipes/mypaint/files/scons-adapt.patch
>> new file mode 100644
>> index 0000000..dcad6db
>> --- /dev/null
>> +++ b/recipes/mypaint/files/scons-adapt.patch
>> @@ -0,0 +1,20 @@
>> +Author: Christian Charreyre<christian.charreyre at cioinfoindus.fr>
>> +
>> +Adapt scons to OE
>> +Index: mypaint-0.9.1/lib/SConscript
>> +===================================================================
>> +--- mypaint-0.9.1.orig/lib/SConscript    2011-04-21 11:31:29.000000000
>> +0200
>> ++++ mypaint-0.9.1/lib/SConscript    2011-04-21 11:35:30.000000000 +0200
>> +@@ -1,5 +1,11 @@
>> + Import('env')
>> +-import sys
>> ++import sys, os
>> ++env.Replace(SHCXX = os.environ['CXX'])
>> ++env.Replace(SHLINK = os.environ['CXX'])
>> ++env.Replace(LDMODULEFLAGS = os.environ['TARGET_LDFLAGS'])
>> ++env.Append(LDMODULEFLAGS = ['-Wl,--hash-style=gnu'])
>> ++env.Append(LDMODULEFLAGS = ['-shared'])
>> ++env.Replace(PATH = os.environ['PATH'])
> Did you send this upstream so that they can improve the build system for
> cross compilation?
I think I can't send it upstream, because this is specific to OE 
(TARGET_LDFLAGS for instance is defined by OE).
In fact I didn't knew scons at all, and there are very few recipes with 
this compilation system, so I have tried to have mypaint successfully 
cross-compiled in OE with this patch.
If someone can improve that, it would be fine.
>> +
>> + # For the record: I know that scons supports swig. But it doesn't scan
>> for #include in the generated code.
>> + #
>> diff --git a/recipes/mypaint/mypaint_0.9.1.bb
>> b/recipes/mypaint/mypaint_0.9.1.bb
>> new file mode 100644
>> index 0000000..a7e5077
>> --- /dev/null
>> +++ b/recipes/mypaint/mypaint_0.9.1.bb
>> @@ -0,0 +1,36 @@
>> +DESCRIPTION="fast and easy open-source graphics application for digital
>> painters"
>> +LICENSE="GPL"
>> +SECTION = "x11/applications"
>> +PRIORITY = "optional"
>> +HOMEPAGE = "http://mypaint.intilinux.com/"
>> +
>> +RDEPENDS_${PN} = "python-numpy python-image python-json"
> Can you sort those lexicographically.
OK, noted.
>> +
>> +SRC_URI = "http://download.gna.org/mypaint/${PN}-${PV}.tar.bz2 \
>> +    file://scons-adapt.patch \
>> +"
>> +
>> +inherit distutils scons
>> +
>> +do_compile() {
>> +    STAGING_INCDIR=${STAGING_INCDIR} \
>> +    STAGING_LIBDIR=${STAGING_LIBDIR} \
>> +    BUILD_SYS=${BUILD_SYS} \
>> +    HOST_SYS=${HOST_SYS} \
>> +    PATH=${PATH} scons_do_compile
>> +}
> I do not know Scons, but it looks strange that the above is needed and
> not done automatically.
Same reply as above, there are few recipes with scons, I've found this 
tip in one recipe and applied it here to get the build completed.
>> +
>> +do_install() {
>> +    BUILD_SYS=${BUILD_SYS} \
>> +    HOST_SYS=${HOST_SYS} \
>> +    scons_do_install
>> +}
>> +
>> +FILES_${PN} += "${datadir}"
>> +
>> +SRC_URI[md5sum] = "6249a16359a438d6dc658f5765b35515"
>> +SRC_URI[sha256sum] =
>> "407b599f62fb0d6e711fee57d22e64d3aec88825364fb5f7f73b9f0940aa7aed"
> Please put it right at the top under `SRC_URI`.
I don't understand what is the problem. What part has to be placed under 
'SRC_URI' ?
>> +
>> +INSANE_SKIP_${PN} = True
> Please note in the commit message or in the recipe as a comment what
> error you get/why this is needed.
OK, noted.
>> +
>> +
> Please remove the empty lines at the end.
OK, noted.
>
> Thanks,
>
> Paul
Thanks for your comments that will improve my future submissions.
>
>
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel at lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

-------------- next part --------------
A non-text attachment was scrubbed...
Name: christian_charreyre.vcf
Type: text/x-vcard
Size: 277 bytes
Desc: not available
URL: <http://lists.openembedded.org/pipermail/openembedded-devel/attachments/20110428/0cc6b5d8/attachment-0002.vcf>


More information about the Openembedded-devel mailing list