[oe] FW: [PATCH] bluez4_4.95.bb: new recipe for building bluez-4.95

Sun Nov 13 07:55:51 UTC 2011

Ping to: [oe] [PATCH] bluez4_4.95.bb: new recipe for building bluez-4.95

Regards,
Vita Preskovsky

-----Original Message-----
From: openembedded-devel-bounces at lists.openembedded.org [mailto:openembedded-devel-bounces at lists.openembedded.org] On Behalf Of Preskovsky, Vita
Sent: Monday, October 31, 2011 11:19 PM
To: Paul Menzel
Cc: openembedded-devel at lists.openembedded.org
Subject: Re: [oe] [PATCH] bluez4_4.95.bb: new recipe for building bluez-4.95

Paul,

I'm subscribed to the list and receive mails from it, but I'm new to arago and openembeded.
I have a question about the right way to define DEFAULT_PREFERENCE in recipes.
In all bluez4_4.x recipes there is definition of default preference as
DEFAULT_PREFERENCE = "-1", and bitbake takes the last one.
In the last recipe version 4.89 they defined also angstrom default preference:
DEFAULT_PREFERENCE_angstrom = "1"
So everyone, working under angstrom, as we do, compiles with the version 4.89.
We chose to work with version 4.95 (despite existence of 4.96) but we don't want this recipe will be compiled as default, only for those who will compile this recipe for specific machine as am180x-evm.
So instead of DEFAULT_PREFERENCE_angstrom I can define DEFAULT_PREFERENCE_am180x-evm = "1", that meets our goal.
Is this way more appropriate than I did previously? 

Thank you,
Vita.

-----Original Message-----
From: Paul Menzel [mailto:paulepanter at users.sourceforge.net] 
Sent: Sunday, October 30, 2011 1:42 PM
To: openembedded-devel at lists.openembedded.org
Cc: Preskovsky, Vita; Vita Preskovsky
Subject: Re: [oe] [PATCH] bluez4_4.95.bb: new recipe for building bluez-4.95

Dear Vita,

are you reading the list or do I have to add your address to the CC field?

Am Sonntag, den 30.10.2011, 10:58 +0100 schrieb Preskovsky, Vita:

> Please see my answers bellow.

thank you for your answer.

> -----Original Message-----

> Sent: Tuesday, October 11, 2011 4:32 PM

> Subject: Re: [oe] [PATCH] bluez4_4.95.bb: new recipe for building 
> bluez-4.95
> 
> Dear Vita,
> 
> 
> thank you for your patch.
> 
> Am Dienstag, den 11.10.2011, 11:08 +0200 schrieb Vita Preskovsky:
> 
> Using
> 
>         bluez4: Add version 4.95
> 
> as the commit summary would be cleaner. Use `git commit --amend` to change that.
> 
> Additionally you should elaborate more in the commit message what the patch is doing.
> 
> 1. Have you tested it? Is it just an upgrade because of or does 
> version
> 4.91 in the repository have any short comings so that it should be removed?
>
> I tested this recipe, and the reason I added this version is that we want to use this version of Bluez in our release.

I guess you tested it using Ȧngström, so the old recipe can be removed.
Please note that oe-core already has version 4.96.

> > Signed-off-by: Vita Preskovsky <vitap at ti.com>
> > ---
> >  recipes/bluez/bluez4_4.95.bb |   30 ++++++++++++++++++++++++++++++
> >  1 files changed, 30 insertions(+), 0 deletions(-)  create mode 
> > 100644 recipes/bluez/bluez4_4.95.bb
> > 
> > diff --git a/recipes/bluez/bluez4_4.95.bb 
> > b/recipes/bluez/bluez4_4.95.bb new file mode 100644 index 
> > 0000000..a682d6a
> > --- /dev/null
> > +++ b/recipes/bluez/bluez4_4.95.bb
> > @@ -0,0 +1,30 @@
> > +require bluez4.inc
> > +SRC_URI = "\
> > +  http://www.kernel.org/pub/linux/bluetooth/bluez-${PV}.tar.gz \
> 
> 2. As far as I know kernel.org after the compromise presently does not provide that archive. So the recipe will not build unless there is a mirror somewhere. Do you know of a mirror?
> 
> There are two possible locations for this package:
> 	http://www.mirrorservice.org/sites/ftp.kernel.org/pub/linux/bluetooth/bluez-4.95.tar.gz
>  	
> http://www.angstrom-distribution.org/unstable/sources/bluez-4.95.tar.g
> z I changed the SRC_URI in my working directory for the first location 
> and it worked fine. So with what location do you prefer the recipe will be released?

I am unsure about the policy myself and I do not think an “OE mirror”
has that particular archive.

> > + file://bluetooth.conf \
> > +"
> > +
> > +SRC_URI[md5sum] = "341294b2849a04a4afff5c96bfbf30b2"
> > +SRC_URI[sha256sum] = "d6ea9de410fc2bcd2620d709c2202893b218e2e6a55c3c0ce6bebd27fa4120f6"
> > +
> > +DEFAULT_PREFERENCE = "-1"
> 
> 3. Why?
> > +DEFAULT_PREFERENCE_angstrom = "1"
> 
> 4. Have you talked to the Ȧngström maintainers?
> 
> These two variables are defined in the previous version of bluez4 recipe: bluez4_4.89.bb. Without these two definitions in my recipe bitbake prefers the old recipe upon mine, which doesn't leave me a choice but to define these variables.

The “correct” way would be to first add that recipe and then afterward sent a patch to change the preferences which the Ȧngström maintainers have to acknowledge.

> > +
> > +DEPENDS += "libsndfile1"
> > +
> > +PR = "${INC_PR}.0"
> > +
> > +# Not all distros have a recent enough udev BTUDEV = " 
> > +--disable-udevrules"
> > +BTUDEV_angstrom = " --enable-udevrules"
> > +BTUDEV_shr = " --enable-udevrules"
> > +
> > +EXTRA_OECONF += "${BTUDEV}"
> > +do_configure_append(){
> > +	echo "#define LIBUDEV_I_KNOW_THE_API_IS_SUBJECT_TO_CHANGE" >> 
> > +${S}/config.h }
> > +
> > +FILES_${PN}-dbg += "\
> > +  ${base_libdir}/udev/.debug \
> > +  ${libdir}/*/.debug \
> > +"
> 
> Otherwise this recipe looks good. Is it planned for branch 2011.03-maintenance?

Thanks,

Paul
_______________________________________________
Openembedded-devel mailing list
Openembedded-devel at lists.openembedded.org
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel