[PATCH 0/1] ncurses packaging fix

[PATCH 0/1] ncurses packaging fix

[Scott Garman]

Hello,

This fixes the ncurses library packaging. Thanks to Andreas Oberritter
for pointing it out.

Scott

The following changes since commit 81379dc141e245307ca391a02b7490b199128991:

  dates: Fix unpackaged files warning (2012-03-13 23:18:19 +0000)

are available in the git repository at:
  git://git.pokylinux.org/poky-contrib sgarman/ncurses-pkg-fix-oe
  http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=sgarman/ncurses-pkg-fix-oe

Scott Garman (1):
  ncurses: fix packaging of moved libraries

 meta/recipes-core/ncurses/ncurses.inc |   18 ++++++++++++++++--
 1 files changed, 16 insertions(+), 2 deletions(-)

-- 
1.7.5.4

[PATCH 1/1] ncurses: fix packaging of moved libraries

[Scott Garman]

The recent move of some of the ncurses libraries to base_libdir
does not work with our automatic library packaging, so this
explicity creates packages for libncurses, libncursesw, and
libtinfo.

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
---
 meta/recipes-core/ncurses/ncurses.inc |   18 ++++++++++++++++--
 1 files changed, 16 insertions(+), 2 deletions(-)

diff --git a/meta/recipes-core/ncurses/ncurses.inc b/meta/recipes-core/ncurses/ncurses.inc
index e450868..155c0bd 100644
--- a/meta/recipes-core/ncurses/ncurses.inc
+++ b/meta/recipes-core/ncurses/ncurses.inc
@@ -6,7 +6,7 @@ LIC_FILES_CHKSUM = "file://ncurses/base/version.c;beginline=1;endline=27;md5=cbc
 SECTION = "libs"
 DEPENDS = "ncurses-native"
 DEPENDS_virtclass-native = ""
-INC_PR = "r5"
+INC_PR = "r6"
 
 inherit autotools binconfig multilib_header
 
@@ -220,6 +220,9 @@ pkg_prerm_ncurses-tools () {
 BBCLASSEXTEND = "native nativesdk"
 
 PACKAGES += " \
+  ${PN}-libncurses \
+  ${PN}-libncursesw \
+  ${PN}-libtinfo \
   ${PN}-tools \
   ${PN}-terminfo \
   ${PN}-terminfo-base \
@@ -231,7 +234,18 @@ FILES_${PN} = "\
   ${bindir}/ncurses5-config \
   ${bindir}/ncursesw5-config \
   ${datadir}/tabset \
-  ${base_libdir}/* \
+"
+
+FILES_${PN}-libncurses = "\
+  ${base_libdir}/libncurses.so.* \
+"
+
+FILES_${PN}-libncursesw = "\
+  ${base_libdir}/libncursesw.so.* \
+"
+
+FILES_${PN}-libtinfo = "\
+  ${base_libdir}/libtinfo.so.* \
 "
 
 # This keeps only tput/tset in ncurses
-- 
1.7.5.4

Re: [PATCH 1/1] ncurses: fix packaging of moved libraries

[Darren Hart]

On 03/13/2012 05:31 PM, Scott Garman wrote:
> The recent move of some of the ncurses libraries to base_libdir
> does not work with our automatic library packaging, so this
> explicity creates packages for libncurses, libncursesw, and
> libtinfo.
> 
> Signed-off-by: Scott Garman <scott.a.garman@intel.com>
> ---
>  meta/recipes-core/ncurses/ncurses.inc |   18 ++++++++++++++++--
>  1 files changed, 16 insertions(+), 2 deletions(-)
> 
> diff --git a/meta/recipes-core/ncurses/ncurses.inc b/meta/recipes-core/ncurses/ncurses.inc
> index e450868..155c0bd 100644
> --- a/meta/recipes-core/ncurses/ncurses.inc
> +++ b/meta/recipes-core/ncurses/ncurses.inc
> @@ -6,7 +6,7 @@ LIC_FILES_CHKSUM = "file://ncurses/base/version.c;beginline=1;endline=27;md5=cbc
>  SECTION = "libs"
>  DEPENDS = "ncurses-native"
>  DEPENDS_virtclass-native = ""
> -INC_PR = "r5"
> +INC_PR = "r6"
>  
>  inherit autotools binconfig multilib_header
>  
> @@ -220,6 +220,9 @@ pkg_prerm_ncurses-tools () {
>  BBCLASSEXTEND = "native nativesdk"
>  
>  PACKAGES += " \
> +  ${PN}-libncurses \
> +  ${PN}-libncursesw \

This one needs to be dependent on ENABLE_WIDEC

> +  ${PN}-libtinfo \
>    ${PN}-tools \
>    ${PN}-terminfo \
>    ${PN}-terminfo-base \
> @@ -231,7 +234,18 @@ FILES_${PN} = "\
>    ${bindir}/ncurses5-config \
>    ${bindir}/ncursesw5-config \
>    ${datadir}/tabset \
> -  ${base_libdir}/* \
> +"
> +
> +FILES_${PN}-libncurses = "\
> +  ${base_libdir}/libncurses.so.* \
> +"
> +
> +FILES_${PN}-libncursesw = "\
> +  ${base_libdir}/libncursesw.so.* \
> +"

This file doesn't exist if !ENABLE_WIDEC (which is the case with
poky-tiny for example).

> +
> +FILES_${PN}-libtinfo = "\
> +  ${base_libdir}/libtinfo.so.* \
>  "
>  
>  # This keeps only tput/tset in ncurses

-- 
Darren Hart
Intel Open Source Technology Center
Yocto Project - Linux Kernel

Re: [PATCH 1/1] ncurses: fix packaging of moved libraries

[Martin Jansa]

[-- Attachment #1: Type: text/plain, Size: 2434 bytes --]

On Wed, Mar 14, 2012 at 07:57:56AM -0700, Darren Hart wrote:
> 
> 
> On 03/13/2012 05:31 PM, Scott Garman wrote:
> > The recent move of some of the ncurses libraries to base_libdir
> > does not work with our automatic library packaging, so this
> > explicity creates packages for libncurses, libncursesw, and
> > libtinfo.
> > 
> > Signed-off-by: Scott Garman <scott.a.garman@intel.com>
> > ---
> >  meta/recipes-core/ncurses/ncurses.inc |   18 ++++++++++++++++--
> >  1 files changed, 16 insertions(+), 2 deletions(-)
> > 
> > diff --git a/meta/recipes-core/ncurses/ncurses.inc b/meta/recipes-core/ncurses/ncurses.inc
> > index e450868..155c0bd 100644
> > --- a/meta/recipes-core/ncurses/ncurses.inc
> > +++ b/meta/recipes-core/ncurses/ncurses.inc
> > @@ -6,7 +6,7 @@ LIC_FILES_CHKSUM = "file://ncurses/base/version.c;beginline=1;endline=27;md5=cbc
> >  SECTION = "libs"
> >  DEPENDS = "ncurses-native"
> >  DEPENDS_virtclass-native = ""
> > -INC_PR = "r5"
> > +INC_PR = "r6"
> >  
> >  inherit autotools binconfig multilib_header
> >  
> > @@ -220,6 +220,9 @@ pkg_prerm_ncurses-tools () {
> >  BBCLASSEXTEND = "native nativesdk"
> >  
> >  PACKAGES += " \
> > +  ${PN}-libncurses \
> > +  ${PN}-libncursesw \
> 
> This one needs to be dependent on ENABLE_WIDEC

isn't just
ALLOW_EMPTY_${PN}-libncursesw = "1"
enough?

> 
> > +  ${PN}-libtinfo \
> >    ${PN}-tools \
> >    ${PN}-terminfo \
> >    ${PN}-terminfo-base \
> > @@ -231,7 +234,18 @@ FILES_${PN} = "\
> >    ${bindir}/ncurses5-config \
> >    ${bindir}/ncursesw5-config \
> >    ${datadir}/tabset \
> > -  ${base_libdir}/* \
> > +"
> > +
> > +FILES_${PN}-libncurses = "\
> > +  ${base_libdir}/libncurses.so.* \
> > +"
> > +
> > +FILES_${PN}-libncursesw = "\
> > +  ${base_libdir}/libncursesw.so.* \
> > +"
> 
> This file doesn't exist if !ENABLE_WIDEC (which is the case with
> poky-tiny for example).
> 
> > +
> > +FILES_${PN}-libtinfo = "\
> > +  ${base_libdir}/libtinfo.so.* \
> >  "
> >  
> >  # This keeps only tput/tset in ncurses
> 
> -- 
> Darren Hart
> Intel Open Source Technology Center
> Yocto Project - Linux Kernel
> 
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core

-- 
Martin 'JaMa' Jansa     jabber: Martin.Jansa@gmail.com

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 205 bytes --]

Re: [PATCH 1/1] ncurses: fix packaging of moved libraries

[Paul Eggleton]

On Wednesday 14 March 2012 16:00:33 Martin Jansa wrote:
> isn't just
> ALLOW_EMPTY_${PN}-libncursesw = "1"
> enough?

I suspect we don't care if the libncursesw package isn't created when wide 
character support is disabled, if so, we don't even need that.

Cheers,
Paul

-- 

Paul Eggleton
Intel Open Source Technology Centre

Re: [PATCH 1/1] ncurses: fix packaging of moved libraries

[Darren Hart]

On 03/14/2012 08:07 AM, Paul Eggleton wrote:
> On Wednesday 14 March 2012 16:00:33 Martin Jansa wrote:
>> isn't just
>> ALLOW_EMPTY_${PN}-libncursesw = "1"
>> enough?
> 
> I suspect we don't care if the libncursesw package isn't created when wide 
> character support is disabled, if so, we don't even need that.

OK, so if a package is defined and none of the files exist, bitbake will
simply not create the package - and no errors or warnings will be raised?

If so, then kindly disregard my comments.

-- 
Darren Hart
Intel Open Source Technology Center
Yocto Project - Linux Kernel

Re: [PATCH 1/1] ncurses: fix packaging of moved libraries

[Paul Eggleton]

On Wednesday 14 March 2012 08:18:34 Darren Hart wrote:
> On 03/14/2012 08:07 AM, Paul Eggleton wrote:
> > On Wednesday 14 March 2012 16:00:33 Martin Jansa wrote:
> >> isn't just
> >> ALLOW_EMPTY_${PN}-libncursesw = "1"
> >> enough?
> > 
> > I suspect we don't care if the libncursesw package isn't created when wide
> > character support is disabled, if so, we don't even need that.
> 
> OK, so if a package is defined and none of the files exist, bitbake will
> simply not create the package - and no errors or warnings will be raised?

That's correct.

Cheers,
Paul

-- 

Paul Eggleton
Intel Open Source Technology Centre

Re: [PATCH 1/1] ncurses: fix packaging of moved libraries

[Darren Hart]

On 03/14/2012 08:00 AM, Martin Jansa wrote:
> On Wed, Mar 14, 2012 at 07:57:56AM -0700, Darren Hart wrote:
>>
>>
>> On 03/13/2012 05:31 PM, Scott Garman wrote:

...

>>>  PACKAGES += " \
>>> +  ${PN}-libncurses \
>>> +  ${PN}-libncursesw \
>>
>> This one needs to be dependent on ENABLE_WIDEC
> 
> isn't just
> ALLOW_EMPTY_${PN}-libncursesw = "1"
> enough?

Perhaps. ALLOW_EMPTY is not defined in poky/documentation/* nor in
poky/conf/documentation.conf.

(I realize this oe-core and not poky, but this is the best reference I
know of)

Could you provide a terse description of how ALLOW_EMPTY* is meant to be
used, its effects, and any needful cautionary text? Scott can then add
it to the above documentation.

-- 
Darren Hart
Intel Open Source Technology Center
Yocto Project - Linux Kernel

Re: [PATCH 1/1] ncurses: fix packaging of moved libraries

[Paul Eggleton]

On Wednesday 14 March 2012 08:12:23 Darren Hart wrote:
> Perhaps. ALLOW_EMPTY is not defined in poky/documentation/* nor in
> poky/conf/documentation.conf.
> 
> (I realize this oe-core and not poky, but this is the best reference I
> know of)
> 
> Could you provide a terse description of how ALLOW_EMPTY* is meant to be
> used, its effects, and any needful cautionary text? Scott can then add
> it to the above documentation.

ALLOW_EMPTY (which like all package-controlling variables should always be 
used in conjunction a package name override, e.g. ALLOW_EMPTY_${PN}) just 
specifies that empty packages are actually produced instead of the default 
behaviour where they are skipped. This is mostly useful where there is an 
RDEPENDS or other runtime hard-requirement on the package existing.

Cheers,
Paul

-- 

Paul Eggleton
Intel Open Source Technology Centre

Re: [PATCH 0/1] ncurses packaging fix

[Martin Jansa]

[-- Attachment #1: Type: text/plain, Size: 1186 bytes --]

On Tue, Mar 13, 2012 at 05:31:24PM -0700, Scott Garman wrote:
> Hello,
> 
> This fixes the ncurses library packaging. Thanks to Andreas Oberritter
> for pointing it out.

Why not use another do_split_packages for base_libdir?

Can it be used only once per recipe? are there any libs left in libdir?

Cheers,

> 
> Scott
> 
> The following changes since commit 81379dc141e245307ca391a02b7490b199128991:
> 
>   dates: Fix unpackaged files warning (2012-03-13 23:18:19 +0000)
> 
> are available in the git repository at:
>   git://git.pokylinux.org/poky-contrib sgarman/ncurses-pkg-fix-oe
>   http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=sgarman/ncurses-pkg-fix-oe
> 
> Scott Garman (1):
>   ncurses: fix packaging of moved libraries
> 
>  meta/recipes-core/ncurses/ncurses.inc |   18 ++++++++++++++++--
>  1 files changed, 16 insertions(+), 2 deletions(-)
> 
> -- 
> 1.7.5.4
> 
> 
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core

-- 
Martin 'JaMa' Jansa     jabber: Martin.Jansa@gmail.com

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 205 bytes --]

Re: [PATCH 0/1] ncurses packaging fix

[Scott Garman]

On 03/14/2012 12:51 AM, Martin Jansa wrote:
> On Tue, Mar 13, 2012 at 05:31:24PM -0700, Scott Garman wrote:
>> Hello,
>>
>> This fixes the ncurses library packaging. Thanks to Andreas Oberritter
>> for pointing it out.
>
> Why not use another do_split_packages for base_libdir?
>
> Can it be used only once per recipe? are there any libs left in libdir?

Hi Martin,

This is something worth looking into, but due to the amount of testing I 
feel it would need, I'm going to have to defer it. Our next release is 
coming up fast, and I've been asked to focus on recipe upgrades for the 
rest of this week. I have filed bug #2106 to keep track of it.

https://bugzilla.yoctoproject.org/show_bug.cgi?id=2106

Certainly anyone who has time to do proper testing on this can grab 
ownership of the bug in the meantime.

Scott

-- 
Scott Garman
Embedded Linux Engineer - Yocto Project
Intel Open Source Technology Center

Re: [PATCH 0/1] ncurses packaging fix

[Koen Kooi]

Op 14 mrt. 2012, om 01:31 heeft Scott Garman het volgende geschreven:

> Hello,
> 
> This fixes the ncurses library packaging. Thanks to Andreas Oberritter
> for pointing it out.
> 
> Scott
> 
> The following changes since commit 81379dc141e245307ca391a02b7490b199128991:
> 
>  dates: Fix unpackaged files warning (2012-03-13 23:18:19 +0000)
> 
> are available in the git repository at:
>  git://git.pokylinux.org/poky-contrib sgarman/ncurses-pkg-fix-oe
>  http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=sgarman/ncurses-pkg-fix-oe
> 
> Scott Garman (1):
>  ncurses: fix packaging of moved libraries

You're missing a PR bump for the affected packages.

regards,

Koen

Re: [PATCH 0/1] ncurses packaging fix

[Scott Garman]

On 03/14/2012 01:22 AM, Koen Kooi wrote:
>
> Op 14 mrt. 2012, om 01:31 heeft Scott Garman het volgende geschreven:
>
>> Hello,
>>
>> This fixes the ncurses library packaging. Thanks to Andreas Oberritter
>> for pointing it out.
>>
>> Scott
>>
>> The following changes since commit 81379dc141e245307ca391a02b7490b199128991:
>>
>>   dates: Fix unpackaged files warning (2012-03-13 23:18:19 +0000)
>>
>> are available in the git repository at:
>>   git://git.pokylinux.org/poky-contrib sgarman/ncurses-pkg-fix-oe
>>   http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=sgarman/ncurses-pkg-fix-oe
>>
>> Scott Garman (1):
>>   ncurses: fix packaging of moved libraries
>
> You're missing a PR bump for the affected packages.

My patch included incrementing INC_PR, but I'm guessing you're pointing 
this out because overnight Paul Eggleton got a patch into master which 
incremented INC_PR to the same number as I did in this patch. So I'm 
going to respin the branch with another increase to INC_PR.

If I'm mistaken, please let me know.

Thanks,

Scott

-- 
Scott Garman
Embedded Linux Engineer - Yocto Project
Intel Open Source Technology Center