Patchwork apr: add libtool to SSTATE_SCAN_FILES

login
register
mail settings
Submitter Randy MacLeod
Date March 4, 2013, 1:34 a.m.
Message ID <1362360870-12148-1-git-send-email-randy.macleod@windriver.com>
Download mbox | patch
Permalink /patch/45457/
State Accepted
Commit a2d97d559815a639a6852d5c7ac830489fe30b32
Headers show

Comments

Randy MacLeod - March 4, 2013, 1:34 a.m.
From: Jackie Huang <jackie.huang@windriver.com>

The hardcoded paths in libtool will cause apache2 compile fail at
link stage when the package apr is installed from sstate cache, so
add libtool to SSTATE_SCAN_FILES.

[ CQID: WIND00405372 ]

Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
---
 meta/recipes-support/apr/apr_1.4.6.bb | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
Martin Jansa - March 4, 2013, 8:18 a.m.
On Sun, Mar 03, 2013 at 08:34:30PM -0500, randy.macleod@windriver.com wrote:
> From: Jackie Huang <jackie.huang@windriver.com>
> 
> The hardcoded paths in libtool will cause apache2 compile fail at
> link stage when the package apr is installed from sstate cache, so
> add libtool to SSTATE_SCAN_FILES.
> 
> [ CQID: WIND00405372 ]

You can drop this.

> 
> Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
> Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
> ---
>  meta/recipes-support/apr/apr_1.4.6.bb | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/meta/recipes-support/apr/apr_1.4.6.bb b/meta/recipes-support/apr/apr_1.4.6.bb
> index b7220e0..77dffee 100644
> --- a/meta/recipes-support/apr/apr_1.4.6.bb
> +++ b/meta/recipes-support/apr/apr_1.4.6.bb
> @@ -9,7 +9,7 @@ LIC_FILES_CHKSUM = "file://LICENSE;md5=0c35ff3c4c83b89d2f076e315caac28b \
>  
>  BBCLASSEXTEND = "native"
>  
> -PR = "r1"
> +PR = "r2"
>  
>  SRC_URI = "${APACHE_MIRROR}/apr/${BPN}-${PV}.tar.bz2 \
>             file://configure_fixes.patch \
> @@ -43,7 +43,7 @@ do_install_append() {
>  	cp ${S}/${HOST_SYS}-libtool ${D}${datadir}/build-1/libtool
>  }
>  
> -SSTATE_SCAN_FILES += "apr_rules.mk"
> +SSTATE_SCAN_FILES += "apr_rules.mk libtool"
>  
>  SYSROOT_PREPROCESS_FUNCS += "apr_sysroot_preprocess"
>  
> -- 
> 1.8.1
> 
> 
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Richard Purdie - March 4, 2013, 10:11 a.m.
On Mon, 2013-03-04 at 09:18 +0100, Martin Jansa wrote:
> On Sun, Mar 03, 2013 at 08:34:30PM -0500, randy.macleod@windriver.com wrote:
> > From: Jackie Huang <jackie.huang@windriver.com>
> > 
> > The hardcoded paths in libtool will cause apache2 compile fail at
> > link stage when the package apr is installed from sstate cache, so
> > add libtool to SSTATE_SCAN_FILES.
> > 
> > [ CQID: WIND00405372 ]
> 
> You can drop this.

Is this really a big problem? I can understand people wanting to be able
to track commits back to defect reports...

Cheers,

Richard
Martin Jansa - March 4, 2013, 10:33 a.m.
On Mon, Mar 04, 2013 at 10:11:27AM +0000, Richard Purdie wrote:
> On Mon, 2013-03-04 at 09:18 +0100, Martin Jansa wrote:
> > On Sun, Mar 03, 2013 at 08:34:30PM -0500, randy.macleod@windriver.com wrote:
> > > From: Jackie Huang <jackie.huang@windriver.com>
> > > 
> > > The hardcoded paths in libtool will cause apache2 compile fail at
> > > link stage when the package apr is installed from sstate cache, so
> > > add libtool to SSTATE_SCAN_FILES.
> > > 
> > > [ CQID: WIND00405372 ]
> > 
> > You can drop this.
> 
> Is this really a big problem? I can understand people wanting to be able
> to track commits back to defect reports...

Not big problem, but isn't link to cgit in their tracking system enough?

That's what I'm using for SHR bugs now, because I cannot assume that people
using oe-core will know where bug tracker for [SHR #666] is.

I cannot find public windriver bug tracker where I can search 
WIND00405372 to find longer description why this change is needed (I
know why in this case) - that's why I find yocto bug tags useful, but
not WIND00*.
Richard Purdie - March 4, 2013, 10:38 a.m.
On Mon, 2013-03-04 at 11:33 +0100, Martin Jansa wrote:
> On Mon, Mar 04, 2013 at 10:11:27AM +0000, Richard Purdie wrote:
> > On Mon, 2013-03-04 at 09:18 +0100, Martin Jansa wrote:
> > > On Sun, Mar 03, 2013 at 08:34:30PM -0500, randy.macleod@windriver.com wrote:
> > > > From: Jackie Huang <jackie.huang@windriver.com>
> > > > 
> > > > The hardcoded paths in libtool will cause apache2 compile fail at
> > > > link stage when the package apr is installed from sstate cache, so
> > > > add libtool to SSTATE_SCAN_FILES.
> > > > 
> > > > [ CQID: WIND00405372 ]
> > > 
> > > You can drop this.
> > 
> > Is this really a big problem? I can understand people wanting to be able
> > to track commits back to defect reports...
> 
> Not big problem, but isn't link to cgit in their tracking system enough?
> 
> That's what I'm using for SHR bugs now, because I cannot assume that people
> using oe-core will know where bug tracker for [SHR #666] is.

I am quite happy for the meaning of some of these to be documented
somewhere with a link to the appropriate bugzillas or other systems.

> I cannot find public windriver bug tracker where I can search 
> WIND00405372 to find longer description why this change is needed (I
> know why in this case) - that's why I find yocto bug tags useful, but
> not WIND00*.

As per the commit guidelines, these are additional information and no
excuse for not writing a self-standing commit message. Some will be more
useful than others but I'm not sure it makes sense to remove them just
for the sake of it.

Cheers,

Richard
Khem Raj - March 6, 2013, 3:13 a.m.
On Mon, Mar 4, 2013 at 2:11 AM, Richard Purdie
<richard.purdie@linuxfoundation.org> wrote:
>> > [ CQID: WIND00405372 ]
>>
>> You can drop this.
>
> Is this really a big problem? I can understand people wanting to be able
> to track commits back to defect reports...

it would not make sense for most of people in community. Having
reference to yocto project
bug reports is still some information since people can go and look at the bug.
 each one of us have one or another way of internal tracking,
Trevor Woerner - March 6, 2013, 4:04 p.m.
Ah... so CQID is an internal WindRiver thing?
Mark Hatle - March 6, 2013, 6:07 p.m.
On 3/6/13 10:04 AM, Trevor Woerner wrote:
> Ah... so CQID is an internal WindRiver thing?

Correct.  This is our internal bug tracking number.  (Our customers also get 
access to some of the numbers as well..)

We include this number of items we've ported from our product back to oe-core so 
that when we uprev it is easier for us to verify that certain problems have been 
corrected.

If the community decided these numbers are not acceptable we'll stop including 
them.  (If you note, not all of our contributions include those numbers.. 
generally only bug fixes do.)

--Mark

> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
>

Patch

diff --git a/meta/recipes-support/apr/apr_1.4.6.bb b/meta/recipes-support/apr/apr_1.4.6.bb
index b7220e0..77dffee 100644
--- a/meta/recipes-support/apr/apr_1.4.6.bb
+++ b/meta/recipes-support/apr/apr_1.4.6.bb
@@ -9,7 +9,7 @@  LIC_FILES_CHKSUM = "file://LICENSE;md5=0c35ff3c4c83b89d2f076e315caac28b \
 
 BBCLASSEXTEND = "native"
 
-PR = "r1"
+PR = "r2"
 
 SRC_URI = "${APACHE_MIRROR}/apr/${BPN}-${PV}.tar.bz2 \
            file://configure_fixes.patch \
@@ -43,7 +43,7 @@  do_install_append() {
 	cp ${S}/${HOST_SYS}-libtool ${D}${datadir}/build-1/libtool
 }
 
-SSTATE_SCAN_FILES += "apr_rules.mk"
+SSTATE_SCAN_FILES += "apr_rules.mk libtool"
 
 SYSROOT_PREPROCESS_FUNCS += "apr_sysroot_preprocess"