[PATCH] sstatesig: don't squash dependencies of useradd packages

[PATCH] sstatesig: don't squash dependencies of useradd packages

[Maxin B. John]

Without this change, sstate dependency filter squashes dependencies
of packages which inherits useradd class. That causes failures while
rebuilding it for another MACHINE.

Fixes [YOCTO #8078]

Signed-off-by: Maxin B. John <maxin.john@intel.com>
---
 meta/lib/oe/sstatesig.py | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/meta/lib/oe/sstatesig.py b/meta/lib/oe/sstatesig.py
index 8224e3a..88b2290 100644
--- a/meta/lib/oe/sstatesig.py
+++ b/meta/lib/oe/sstatesig.py
@@ -17,6 +17,9 @@ def sstate_rundepfilter(siggen, fn, recipename, task, dep, depname, dataCache):
     def isAllArch(fn):
         inherits = " ".join(dataCache.inherits[fn])
         return "/allarch.bbclass" in inherits
+    def isUserAdd(fn):
+        inherits = " ".join(dataCache.inherits[fn])
+        return "/useradd.bbclass" in inherits
     def isImage(fn):
         return "/image.bbclass" in " ".join(dataCache.inherits[fn])
 
@@ -39,6 +42,10 @@ def sstate_rundepfilter(siggen, fn, recipename, task, dep, depname, dataCache):
 
     # Only target packages beyond here
 
+    # Don't squash the dependencies of packages inheritting useradd class
+    if  isUserAdd(fn):
+        return True
+
     # allarch packagegroups are assumed to have well behaved names which don't change between architecures/tunes
     if isPackageGroup(fn) and isAllArch(fn) and not isNative(depname):
         return False  
-- 
2.4.0

Re: [PATCH] sstatesig: don't squash dependencies of useradd packages

[Maxin B. John]

Ping.

On Mon, Dec 12, 2016 at 03:48:04PM +0200, Maxin B. John wrote:
> Without this change, sstate dependency filter squashes dependencies
> of packages which inherits useradd class. That causes failures while
> rebuilding it for another MACHINE.
> 
> Fixes [YOCTO #8078]
> 
> Signed-off-by: Maxin B. John <maxin.john@intel.com>
> ---
>  meta/lib/oe/sstatesig.py | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/meta/lib/oe/sstatesig.py b/meta/lib/oe/sstatesig.py
> index 8224e3a..88b2290 100644
> --- a/meta/lib/oe/sstatesig.py
> +++ b/meta/lib/oe/sstatesig.py
> @@ -17,6 +17,9 @@ def sstate_rundepfilter(siggen, fn, recipename, task, dep, depname, dataCache):
>      def isAllArch(fn):
>          inherits = " ".join(dataCache.inherits[fn])
>          return "/allarch.bbclass" in inherits
> +    def isUserAdd(fn):
> +        inherits = " ".join(dataCache.inherits[fn])
> +        return "/useradd.bbclass" in inherits
>      def isImage(fn):
>          return "/image.bbclass" in " ".join(dataCache.inherits[fn])
>  
> @@ -39,6 +42,10 @@ def sstate_rundepfilter(siggen, fn, recipename, task, dep, depname, dataCache):
>  
>      # Only target packages beyond here
>  
> +    # Don't squash the dependencies of packages inheritting useradd class
> +    if  isUserAdd(fn):
> +        return True
> +
>      # allarch packagegroups are assumed to have well behaved names which don't change between architecures/tunes
>      if isPackageGroup(fn) and isAllArch(fn) and not isNative(depname):
>          return False  
> -- 
> 2.4.0
> 
> -- 
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [PATCH] sstatesig: don't squash dependencies of useradd packages

[Richard Purdie]

On Mon, 2017-01-23 at 16:55 +0200, Maxin B. John wrote:
> Ping.

This never merged as I was never 100% convinced this was the correct
way to solve this problem. I'd be interested to understand if the
sstatesig tests in oe-selftest worked after this.

I'm hoping this issue has also now been resolved in a different way
though the switch to recipe specific sysroots. Would you be able test
if recipe specific sysroots have addressed this issue?

Cheers,

Richard

Re: [PATCH] sstatesig: don't squash dependencies of useradd packages

[Maxin B. John]

Hi RP,

On Mon, Jan 23, 2017 at 03:59:31PM +0000, Richard Purdie wrote:
> On Mon, 2017-01-23 at 16:55 +0200, Maxin B. John wrote:
> > Ping.
> 
> This never merged as I was never 100% convinced this was the correct
> way to solve this problem. I'd be interested to understand if the
> sstatesig tests in oe-selftest worked after this.
>
> I'm hoping this issue has also now been resolved in a different way
> though the switch to recipe specific sysroots. Would you be able test
> if recipe specific sysroots have addressed this issue?

Thanks for looking into this. Tested this (bug #8078) on the latest master
with recipe specific sysroots modification and found that the error still exists.  

> Cheers,
> Richard

Best Regards,
Maxin