On 2015-01-15 12:38, Bob Cochran wrote:
On 01/15/2015 12:46 PM, Gary Thomas wrote:
This commit breaks a build when an image contains udev-extraconf

commit d7df92b6f5c1a673341d54d6b064ad2fd7b16d74
Author: Ting Liu <[email protected]>
Date:   Wed Jan 7 17:54:59 2015 +0800

     udev-extraconf: replace bbappend with own recipe

     * introduce variable RULE for DPAA FMan ethernet ports name rule.
     * use 72-fsl-dpaa-persistent-networking.rules for t1024, same as
e6500.

     Signed-off-by: Ting Liu <[email protected]>

Since this change, I get this error when building:
  * check_data_file_clashes: Package udev-rules-qoriq wants to install
file
/local/diskstation-cutting-edge_2015-01-08/tmp/work/p1022ds-amltd-linux-gnuspe/amanda-server-image/1.0-r0/rootfs/etc/udev/mount.blacklist

         But that file is already provided by package  * udev-extraconf


I'm wondering where the root cause is located.  The recipe isn't installing the 
blacklist file.  It's just appending to it (if it already exists), so why is 
there a clash?  Does
the recipe need modification or is the problem in the framework?

One recipe doesn't get to modify the results of a different recipe
so the do_install() step is really creating a second copy of the
blacklist file (which if you don't have udev-extraconf installed
isn't used anyway)


As an aside, should the udev-rules-qoriq recipe see if the blacklists already 
exist before issuing the append (echo) statements?

Lastly, I think I like having the separate recipe rather than the 
udev-extraconf append since I can easily just leave it out of my image, which 
I'm currently doing.

Well I would like to have udev-extraconf in my image (that's
what choices are all about) and this version of udev-rules-qoriq
is making that impossible :-(



I don't see why this was changed from using .bbappend for
the udev-extraconf - it seems wrong headed to me.


--
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------
--
_______________________________________________
meta-freescale mailing list
[email protected]
https://lists.yoctoproject.org/listinfo/meta-freescale

Reply via email to