[poky] Errors attempting to build AT91SAM9x5 BSP

[Paul Eggleton]

On Tuesday 13 November 2012 14:17:06 Bryan Evenson wrote:
> Thanks for the information.  From my reading of the BSP manual I'd thought I
> was supposed to increment the LCONF_VERSION each time I made a change to
> the bblayers.conf file; I appreciate the clarification.

Could you suggest a change to the BSP manual that would help avoid this 
confusion?
 
> I searched for the error message I was getting, and I think it was
> originating in this section of sanity.bbclass:
> 
>     # Check bblayers.conf is valid
>     current_lconf = sanity_data.getVar('LCONF_VERSION', True)
>     lconf_version = sanity_data.getVar('LAYER_CONF_VERSION', True)
>     if current_lconf != lconf_version:
>         try:
>             bb.build.exec_func("check_bblayers_conf", sanity_data)
>             if sanity_data.getVar("SANITY_USE_EVENTS", True) == "1":
>                 bb.event.fire(bb.event.SanityCheckFailed("Your
> conf/bblayers.conf has been automatically updated. Please close and
> re-run."), sanity_data) return
>             else:
>                 bb.note("Your conf/bblayers.conf has been automatically
> updated. Please re-run %s." % os.path.basename(sys.argv[0])) sys.exit(0)
>         except Exception:
>             messages = messages + "Your version of bblayers.conf was
> generated from an older version of bblayers.conf.sample and there have been
> updates made to this file. Please compare the two files and merge any
> changes before continuing.\nMatching the version numbers will remove this
> message.\n\"meld conf/bblayers.conf
> ${COREBASE}/meta*/conf/bblayers.conf.sample\" is a good way to visualise
> the changes.\n"

Right, I guess (without having tested it recently) what I expected to happen 
was the check_bblayers_conf function would have raised an exception if it 
couldn't do anything with the file; currently it doesn't.

> Would it be helpful if the current_lconf != lconf_version check changed to
> have different messages when current_lconf > lconf_version and when
> current_lcong < lconf_version?

I think that would be a good idea, yes.

Cheers,
Paul

-- 

Paul Eggleton
Intel Open Source Technology Centre