Re: [PATCH 1/1] bitbake: runqueue: terminate build safely while RunQueueExecuteScenequeue init failed

[Richard Purdie <richard.purdie@linuxfoundation.org>]

On Wed, 2014-08-27 at 18:44 +0800, Hongxu Jia wrote:
> While RunQueueExecuteScenequeue init failed, the exit of build is mess.
> Here is the example while bb.fatal invoked in RunQueueExecuteScenequeue:
> ...
> Traceback (most recent call last):
> |  File "/home/jiahongxu/yocto/poky/bitbake/lib/bb/runqueue.py",
>   line 1824, in RunQueueExecuteScenequeue.__init__(rq=<bb.runqueue.RunQueue
>   instance at 0x7f87dd7d5050>):
> |                 locs = { "sq_fn" : sq_fn, "sq_task" : sq_taskname,
>   "sq_hash" : sq_hash, "sq_hashfn" : sq_hashfn, "d" : self.cooker.data }
> |    >            valid = bb.utils.better_eval(call, locs)
> 
> |    def better_eval(source, locals):
> |    >    return eval(source, get_context(), locals)
> 
> |  File "<string>", line 1, in <module>
> |  File "/home/jiahongxu/yocto/poky/bitbake/lib/bb/__init__.py", line 102,
>   in fatal:
> |         logger.critical(''.join(args))
> |    >    raise BBHandledException()
> 
> BBHandledException
> ...
> 
> We should terminate build safely while RunQueueExecuteScenequeue init failed.
> 
> Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
> ---
>  bitbake/lib/bb/runqueue.py | 7 ++++++-
>  1 file changed, 6 insertions(+), 1 deletion(-)
> 
> diff --git a/bitbake/lib/bb/runqueue.py b/bitbake/lib/bb/runqueue.py
> index e13dc57..c7b1dc0 100644
> --- a/bitbake/lib/bb/runqueue.py
> +++ b/bitbake/lib/bb/runqueue.py
> @@ -1046,7 +1046,12 @@ class RunQueue:
>                  self.state = runQueueComplete
>              else:
>                  self.start_worker()
> -                self.rqexe = RunQueueExecuteScenequeue(self)
> +                try:
> +                    self.rqexe = RunQueueExecuteScenequeue(self)
> +                except:
> +                    # The RunQueueExecuteScenequeue init failure
> +                    # terminated the build safely
> +                    self.state = runQueueComplete
>  

If the init fails, shouldn't we show a traceback and exit with an error
status? Does this happen in a normal build and if so, is there an
underlying cause we should fix?

I'm worried this would let errors go unnoticed. Except clauses in
general should catch specific errors not every exception so this really
can't merge in this form, sorry.

Cheers,

Richard