[Dev] Yocto Tizen - clean, continue option
geoffroy.vancutsem at intel.com
Wed Jan 21 23:21:20 GMT 2015
> -----Original Message-----
> From: Dev [mailto:dev-bounces at lists.tizen.org] On Behalf Of Saul Wold
> Sent: Wednesday, January 21, 2015 3:45 PM
> To: ds23.jung at samsung.com
> Cc: dev at lists.tizen.org; 유덕인
> Subject: Re: [Dev] Yocto Tizen - clean, continue option
> On 01/20/2015 08:36 PM, 정대순 wrote:
> > Hi,
> > In before e-mail, My questions are separation.
> > In firt question, I want to clean to target.(all packages)
> > So, I input 'bitbake -c cleanall tizen-common-core-image-crosswalk-dev', and
> > 'bitbake tizen-common-core-image-crosswalk-dev'.
> > But, build started 5748 task. (msg : Currently 1 running tasks (5749 of 5752)),
> > not first task.
> > Is it right? In my opinion, just clean some of packages.
> Not quite, the 3 clean tasks (clean, cleansstate, cleanall) are tasks
> against recipes or targets, they don't affect the dependencies of the
> target. So the above cleanall just cleans the specific target listed on
> the command line.
I have been struggling with this concept a bit myself. The standard clean option clears the workspace and sysroot (and any artifact that's in there), but what does that mean when the target is something different than a package (i.e. a complete image such as tizen-ivi-core)?
> > (When I run the build after git clone, started '1 of 5752' task.)
> > How to build started in init status? Delete a 'tmp-glibc' directory?
> I would ask why do you need to start from the "initial" state, if you
> change a recipe it will rebuild that target. Deleting the tmp-glibc is a
> pretty big hammer.
My own trick is to initialise a new different build dir by re-running the script... that's also a pretty big hammer :-)
There are times when you don't know if or up to what point you can trust the cache... this is when I feel I need to go back to an initial state.
> > In second question, I just curious about build break.
> > When build break occur, if I run 'bitbake tizen-common-core-image-
> > again, does it continue break point?
> It will continue, from the failure, but unless you fix the failure it
> will fail again!
This is exactly what I saw the other day. What confused me quite a bit (and made me go back to the initial state) is that when I enabled more debug output (-D), it proceeded to building more packages after the rpm-native failure, something it had never done before. Maybe this is the normal behaviour but it's not very intuitive to the Yocto newbie that I am (and yes, I still have to read most of the documentation on Yocto).
Intel Corporation NV/SA
Kings Square, Veldkant 31
RPM (Bruxelles) 0415.497.718.
Citibank, Brussels, account 570/1031255/09
More information about the Dev