what OE/YP layers should be considered "supported"?


Robert P. J. Day
 

related to something that richard purdie mentioned on the OE list,
if one wanted to do a YP-wide "cleanup" of some indeterminate form,
what are the layers that would be considered mandatory to cover in
such a cleanup?

no-brainers would, of course, include:

* oe-core
* meta-openembedded

beyond that, what else? i would think:

* meta-virtualization
* meta-java
* meta-security
* ... more? ...

and there are the vendor layers:

* meta-intel
* meta-freescale
* meta-qcom
* meta-boundary

where can one stop?

rday


Armin Kuster
 

Helllo Robert,

On 5/4/21 2:03 PM, Robert P. J. Day wrote:
related to something that richard purdie mentioned on the OE list,
if one wanted to do a YP-wide "cleanup" of some indeterminate form,
what are the layers that would be considered mandatory to cover in
such a cleanup?
I don't have an context in the email you are referring to. Can you
include it please?
Supported will mean different things to different people.

-Armin
no-brainers would, of course, include:

* oe-core
* meta-openembedded

beyond that, what else? i would think:

* meta-virtualization
* meta-java
* meta-security
* ... more? ...

and there are the vendor layers:

* meta-intel
* meta-freescale
* meta-qcom
* meta-boundary

where can one stop?

rday



Robert P. J. Day
 

On Tue, 4 May 2021, akuster808 wrote:

Helllo Robert,

On 5/4/21 2:03 PM, Robert P. J. Day wrote:
related to something that richard purdie mentioned on the OE list,
if one wanted to do a YP-wide "cleanup" of some indeterminate form,
what are the layers that would be considered mandatory to cover in
such a cleanup?
I don't have an context in the email you are referring to. Can you
include it please? Supported will mean different things to different
people.

-Armin
granted, that Q was a bit vague ... based on a suggestion of
richard's, i was going to do an "audit" of OE/YP layers to see the
effect of doing a particular minor cleanup, but it's not clear the
requirement for wide-sweepingness that would represent a sufficient
audit.

would that necessarily cover every single layer identified at
layers.openembedded.org? that seems a bit onerous. so is there some
moderately vague definition of what constitutes a minimally
representative collection of layers?

rday