FILE_PR and my requirements Re: Some open issues

[Holger Freyther <zecke@selfish.org>]

On Saturday 18 October 2008 14:32:37 Holger Freyther wrote:

More in depth:

Issues with the current approach:
	- OE => poky is more hard to sync as one needs to watch PR
	- We need to make sure that people do not introduce PR but use FILE_PR

Alternatives:
	- Python black magic(*)
	- A new set of variables + bitbake change

Alternative Python black magic:
	- I strongly oppose any new transparent python magic. I constantly see people 
confused by what we do behind the scenes. I want that PF (so WORKDIR) 
containts the PN, PV, 'PR' of the resulting package. It can not be that you 
have a PR r10.1 as package but only a r10 in your work directory.
	- If I'm not mistaken about the above I will not be happy with such a 
solution


Alternative a new set of variables:
	- Revert to make PR as it was before. This adds legacy...
	- Introduce PKG_PR which is the composition of PR and DISTRO_PR
	- Change PF to be PN-PV-PR (this influences the stamps)
	- Change Bitbake to cache PKG_PR and prefer it over PR.


comments?