[oe] MACHINE_KERNEL_PR overrides PR for modules

[Frans Meulenbroeks]

2011/2/20 Denis Dydychkin <nyrl at mail.ru>:
>
>
> When recipe inherits module.bbclass, it will get its PR overwritten if
> MACHINE_KERNEL_PR is defined. The outcome is that bumping PR within recipe has
> no effect, as well as SRCREV does not work anymore, which is especially
> important when working with latest version from repos.
>
> The following patch fixes this problem by appending MACHINE_KERNEL_PR to PR
> instead of replacing it. "+k" is added to get resulting PR looks like "r1+kr22"
> or "r1+svnr222+kr33" since MACHINE_KERNEL_PR is usually defined as "rXX".
>
>
> --- classes/module-base.bbclass 2011-02-17 16:14:45.792425264 +0300
> +++ classes/module-base.bbclass 2011-02-17 16:36:00.088425246 +0300
> @@ -11,7 +11,8 @@
>     machine_kernel_pr = bb.data.getVar('MACHINE_KERNEL_PR', d, True)
>
>     if machine_kernel_pr:
> -       bb.data.setVar('PR', machine_kernel_pr, d)
> +       pr = bb.data.getVar('PR', d, True) + "+k" + machine_kernel_pr
> +       bb.data.setVar('PR', pr, d)
>  }
>
>  export KERNEL_VERSION = "${@base_read_file('${STAGING_KERNEL_DIR}/kernel-abiversion')}"

Actually also just noticed this problem after building iscsi-target
for openrd (r19) and calamari (which to my surprise gave r1)

The patch does look ok to me, but I can imagine that we also want to
bump PE or so (to avoid that the new PR from this code is smaller than
the old PR
(acutally thinking about it, if we put kernel pr first (without the k)
and then something like m followed by PR, it might not require bumping
PE).

Frans