-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Paul Sokolovsky schreef:

> -K_MAJOR = "2"
> -K_MINOR = "6"
> -K_MICRO = "15"
> -HHV     = "0"
> +K_MAJOR ?= "2"
> +K_MINOR ?= "6"
> +K_MICRO ?= "15"
> +HHV     ?= "0"
> 
>   Of course, ideally it should be given just CVS date and be able to
> infer corresponding kernel version from the checkout, but I'm not sure
> it can be done with OE, as there's going to be circular dependency -
> it must provide PV to be selected for specific build, and cannot (and
> of course shouldn't) do any actions (like checkout) unless it is
> selected.

A recipe with a floating cvs is is supposed to be valid against the scm
repo at that moment (give or take a day, thanks to timezones), so the
version should match what it has in cvs now, and you'll have no problem
developing.

However...

... if you lock down a cvsdate (for whatever reason) this is only valid
if you keep the file 'stale'. By keeping the file 'stale' you prevent
people from working against latest cvs and run into the versioning and
therefore dependency problems you have.

Basically you have to proper options to solve that:

1) remove the SRCDATE lock for the kernel, or slide it forward, while
updating the recipe to match 'todays' cvs
2) add a snapshot recipe that build for a fixed that (e.g. when the
patches went in)

This problem illustrates why I don't like locking down SRCDATEs in
config files; it may work, but was never intended to. I had a discussion
with Richard about SRCDATE, time-limited patches and reproducability on
#oe this afternoon. We still disagree, but we made a compromize and are
going to work from there :)
With the new bitbake, files are cheap so you can split out the common
parts of the kernel recipe into a .inc and make your own snapshot. I
still think it's bad form to influence package versions from a
MACHINE.conf and consider that, and the cause a bug.

This also illustrates why it is usefull to have a development branch
(aptly named .dev in OE) and a release branch, so you can cater the
bleeding edge people and have proper, stable release management at the
same time.

I hope this clears up why I called the bugs bugs and why I won't apply
that patch to .dev.
I'd also like to mention that problems like this (modules missing deps)
are usually identified and solved within minutes if you mention it on
#oe or a bit slower if you mail the mailinglist. It *really* pays off to
touch base with the OpenEmbedded project :)

regards,

Koen


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (Darwin)

iD8DBQFEsVidMkyGM64RGpERAvIqAJwLOSa2Q/MFauQ1xUeH5be3uHXeCQCdG5tG
UBKw3A7/9kUaN7Gasde41Rc=
=179z
-----END PGP SIGNATURE-----

_______________________________________________
Oe mailing list
[email protected]
https://www.handhelds.org/mailman/listinfo/oe

Reply via email to