Re: OE recipe tree quality

[Thomas Zimmermann <ml@vdm-design.de>]

Am Donnerstag 29 Juli 2010, 10:50:53 schrieb Frans Meulenbroeks:
> Dear all,
> 
> Given the discussions on quality that sometimes pop up (and also triggered
> by Robert's message), I decided to kick off a bitbake -k world.
> Below are some initial findings. I would like to use this as a starter to
> kickoff the discusson on quality and what to do with broken stuff.
> And also I hope that people who feel attached to a recipe mentioned below
> will take action.
> 
> My setting:
> local.conf:
> IMAGE_LINGUAS = ""
> ENABLE_BINARY_LOCALE_GENERATION = "0"
> 
> MACHINE = "beagleboard"
> DISTRO = "angstrom-2008.1"
> 
> EXTRA_IMAGECMD_jffs2 = "--pad --eraseblock=0x20000"
> MKUBIFS_ARGS = "-m 2048 -e 129024 -c 1024"
> 
> IMAGE_FSTYPES = "tar squashfs ubifs jffs2"
> IMAGE_KEEPROOTFS = "1"
> OE_ALLOW_INSECURE_DOWNLOADS = "yes"
> 
> I opted for angstromg and beagleboard as I feel this is the most often
> build configuration, so it should have the least issues.
> host is ubuntu 10.04 on 64 bit. tree is oe dev git head (as of jul 29)
> 
> All results are not in yet (there are 71047 taksks, I am at 159 :-)
> However the parsing already gave some interesting output.
> I've ordered them somehwat:
> 
> Nothing PROVIDES errors
> ===================
> ...
Some of these 'Nothing PROVIDES' errors are because you are using angstrom as 
distro and some recipes are blacklisted there (e.g. fso-apm), because of that 
a lot of the shr specific recipes aren't buildable.

I think one should do such a test with minimal distro because nothing is 
blacklisted there.

Other recipes like etk or epsilon were moved to obsolete, but recipes 
depending on them weren't. So i think recipes depending on etk and epsilon can 
be moved to obsolete too.

Regards
Thomas