**** BEGIN LOGGING AT Mon Aug 19 02:59:59 2013
Aug 19 09:46:10 <ant_work> bluelightning: hello
Aug 19 09:46:27 <bluelightning> hi ant_work
Aug 19 09:47:09 <ant_work> I have in front of me the logs of some days ago about rebuilding failures from sstate
Aug 19 09:47:32 <ant_work> I have a fresh build as well so later I'll plan to debug that
Aug 19 09:47:44 <ant_work> my question for you:
Aug 19 09:48:30 <ant_work> the case was about klibc and the ipk control file does specify it Depends on linux-libc-headers
Aug 19 09:49:04 <ant_work> now, the produced package is called linux-libc-headers-dev
Aug 19 09:49:30 <ant_work> it's control file inside the ipk doesn't specify any Depends
Aug 19 09:49:50 <ant_work> isn't it strange?
Aug 19 09:51:04 <bluelightning> doesn't sound right to me
Aug 19 09:53:59 <ant_work> the recipe has just DEPENDS += "unifdef-native"
Aug 19 09:54:18 <ant_work> argh
Aug 19 09:54:19 <ant_work> INHIBIT_DEFAULT_DEPS = "1"
Aug 19 09:54:30 <ant_work> spotted
Aug 19 09:55:33 <ant_work> then I have to find an additional target to add tc as dependency of klibc
Aug 19 09:56:54 <ant_work> or just remove that INHIBIT
Aug 19 09:57:04 <ant_work> introduced by Bruce # 	ccf3bf70a1a83ac55e47acfa4759838f7f4834ae
Aug 19 09:57:42 <bluelightning> oh, we're talking the dependencies of linux-libc-headers  itself ?
Aug 19 09:57:49 <ant_work> yes
Aug 19 09:58:15 <ant_work> I imagine one has already th etoolchain if he needs th eextra headers ;)
Aug 19 09:58:37 <bluelightning> does it actually need the specific libc to be used?
Aug 19 09:59:53 <ant_work> no, thi sis a sanitized form of kernel's make headers_install
Aug 19 10:00:02 <ant_work> imho is agnostic
Aug 19 10:00:54 <bluelightning> then, what is missing?
Aug 19 10:01:49 <ant_work> some basic headers (iirc limits.h)
Aug 19 10:07:01 <ant_work> I have not yet built klibc so thi stime I'll try to rebuild linux-libc-headers from sstate and log its failure
Aug 19 10:08:42 <ant_work> last I've looked into it, the code in base.bbclass et al. did not add linux-libc-headers as extradep
Aug 19 10:09:01 <ant_work> so it is not considered part of the standard toolchain
Aug 19 10:09:07 <ant_work> (in debian world)
Aug 19 10:09:34 <ant_work> Gentoo is different, I have to make my mind about that but finally I think this INHIBIT in the recipe can be removed
Aug 19 10:11:40 <ant_work> ofc your'e free to dig in the code building "DEFAULT_DEPS" ;
Aug 19 16:27:06 <ant_home> bluelightning: wrt wiping tmpdir, what about sstate-control ?
Aug 19 16:27:35 <bluelightning> ant_home: not sure...
Aug 19 16:27:48 <bluelightning> depends if anything looks at that in the absence of a stamp
Aug 19 16:27:54 <ant_home> usually I blank tmp-eglibc
Aug 19 16:28:19 <ant_home> sstate-cache is out as per defaults
Aug 19 16:29:01 <ant_home> seems there are only manifests
Aug 19 16:29:23 <ant_home> Gentoo handles recipe + digest together
Aug 19 16:29:47 <ant_home> err MAnifest
Aug 19 16:31:17 <ant_home> ok, I'll keep it for the first rebuild of linux-libc-headers
Aug 19 16:39:30 <ant_home> pointless
Aug 19 16:43:27 <ant_home> ok, rebuilds cleanly
Aug 19 16:43:33 <ant_home> now klibc
Aug 19 16:45:09 <ant_home> succeeded
Aug 19 16:45:12 <ant_home> hmm
**** ENDING LOGGING AT Tue Aug 20 02:59:59 2013