Op 22 sep. 2011, om 10:58 heeft Koen Kooi het volgende geschreven:
> When enabled fetch2 will generate tarballs with the SRCREV embbeded like so:
>
>
> git2_git.pingu.fi.xf86-video-omapfb_28c006c94e57ea71df11ec4fff79d7ffcfc4860f.tar.gz
>
> Signed-off-by: Koen Kooi <k...@dominion.thruhere.net>
> ---
> lib/bb/fetch2/git.py | 2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/lib/bb/fetch2/git.py b/lib/bb/fetch2/git.py
> index fb6125c..f889562 100644
> --- a/lib/bb/fetch2/git.py
> +++ b/lib/bb/fetch2/git.py
> @@ -93,7 +93,7 @@ class Git(FetchMethod):
>
> ud.nocheckout = ud.parm.get("nocheckout","0") == "1"
>
> - ud.rebaseable = ud.parm.get("rebaseable","0") == "1"
> + ud.rebaseable = ud.parm.get("rebaseable","0") == "1" or
> ((data.getVar("BB_GENERATE_MIRROR_VERSIONED_TARBALLS", d, True) or "0") !=
> "0")
Apart from the objections raised by Richard I've found 2 other problems with it:
1) it will force a cache invalidation each build
2) fetch2 witll *always* try to access the remote git repo, even if the SRCREV
is already present in the local copy
Those aren't really bugs in my patch, but in the rebasable fetch2 code.
_______________________________________________
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
