On Jan 31, 2013, at 1:17 PM, Harvey Chapman <hchapman-oec...@3gfp.com> wrote:
> On Jan 31, 2013, at 9:45 AM, "Robert P. J. Day" <rpj...@crashcourse.ca> wrote:
>
>> and i'm trying to build a core-image-minimal. with regular
>> (git-allowed) net access, the build works fine. but i haven't yet
>> found any combination of BB_SRCREV_POLICY or BB_FETCH_PREMIRRORONLY or
>> anything else that let's me "fetchall", then disconnect from the net
>> to do the build.
>
> This really seems like a bug in bitbake/lib/bb/fetch2/git.py to me. It should
> honor BB_NO_NETWORK by trying to see if it has the tag locally before it runs
> the ls-remote to see if the tag has changed. I was going to suggest this
> earlier in the thread, but due to my relatively new experience with layered
> OE and the newer bitbake, I refrained. However, I was just inside that code
> updating the mercurial.py fetcher to actually work for file:// urls. At some
> point in the next few weeks, I'll update it further to try and be on par with
> git.py and to at least match the functionality since mercurial.py does things
> differently. It makes it difficult to understand them both if you only know
> how to use one.
I made some minor tweaks to git.py so _latest_revision() now respects
BB_NO_NETWORK. However, I noticed that fetch2.latest_revision() contains:
revs = bb.persist_data.persist('BB_URI_HEADREVS', d)
key = self.generate_revision_key(url, ud, d, name)
try:
return revs[key]
except KeyError:
revs[key] = rev = self._latest_revision(url, ud, d, name)
return rev
What does BB_URI_HEADREVS do and could it be part of the solution to this
problem?
_______________________________________________
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
