Bug 1186 - Recent bitbake change results in git fetch during parse
Summary: Recent bitbake change results in git fetch during parse
Status: RESOLVED FIXED
Alias: None
Product: BitBake
Classification: Build System, Metadata & Runtime
Component: bitbake (show other bugs)
Version: unspecified
Hardware: All Multiple
: Undecided critical
Target Milestone: ---
Assignee: Paul Eggleton
QA Contact:
URL:
Whiteboard:
Depends on:
Blocks:
 
Reported: 2011-06-21 09:21 UTC by Paul Eggleton
Modified: 2011-06-24 10:08 UTC (History)
4 users (show)

See Also:
OS type for building Yocto: ---
Type of Regression: ---
Verified:
Documentation change: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Paul Eggleton 2011-06-21 09:21:28 UTC
With my change 81e2f52fb8e (bitbake upstream commit 7d2363f3535) we are now expanding PROVIDES for skipped recipes. One of those skipped recipes is u-boot, expanding PROVIDES expands SRCPV, and unfortunately because u-boot is fetched from git and its SRCPV isn't defined at that stage, it attempts to do a fetch to find out the latest revision. This has resulted in bitbake freezing during parsing on autobuilder03. (Thanks to Dexuan for unravelling the above).

u-boot gets skipped for most of the machines we build for, but I'm guessing this could have occurred if u-boot was not skipped and if that's true then my change is not at fault here - it's either a recipe bug or we should not be allowing this chain of events to get as far as the fetcher during the parsing stage.

(FWIW the purpose of expanding PROVIDES here for a skipped recipe is to know what the recipe would have provided, so in future when we improve NoProvider error reporting we can search this info to see why the target the user is requesting to be built was unavailable.)
Comment 1 Beth Flanagan 2011-06-21 09:25:52 UTC
I'm marking this up to critical as this bug brings down a critical bit of infrastructure (one of our autobuilders that is behind a firewall and relies on git proxy, sits there and spins with 81e2f52fb8e, rendering it useless.)
Comment 2 Paul Eggleton 2011-06-21 11:17:22 UTC
Fix posted to oe-core ML:

http://thread.gmane.org/gmane.comp.handhelds.openembedded.core/2551
Comment 3 Paul Eggleton 2011-06-23 02:37:57 UTC
Fix merged in revision 3b79156be8300ebf348e6a8e14c8e7facd66a5b6.