Bug 4250 - buildhistory reports bad/incomplete values for meta-toolchain-sdk
Summary: buildhistory reports bad/incomplete values for meta-toolchain-sdk
Status: VERIFIED NOTABUG
Alias: None
Product: OE-Core
Classification: Build System, Metadata & Runtime
Component: configuration (show other bugs)
Version: 1.4
Hardware: x86 Multiple
: Undecided normal
Target Milestone: ---
Assignee: Paul Eggleton
QA Contact:
URL:
Whiteboard:
Depends on:
Blocks:
 
Reported: 2013-04-11 12:46 UTC by Stoicescu Cornel
Modified: 2013-04-12 07:42 UTC (History)
0 users

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 Stoicescu Cornel 2013-04-11 12:46:46 UTC
/build/tmp/buildhistory/sdk/poky-eglibc-*/- sdk-info.txt has osme issues:
     * SDKMACHINE has no value if it is not specifically set in local.conf(it should add the default value if it is not set in local.conf)
     * SDKSIZE has no measurement unit set (ex: SDKSIZE = 587876)
     * SDKIMAGE_FEATURES has no value
     * BAD_RECOMMENDATIONS has no value (maybe this is empty because there are no BAD_RECOMMENDATIONS set?)
Comment 1 Paul Eggleton 2013-04-11 14:04:57 UTC
Dealing with these issues in turn:

* SDKMACHINE gets auto-detected by the build system if not set, but the system does not set the variable's value after doing that. Perhaps it should, but this is not an issue in buildhistory.
* SDKSIZE doesn't have a unit I agree, but nor do any of the sizes reported by buildhistory i.e. the ones for packages and images. Perhaps they should, but we'd also have to remove this when reading the number back in.
* SDKIMAGE_FEATURES isn't currently used by SDKs except for when you're doing "bitbake -c populate_sdk imagename", so if you're looking for the value when building meta-toolchain for example then the value will probably be blank. Once bug 3441 is implemented this limitation should no longer be present, however until then what buildhistory is reporting is correct here.
* As for BAD_RECOMMENDATIONS, the default value is blank so it's correct; to be honest though it's only implemented for ipk and even then it is only implemented for images so its utility at the moment in this file is questionable.
Comment 2 Paul Eggleton 2013-04-11 16:37:12 UTC
Given the above, I'm marking this as NOTABUG. If we want to make general enhancements to buildhistory output we should definitely do that, but this is specific to the SDK.
Comment 3 Stoicescu Cornel 2013-04-12 07:42:44 UTC
This bug was filed after verifying bug 3964 (enhancement) and the above issues ware pointed out(this is why it's a buildhistory bug specific to SDK).

As per Paul's comments I am setting this to verified - not a bug.