Bug 2787 - buildhistory.bbclass should track runtime package name changes
Summary: buildhistory.bbclass should track runtime package name changes
Status: VERIFIED FIXED
Alias: None
Product: OE-Core
Classification: Build System, Metadata & Runtime
Component: core (show other bugs)
Version: unspecified
Hardware: x86 Multiple
: Low enhancement
Target Milestone: 1.3 M4
Assignee: Paul Eggleton
QA Contact:
URL:
Whiteboard:
Depends on:
Blocks:
 
Reported: 2012-07-18 09:45 UTC by Martin Jansa
Modified: 2012-10-17 12:10 UTC (History)
5 users (show)

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

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Martin Jansa 2012-07-18 09:45:04 UTC 
buildhistory is great tool to test metadata changes to see if they have some unexpected packaging side-effects.

One of common side-effects is that new recipe version includes new binary/lib in package and without LEAD_SONAME we don't get runtime package named as it was before (when using debian.bbclass). 

Or when .so version is changed and e.g. libudev0 package is renamed to libudev1.

Such changes should be clearly documented by buildhistory.

Now there is only buildtime name (e.g. libudev) and 
packages/armv7a-vfp-neon-oe-linux-gnueabi/udev/libudev/latest doesn't show if it's called libudev0 or libudev1 in runtime (also base of .ipk filename).

If possible add
RUNTIME_NAME = "libudev0"
to latest file, or something like that
Comment 1 Martin Jansa 2012-07-20 13:08:30 UTC 
Also runtime version (PKGV) should be tracked..

e.g. systemd is using gitpkgv class for consistent git versions across multiple builders (without need to synchronize cache/bb_persist_data.sqlite3)

inherit gitpkgv
PKGV = "v${GITPKGVTAG}"
PV = "git"

And when SRCREV was upgraded to newer version (with newer GITPKGVTAG) together with PR reset it was incorrectly reporting:

ERROR: Package version for package systemd went backwards which would break package feeds from (0:git-r32 to 0:git-r5)

but from package manager POV it was normal upgrade..
Comment 2 Paul Eggleton 2012-08-06 14:51:29 UTC 
Implementation merged in Poky revision 06ad59cca92f58d4eefa42b532a778f3b83c66a1.
Comment 3 Alexandru Palalau 2012-10-17 12:10:04 UTC 
Verified on danny and master branches