Bug 8790 - Boostrapping package information for image customisation purposes
Summary: Boostrapping package information for image customisation purposes
Status: NEW
Alias: None
Product: Toaster
Classification: Build System, Metadata & Runtime
Component: toaster (show other bugs)
Version: 2.0
Hardware: x86 Multiple
: Medium normal
Target Milestone: Future
Assignee: Toaster default assignee
QA Contact:
URL:
Whiteboard:
Depends on:
Blocks:
 
Reported: 2015-12-11 10:16 UTC by Belen Barros Pena
Modified: 2023-12-14 04:24 UTC (History)
7 users (show)

See Also:
OS type for building Yocto: ---
Type of Regression: ---
Verified:
Documentation change: No (bug/feature does not impact docs)


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Belen Barros Pena 2015-12-11 10:16:53 UTC
Our version one of image customisation functionality will show you a list of packages known to Toaster for a certain project, based on the information generated by builds. This approach has a problem: in a brand new project, the list of packages known can be very small. 

As an example, do the following: create a new project, and then build core-image-minimal. Then create a second project, and build core-image-minimal again. In the second project, create a custom image based on core-image-minimal. The list of packages will include only the 25 packages installed by core-image-minimal: nothing else. As a result, I will find no packages to add.

Ideally, we should find a way of "bootstrapping" the package data, as we do for layers and recipes. What this actually means is hazy, but it could be something like an approximate list of packages produced by the openembedded-core layer for the machines provided by openembedded-core and meta-yocto-bsp.

How we do this, I have no idea. The option discussed in the past was implementing a parsing process in Toaster, as explained on page 5 of this document 

https://wiki.yoctoproject.org/wiki/image/3/31/Image_customisation_in_Toaster.pdf

An alternative would be enhancing the parsing process in the layer index to provide package information.

Ideally, the parsing process would also provide us with package dependency information.