[Toaster] Toaster weekly call notes - 20th May 2015

Barros Pena, Belen belen.barros.pena at intel.com
Wed May 20 10:26:30 PDT 2015


We were: Ed, Stephen, Mihail, Alex, Michael, Belen, Tiago, Levi, Dave and
David

To dos:

* Alex to document the Toaster Test Framework in the wiki, including a
link to the Django unit tests documentation
* Belén: to sync with Aníbal about the layer index work needed to identify
image recipes

Weekly status

David: working on 7154 and 7334, which are about sorting and filtering on
non-member fields. Not a trivial fix. Also reviewed the dropping support
for analysis mode design document (see opens below) and spoke to his
management about getting some extra QA resources for Toaster. He should
have more news next week

Dave: nothing to report

Belen: prepared designs for showing PROVIDES information (6169) and for
dropping support for analysis mode (7711). Collected some conclusions from
the Jenkins interviews (sent to the mailing list). Worked with Levi and
Tiago, UI patch reviews and also got in touch with the Linux Foundation to
get Toaster into their Yocto Project training materials. Currently working
on design for 5811.

Alex: improving tests (submitted patches for review). Also improving the
Toaster API: he has a demo patch that modifies the tables to return both
html and json. Waiting for managed mode refactoring work to complete in
order to finish the URL refactoring

Michael: fixing existing unit tests. Porting some more tables to the new
Toaster table widget, and fixing other small issues caused by refactoring

Mihail: most of the week on other tests, so no much progress on
automation. He'll be on holiday next week. Sent automation plan for review
to the  mailing list (see opens below)

Tiago: setting up environment for Bootstrap customisation with Levi.
Playing with Toaster prototype to finalise his initial designs for the
image customisation feature

Levi: working with Belen on UI, did some studies on tabs, and looking at
Boostrap to theme it. He sent an email to the mailing list about it, since
there are 2 different approaches for customising Bootstrap and we should
make a decision about which one to use.

Stephen: nothing to report

Ed: working on a couple of bugs. He has patches out for review for
removing full paths from variables information, picking up build tasks
when specified by users and to clean the Toaster script from bashisms. Had
a chat with Belen about image customisation and Jenkins. He won't get much
done next week since he has other tasks.

Opens: 

1. QA automation plan:

Alex has a framework in place (TTS) to assist QA as much as possible. The
framework is partially merged and it's ready to use. He has integrated as
many tests as possible in this framework, and he would like the QA and dev
teams to use it and contribute to it.

He is currently working on integrating the Selenium tests QA developed
(that they run and that outcome is collected). This maps to part of step 2
in the QA automation plan Mihail emailed to the mailing list. But step 1
(adapt the existing UI tests to work with the new managed mode interface)
should be done as soon as possible to be able to then integrate them in
TTS. Mihail thinks the data test are already working for managed mode, but
the Selenium tests only work in analysis mode.

Steps 1 and 2 are equally important: Mihail can pick which one should go
first. Just make sure to complete one before moving to the next.

Developers working on features and bugs should submit a unit test for
their work that validates the feature is done, to stop us from merging
patches that cause regressions.

The framework is ready to use, but must be documented in the wiki (AR for
Alex). The Django unit tests are well documented in Django documentation.
Let's link to that in the wiki.

Mihail asked if we are running on time for our features for M1. Alex
thinks we are a bit late, but he needs to check Bugzilla. Mihail will try
to be in sync and have test cases ready by the time features are done.

2. Removing the analysis mode:

Answers to David's questions:

* the permissions system will specify if remote 'agents' have o not
permissions to create Toaster projects
* by default projects log data to a default project id 0 to provide
backwards compatibility with the old analysis mode
* specifying a project id to log the builds is done via an environment
variable at the moment
* the API to pass data from a external builder to Toaster is already in
place
* project names expose more information than an integer, so it's probably
safer to expose the project id to enable the connection to remote builders

3. M1 cut off date is 23rd June

The things that we need to get done by then:

* Code refactoring 
* Fetching data back from AB is in place, but it would be useful to have
it for Jenkins as well
* Identifying image recipes: this is almost done. We have some code in
place for Toaster, but we need to sync with Anibal on the layer index work



On 20/05/2015 09:57, "Barros Pena, Belen" <belen.barros.pena at intel.com>
wrote:

>The weekly call for Toaster contributors happens today at 4pm GMT (8am
>PST/ 11am EST / 6pm EET). You are all welcome to join.
>
>Call Details:
>Tel: 1-888-875-9370
>Bridge: 1
>Passcode: 8235631
>
>
>Opens so far:
>
>* Discuss David's feedback on analysis and build project types. The design
>document is at
>
>https://bugzilla.yoctoproject.org/attachment.cgi?id=2508
>
>
>



More information about the toaster mailing list