UDPATES:

Kasparnet:

 * Working on addressing the excellent feedback from Astea on my PSP
   API pull request (https://github.com/GPII/universal/pull/561 )
     o Shone light on an old issue that was never fixed: most *Enabled
       terms are currently ignored. Filed GPII-2685 for this work
 * Updated pull/507 - i.e. matchmaker stuff - this would be nice to get
   in before (or together with GPII-2685)
 * Helped JJ with the failing Acceptance tests (or pretending to - see
   below item)
 * Causing widespread destruction through the distribution of faulty advice
 * Struggling with getting virtual box working again on my linux box


Ale:

 * I've been involved in the new demo for Gregg.
 * Keep working on the Windows 10 mega set.


Alfredo:

 * CouchDB cluster merged in to GPII-infra repo GPII-2544
 * Update CouchDB to 2.1.1
 * Fixed Letsencrypt certificates
 * Update Jenkins
 * Tried to build master branch of Universal, but the Jenkins doesn't
   allow to filter the events reported by Github.
 * Helping Javi with the development Cloud instance


Antranig:

 * Slowed down by an ear infection this week :-(
     o -> always puts me in mind of this SYNDROME
       http://memory-alpha.wikia.com/wiki/Irumodic_Syndrome
         + That was a great episode
 * Meetings about requirements for survey popups being pitched to Astea
 * Working with ADTKINS shaking down issues with gpii-diff and its
   subsequence algorithm
     o He's reported an annoying failure in updating Kettle to current
       Infusion which appears to vector through node-jqunit, I will
       take a look
 * Helping JJ with test case structure and timeouts
 * Working with KASPARDNITT thinking about what to do about failure of
   architecture to properly handle enabling/disabling of applications
   triggered from the PSP, e.g. Windows Magnifier
 * Some distracted review of UIO+, will do more shortly
 * Can't even tell if someone has spoken or what follows what in
   alphabetical order


Brendan:

 * Handed initial Vault config stuff (Packer/Terraform/etc.) to Ops for
   review
 * Snapshot cross-region cross-account backup (with rekeying) done
   modulo one metadata race condition (?)
 * Working on Dynamo cross-region cross-account backup (without
   spending $$$$)


Cindy

 * GPII-2538: Pull request under code review -
   https://github.com/GPII/universal/pull/558
 * GPII-2511: Pull request is issued -
   https://github.com/GPII/universal/pull/563
 * GPII-2662: ready to send a pull request
 * GPII-2558: In progress.
     o Created a set() API on untrustred settings data source for
       updating preferences
     o Writing tests


*Colin*

 * Interviews with Tyler for ops candidates
 * Recovering from cold
 * Too many meetings


Javi

 * Supporting Ale and JJ with the demo
 * Continued playing around with gpii-infra - I expect to join the
   devops meeting and give a summary of my experience
 * Attended to meetings, survey pop-up, psp, etc
 * Also, continued experimenting a bit more with my chromebook


JJ:

 * Updating pull requests:
     o https://github.com/GPII/universal/pull/554: Morpher/Simplification.
     o https://github.com/GPII/windows/pull/152: Morpher/Simplification.
         + Trying to fix the last acceptance test failing to complete
           Antranig suggestions.
     o https://github.com/GPII/windows/pull/153: Brightness.
 * Involved in new demo.
 * Keep working with Ale in the new settings set  for windows 10.
 * Asking Kasper, Javi and Antranig about the grade composition in
   acceptance tests (thanks you both for your time).


a'

     o


Justin:

 * UIO+ Phase 1
     o Antranig is review GPII-2424 and I'm making changes as needed
         + https://github.com/GPII/gpii-chrome-extension/pull/2
     o Aim is to have it merged and ready for QA by Nov 27
 * UIO+ Phase 2
     o Working on FLUID-6209 (
       https://issues.fluidproject.org/browse/FLUID-6209 )
         + Needed for GPII-2559 (
           https://issues.gpii.net/browse/GPII-2559 )


Kavya

     o Demo QA
         + Issues identified  with elmer screen resolution ,simi and livia
     o Resumed work on PSP Test plan


Michelle:

 * https://issues.gpii.net/browse/GPII-2686: Creating a reusable GPII
   themed dialog for showing informational and error messages.



Sandra:
    - Interviewing candidates for DevOps and for Installation on Demand.
    - Lots of status reporting. We need to determine where we are against the plan for deployment to schools in the summer. I will get input from team and review before showing to Gregg.     - I'm out this Friday. Next week is Thanksgiving holiday in US. I'm out Wednesday through Friday, back on 27-Nov.

Simon:

 * Code review
 * Working with Sepideh on new Visible Nexus sketches
 * Working with Alan and Gregor on fluid-couch-config
     o https://github.com/fluid-project/fluid-couch-config/pull/2


Steve GRundell:
    User listeners - just about done
    Windows 10 S probe: thinking about it
    Listener API: also thinking about it

HENS:

 *      Will definitely fill this out because he is an upstanding dude.


*Tony:*

 * Working on scaling my "diff" work by implementing a proper LCS
   algorithm:
   https://en.wikipedia.org/wiki/Longest_common_subsequence_problem
   Currently working through "massive" test cases which break the
   previous implementation by running the v8 engine out of memory.
 * Submitted a follow-up PR to fix problems introduced in my
   infusion-docs "Markdown-It" pull (and some others that were not my
   fault).  Thanks to Simon and Michelle for noticing the problem.
 * Followed up on snyk vulnerabilities in various packages, working on
   pulls against Kettle and gpii-pouchdb.  Although I can get rid of
   all the high severity vulnerabilities for this pass, I wanted to
   mention that express-pouchdb moved to a monorepo some time after I
   forked their work, which will make maintaining that fork much more
   difficult.
 * In doing so, encountered a regression on the "failure to catch a
   global error" bug we were seeing.  Antranig is investigating.
 * Steve has asked about moving the i18n work in gpii-handlebars
   forward: https://issues.gpii.net/browse/GPII-2444 I plan to do this
   after:
     o I update fluid.stringTemplate to support deep paths:
       https://issues.fluidproject.org/browse/FLUID-6223
     o Before I submit a new pull, I would like to I unstick my current
       infusion pull: https://github.com/fluid-project/infusion/pull/822


Tyler:
    * Nailing down Prometheus/Grafana
    * Interviews

    Yuriy (For the Astea team):

 * Reviewed Kaspar’s pull request
   (https://github.com/GPII/universal/pull/561)
   <https://github.com/GPII/universal/pull/561%29> <- thanks
 * Some refactoring and renaming (introduced “main” and “renderer”
   source subfolders)
 * Fixed hanging of electron tests
     o An email describing the problem and the solution has been sent
         + -> This email was completely excellent
 * GPII-2607: Modify settings from PSP (integration with Kaspar’s work)
 * GPII-2678: Display snapset name in PSP menu from API (client-side
   changes only)
 * GPII-2677: Redesign and UX updates to PSP main window (WIP in a
   separate branch)
     o Awaiting final design specifications




*_AGENDA:_*


GPII-2538: Adds new requirement on auth database contents - from last week

 * "Secure /untrusted-settings endpoint"
 * https://github.com/GPII/universal/pull/558
 * GPII tokens ("keys") will not be able to retrieve settings from the
   Cloud Based Flow Manager unless they are present in the auth database
     o Rejected with an unauthorized error
 * Auth database initialized from:
   
https://github.com/cindyli/universal/blob/GPII-2538/testData/security/TestOAuth2DataStore.json
 * Are there tokens that we should add?
     o My feeling (AB) is that we should simply add tokens
       corresponding to snapsets, but I don't know how to validate this
       feeling
     o Cindy - The database probably should have adifferent data sets
       for the development and production environments
         + For the dev in this pull request, does it make sense to
           import all tokens at
           https://github.com/GPII/universal/tree/master/testData/preferences
           into the auth database so they can be used for
           /untrusted-settings endpoint
         + When the universal goes live to support pilot, we then worry
           about adding tokens corresponding to snapsets
 * When is it safe to merge?
 * Question on fluid-work regarding tokens to be added:
   https://botbot.me/freenode/fluid-work/2017-11-13/?msg=93463996&page=1
 * The only mechanism we currently have for laoding data into
   production is our preferences dataloader script:
   
https://github.com/gpii-ops/docker-preferences-dataloader/blob/master/loadPreferences.sh
 * We need a new task to create mechanisms/process for loading data
   into the production auth database
     o Two phases: initial seeding (standing the database up) and
       runtime addition of data (new user keys created)
     o Initial state
         + Views
     o Runtime addition
         + Snapset keys
         + User keys
     o The production database should not include test data (keys for
       test user preferences)
 * a follow up meeting: kasper, colin, antranig, simon, tyler, cindy,
   javi, alfredo


NPM 5.5  - The destructor of node modules (see items from last week)
Sadly, 5.x is what you need for the new 2-Factor auth feature, and AFAIK we still can't use it.
https://issues.gpii.net/browse/GPII-2673
We believe we can get around this issue by switching from having a git dependency on universal into having a regular npm dependency. This requires the pull request from Javi ( https://github.com/GPII/universal/pull/548 ) to be merged - which will be very disruptive for people using universal (i.e. everyone) ... Therefore we have decided to wait with the merge of Javi's work until we have delivered the build of PSP to usability testing sometime within the next couple of weeks. Until then we will only be supporting npm < 5


_______________________________________________
Architecture mailing list
[email protected]
https://lists.gpii.net/mailman/listinfo/architecture

Reply via email to