PlugFest/Testing -- 17 Jun 2020

<scribe> scribenick: kaz

PRs

McCool: updates from Fujitsu

Matsukura: provided two TDs
... basically same
... direct access and via proxy
... so the IP addresses are different with each other

McCool: relative to the base address?

Matsukura: yes

Daniel: question on the IP addresses
... got a different one today from yesterday's one

McCool: something changed?

Toumura: didn't change myself

McCool: we could use the specific hostname, e.g., jv.local
... to avoid possible changes with the IP addresses

Ege: remember Panasonic had some trouble with the hostname usage

McCool: a possibility is using static addresses...
... anyway, we can't assume DHCP-based addresses are stable

(Fujitsu's PR 21 is merged)

PR 20

McCool: modified the instructions

changes

McCool: dis some more investigation as well
... about init.d setting
... crated a script for initial setting
... vpnbridge to be used as a daemon
... make sure for the VPN to assign TAP address a well
... it seems adding "sleep 10" would work, but this is not stable yet
... please make your contribution too
... please try this script and let me know
... note that my I/F is "eth1" for upstream, and "eth0" for downstream
... if you use wifi connection could be "wlan0"
... seems sometimes worky but sometimes not
... btw, Toumura-san, did you add any changes for the setting yesterday?
... e.g., rebooting the SoftEther server

Toumura: no changes added

McCool: ok
... anyway, please don't use this script asis but check the parameters before you use

(PR 20 merged)

LinkSmart updates

McCool: any news?

Farshid: setup the server to get TDs from the wot-testing repo
... assuming that's ok since the resources are public

McCool: the question is that some of the TDs are behind the local network
... also validation

Farshid: yeah
... trying to update some of them since they didn't have unique IDs

McCool: IDs are now optional for privacy purposes

<FarshidT> https://github.com/farshidtz/wot-testing-directory-sync/runs/780363992?check_suite_focus=true#step:6:35

Farshid: the above example doesn't have ID

Kaz: at least for the PlugFest purposes, we should add IDs

McCool: right
... on the other hand, what about the public purposes
... we could generate some ID automatically by some wrapper

Lagally: all the DB problem on unique ID retrieval
... the point is in this kind of structure, we need some primary key

McCool: the question is that the TD spec says ID is optional

Kaz: should think about the two purposes, 1. PlugFest this week and 2. Spec updates after that, separately

McCool: right

Ege: we could have a DB and wrapper object too
... don't think IDs are mandate for PlugFest

Farshid: IDs themselves are not necessary
... but this issue is something to be discussed in the future
... can generate an ID automatically on my side
... for validation, printing the status if any problems

<McCool> LinkSmart results

<FarshidT> populated directory: https://demo.linksmart.eu/thing-directory/td

Nodegen population

McCool: for the PoC for retail
... any ideas nodegen population from directory

LinkSmart updates - revisited

Kaz: before moving forward to the Nodegen population topic, we should record the possible issue/discussion on IDs as an outcome from the PlugFest, shouldn't we?

McCool: yeah, let me generate a GitHub issue for that purpose

<McCool> Issue 22 on IDs

Lagally: question about mechanism for LinkSmart
... query-based ID?

Farshid: ID within the TDs

Lagally: ok

McCool: 2 issues there
... difference between globally-stable ID vs temporary one

Nodegen population - revisited

McCool: someone working on that?

Farshid: we can give it a try

McCool: the purpose is retail PoC this time
... node from directory
... query against the directory service
... nodegen generates various nodes
... relatively small script to be used for the connection
... would like to volunteer myself but have long backlog queue
... if somebody could work on this, would be great
... anything else?