@jywarren it will be great if you can change https://github.com/publiclab/spectral-workbench/blob/9f9ca3cf759678e00b3283054cdf78d51d364d43/app/controllers/sessions_controller.rb#L5 and https://github.com/publiclab/mapknitter/blob/ccd6317f7752dc5c86259c40d4854cee423bfe85/app/controllers/sessions_controller.rb#L7 to localhost:3000/people at YOUR LOCAL BRANCH and test the system instead of unstable branch pushing and checking.

I am asking you to run mapknitter at localhost:3002, plots2 at localhost:3000 and swb at localhost:3004 instead of unstable branches as there is some routing constraint which routes openid request at unstable.publiclab.org to publiclab.org and we are not able to test our system at unstable branch.

SidharthBansal on 2 May 2019

Step 2 to 4 needs to be executed one after another without delay whenever you feel that it is suitable time.

SidharthBansal on 3 May 2019

I have already pushed the changes to MK and plots2. I am not able to push the changes to the SWB. As there is no unstable branch at SWB.

As written earlier, unstable.publiclab.org gets routed to publiclab.org by openid so it's better to test thing at your local @jywarren. Also, please see the dbms dump at the rails server to know what's going on at the dbs of MK, Plots and SWB.

Thanks.

SidharthBansal on 3 May 2019

OK, testing now, i hope... let's see how it goes! Thanks!

jywarren on 8 May 2019

Also, how are we on the oauth prompt UI on MapKnitter? can you link me to the latest? Thanks!

jywarren on 8 May 2019

I noticed these branches' sessions_controller.rb are pointed at https://unstable.publiclab.org - can you adjust that in SWB? I haven't checked MK but let's be sure there too.

jywarren on 8 May 2019

👍1

SWB looks good, even in not-quite-finished state:

jywarren on 8 May 2019

even in not-quite-finished state:

Frontend is broken into ftos and is implemented by folks. So, we don't need to worry much there. We have really supportive community Jeff this year. :smile:

SidharthBansal on 8 May 2019

For testing at local please follow https://github.com/publiclab/plots2/issues/5652#issuecomment-489143992. Thanks

SidharthBansal on 8 May 2019

how are we on the oauth prompt UI on MapKnitter? can you link me to the latest?

Click on login on the header
OR
go to localhost:3002/login
Both will work.
Frontend is going separately in other prs. I made frontend is separate pr as those prs does not need us to run 3 server simultaneously.

SidharthBansal on 8 May 2019

I noticed these branches' sessions_controller.rb are pointed at https://unstable.publiclab.org - can you adjust that in SWB? I haven't checked MK but let's be sure there too.

Done for both MK and SWB

SidharthBansal on 8 May 2019

OK, i'm almost installed all gems and running the local tests, i'll follow that, thank you.

I clicked Log in in SWB in the above screenshot and was directed to:

http://localhost:3001/session/new?authenticity_token=5BKc7Cnm5kaWKslIv%2BcDMsg5ciDXYRBbmF7CgYLf2G8%3D&back_to=%2F&return_to=%2F

NoMethodError in SessionsController#new
undefined method `split' for nil:NilClass
Rails.root: /home/warren/sites/spectral-workbench

Application Trace | Framework Trace | Full Trace
app/controllers/sessions_controller.rb:90:in `block in openid_authentication'
app/controllers/sessions_controller.rb:88:in `openid_authentication'
app/controllers/sessions_controller.rb:45:in `new'
Request
Parameters:

{"authenticity_token"=>"5BKc7Cnm5kaWKslIv+cDMsg5ciDXYRBbmF7CgYLf2G8=",
 "back_to"=>"/",
 "return_to"=>"/"}

jywarren on 8 May 2019

OK, looks like it's getting nil for identity_url - did I do something wrong?

   88     authenticate_with_open_id(openid_url, :required => [:nickname, :email, :fullname]) do |result, identity_url, registration|                   
   89       dummy_identity_url = identity_url

jywarren on 8 May 2019

Ensure two things. First, the plots2 server is running simultaneously at
localhost: 3000. Second, user session controller should have local
host:3000 instead of publiclab.org at MK and SWK.
Then it will work.

On Wed, May 8, 2019, 11:24 PM Jeffrey Warren notifications@github.com
wrote:

OK, i'm almost installed all gems and running the local tests, i'll follow
that, thank you.

I clicked Log in in SWB in the above screenshot and was directed to:

http://localhost:3001/session/new?authenticity_token=5BKc7Cnm5kaWKslIv%2BcDMsg5ciDXYRBbmF7CgYLf2G8%3D&back_to=%2F&return_to=%2F

NoMethodError in SessionsController#new
undefined method `split' for nil:NilClass
Rails.root: /home/warren/sites/spectral-workbench

Application Trace | Framework Trace | Full Trace
app/controllers/sessions_controller.rb:90:in block in openid_authentication' app/controllers/sessions_controller.rb:88:inopenid_authentication'
app/controllers/sessions_controller.rb:45:in `new'
Request
Parameters:

{"authenticity_token"=>"5BKc7Cnm5kaWKslIv+cDMsg5ciDXYRBbmF7CgYLf2G8=",
"back_to"=>"/",
"return_to"=>"/"}

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-490586460,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFAAEQ6MDSTJRFZBHC37XVLPUMHW7ANCNFSM4HKGL2SQ
.

SidharthBansal on 8 May 2019

Also ensure that all the three repos have my PRS. In case either of three
will not have my PR then it will not work

On Wed, May 8, 2019, 11:26 PM Jeffrey Warren notifications@github.com
wrote:

OK, looks like it's getting nil for identity_url - did I do something
wrong?

88 authenticate_with_open_id(openid_url, :required => [:nickname, :email, :fullname]) do |result, identity_url, registration|
89 dummy_identity_url = identity_url

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-490587055,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFAAEQ3CGNBRZBBS7765BVDPUMH5RANCNFSM4HKGL2SQ
.

SidharthBansal on 8 May 2019

I believe I have all those conditions true... let me double-check...

jywarren on 8 May 2019

oh haha, hadn't started local mysql. 😅

jywarren on 8 May 2019

No problem. Its like time consuming to set and operate these 3 servers
simultaneously. But happy to learn how to execute processes when
applications are running concurrently.

On Wed, May 8, 2019, 11:34 PM Jeffrey Warren notifications@github.com
wrote:

oh haha, hadn't started local mysql. 😅

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-490589973,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFAAEQ7EFI7INO7W66VXDOLPUMI3VANCNFSM4HKGL2SQ
.

SidharthBansal on 8 May 2019

❤1

In case you find any other difficulty in setting it up, then pls feel free
to ping me.
Also, I will request you to check the server logs for all the three repos
before merging those PRS.

On Wed, May 8, 2019, 11:36 PM Sidharth Bansal bansal.sidharthcode@gmail.com
wrote:

No problem. Its like time consuming to set and operate these 3 servers
simultaneously. But happy to learn how to execute processes when
applications are running concurrently.

On Wed, May 8, 2019, 11:34 PM Jeffrey Warren notifications@github.com
wrote:

oh haha, hadn't started local mysql. 😅

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-490589973,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFAAEQ7EFI7INO7W66VXDOLPUMI3VANCNFSM4HKGL2SQ
.

SidharthBansal on 8 May 2019

I mean server log for your local when the request of open id is executed
correctly.

On Wed, May 8, 2019, 11:51 PM Sidharth Bansal bansal.sidharthcode@gmail.com
wrote:

In case you find any other difficulty in setting it up, then pls feel free
to ping me.
Also, I will request you to check the server logs for all the three repos
before merging those PRS.

On Wed, May 8, 2019, 11:36 PM Sidharth Bansal <
[email protected]> wrote:

No problem. Its like time consuming to set and operate these 3 servers
simultaneously. But happy to learn how to execute processes when
applications are running concurrently.

On Wed, May 8, 2019, 11:34 PM Jeffrey Warren notifications@github.com
wrote:

oh haha, hadn't started local mysql. 😅

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-490589973,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFAAEQ7EFI7INO7W66VXDOLPUMI3VANCNFSM4HKGL2SQ
.

SidharthBansal on 8 May 2019

Great. Finally getting time to do this fully.

This is really nice:

However, the URL shows: http://localhost:3000/dashboard (which is still plots2). If I manually get the session openid_return_to from the logs, that URL works. But it's not getting inserted properly into the alert. Aha! Here's the bug:

```html

If you registered in order to use SpectralWorkbench.org or MapKnitter.org, click here to continue »

```

jywarren on 8 May 2019

Solved and pushed a commit!

jywarren on 8 May 2019

I can't test OAuth login from SWB locally, because locally i am not approved:

@SidharthBansal is there any way to test that?

jywarren on 8 May 2019

Please check looking forward email for keys and passwords. Thanks

On Thu, May 9, 2019, 3:12 AM Jeffrey Warren notifications@github.com
wrote:

I can't test OAuth login from SWB locally, because locally i am not
approved:

[image: image]
https://user-images.githubusercontent.com/24359/57410368-a0f76600-71b8-11e9-934c-14880086cb29.png

@SidharthBansal https://github.com/SidharthBansal is there any way to
test that?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-490661676,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFAAEQZZ2DQIOHHVRHO3HA3PUNCMVANCNFSM4HKGL2SQ
.

SidharthBansal on 9 May 2019

I'm trying on GitHub with the right env keys now, and it did lead me through the GitHub OAuth approval, which was great. But I got this error:

App 22225 output: Validation failed: Email has already been taken

So, we should display that, i think?

jywarren on 9 May 2019

Great, it works for unique emails! 🎉

Moving my checklist here:

[x] log in from SWB > PL with existing account
[x] log in from SWB > PL with no existing account
[x] log in from SWB > PL with GitHub account
[x] log in from MK > PL with existing account
[x] log in from MK > PL with no existing account
[x] log in from MK > PL with GitHub account
[ ] need to display validation errors

jywarren on 9 May 2019

Great! All tested and good! I just want to check about showing an alert for the validation error, but otherwise this looks great.

jywarren on 9 May 2019

Tremendous job on getting it all right, Sidharth!

jywarren on 9 May 2019

It's SUCH a smooth process compared to before!!!

jywarren on 9 May 2019

So, we need to wait then until MapKnitter's new version publishes to production before actually doing this deployment. We haven't yet actually published MapKnitter for quite a while. This will also give us time to finish up the UI as well. Does this sound good then?

jywarren on 9 May 2019

Great

On Thu, May 9, 2019, 5:05 AM Jeffrey Warren notifications@github.com
wrote:

So, we need to wait then until MapKnitter's new version publishes to
production before actually doing this deployment. We haven't yet actually
published MapKnitter for quite a while. This will also give us time to
finish up the UI as well. Does this sound good then?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-490689052,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFAAEQ6PIHXHQ2YYKGNISVLPUNPS3ANCNFSM4HKGL2SQ
.

SidharthBansal on 9 May 2019

Actually these three PRS needs to be merged and deployed to production
simultaneously. Other work can be done after these PRS get merged. They can
be merged and deployed at any time. So,
I think UI should be done in separate PRS. What do you think?

On Thu, May 9, 2019, 5:21 AM Sidharth Bansal bansal.sidharthcode@gmail.com
wrote:

Great

On Thu, May 9, 2019, 5:05 AM Jeffrey Warren notifications@github.com
wrote:

So, we need to wait then until MapKnitter's new version publishes to
production before actually doing this deployment. We haven't yet actually
published MapKnitter for quite a while. This will also give us time to
finish up the UI as well. Does this sound good then?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-490689052,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFAAEQ6PIHXHQ2YYKGNISVLPUNPS3ANCNFSM4HKGL2SQ
.

SidharthBansal on 9 May 2019

Great, noting that since https://github.com/publiclab/spectral-workbench/pull/455 is merged, we shouldn't publish Spectral Workbench until this checklist is completed.

jywarren on 9 May 2019

Yeah, I mean to say this. We will deploy changes to the SWB when all 3 prs are merged. Before that we should not set them to production.

Sorry, I used wrong word. I should use deploy instead of merge. Apologies. You did correctly.
We are doing things correctly in correct order.

SidharthBansal on 9 May 2019

Just wanna know one thing, can we merge all the 3 prs and deploy them later simultaneously when the time comes (after bootstrap prs are merged)?
We are not deploying things so this will not break any consistency. What do you suggest Jeff?

SidharthBansal on 9 May 2019

Hmm, maybe! But how long can we wait, maximum?

On Thu, May 9, 2019, 2:22 PM Sidharth Bansal notifications@github.com
wrote:

Just wanna know one thing, can we merge all the 3 prs and deploy them
later simultaneously when the time comes (after bootstrap prs are merged)?
We are not deploying things so this will not break any consistency. What
do you suggest Jeff?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-491012946,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AAAF6JYDP3N3JSUL6UJAIMDPURTTJANCNFSM4HKGL2SQ
.

jywarren on 9 May 2019

I mean like worst case scenario :-)

On Thu, May 9, 2019, 2:41 PM Jeffrey Warren jeff@unterbahn.com wrote:

Hmm, maybe! But how long can we wait, maximum?

On Thu, May 9, 2019, 2:22 PM Sidharth Bansal notifications@github.com
wrote:

Just wanna know one thing, can we merge all the 3 prs and deploy them
later simultaneously when the time comes (after bootstrap prs are merged)?
We are not deploying things so this will not break any consistency. What
do you suggest Jeff?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-491012946,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AAAF6JYDP3N3JSUL6UJAIMDPURTTJANCNFSM4HKGL2SQ
.

jywarren on 9 May 2019

We can wait as much as we want. We just need these 3 prs to be merged at any time before deployment(plots, mk, & swb). Does this makes sense?

SidharthBansal on 9 May 2019

yep, thanks! Just trying to coordinate all the scheduling and proper
checking :-)

On Thu, May 9, 2019 at 2:50 PM Sidharth Bansal notifications@github.com
wrote:

We can wait as much as we want. We just need these 3 prs to be merged at
any time before deployment(plots, mk, & swb). Does this makes sense?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-491022638,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AAAF6J46CKLIUSFMWK3SGYDPURW4JANCNFSM4HKGL2SQ
.

jywarren on 9 May 2019

No hurry!!! Take your time Jeff.
Reviewing some PRs and issues. Trying to coordinate and make groups for people in the meantime.
I hope this will help you a little.
Thanks

SidharthBansal on 9 May 2019

👍1

thank you!

On Thu, May 9, 2019 at 3:04 PM Sidharth Bansal notifications@github.com
wrote:

No hurry!!! Take your time Jeff.
Reviewing some PRs and issues. Trying to coordinate and make groups for
people in the meantime.
I hope this will help you a little.
Thanks

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-491027553,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AAAF6J65GZFNJCAWNART3K3PURYTRANCNFSM4HKGL2SQ
.

jywarren on 9 May 2019

OK, we are going to try to do this tomorrow.

jywarren on 9 May 2019

Thanks @jywarren for merging all the three PRs.

SidharthBansal on 10 May 2019

OK, we weren't able to complete this; a range of issues held us back on MapKnitter; we fixed a number of bugs and configuration issues, but ultimately ran out of time for today. We've scheduled to try again on Monday at 4pm ET (8pm UTC).

We got the point where we are testing that MapKnitter exports work on this map: http://mapknitter-stable.laboratoriopublico.org/maps/test--226

We completed upgrading the slider UI, the Dockerfile, export folder creation, and we got exports to run, although we didn't confirm that they completed.

I believe the spot we're stuck at is getting MapKnitter to boot properly using Passenger in the container, and getting this page to load:

http://mapknitter-stable.laboratoriopublico.org/maps/test--226

Once we resolve that, we can move forward with:

[x] testing that exports generate properly
[x] deploying MapKnitter, SWB, and PublicLab.org together
[x] testing all three for basic login
[ ] testing all three for OAuth login
[x] testing MapKnitter exporting in production

Then we'll be good! That's it for today, folks, thanks for your patience and we'll give it another go on Monday afternoon.

jywarren on 11 May 2019

Great
Best of luck

On Sat, May 11, 2019, 4:44 AM Jeffrey Warren notifications@github.com
wrote:

OK, we weren't able to complete this; a range of issues held us back on
MapKnitter; we fixed a number of bugs and configuration issues, but
ultimately ran out of time for today. We've scheduled to try again on
Monday at 4pm ET (8pm UTC).

We got the point where we are testing that MapKnitter exports work on this
map: http://mapknitter-stable.laboratoriopublico.org/maps/test--226

We completed upgrading the slider UI, the Dockerfile, export folder
creation, and we got exports to run, although we didn't confirm that they
completed.

I believe the spot we're stuck at is getting MapKnitter to boot properly
using Passenger in the container, and getting this page to load:

http://mapknitter-stable.laboratoriopublico.org/maps/test--226

Once we resolve that, we can move forward with:

testing that exports generate properly

deploying MapKnitter, SWB, and PublicLab.org together

testing all three for basic login

testing all three for OAuth login

testing MapKnitter exporting in production

Then we'll be good! That's it for today, folks, thanks for your patience
and we'll give it another go on Monday afternoon.

—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub
https://github.com/publiclab/plots2/issues/5652#issuecomment-491454845,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFAAEQ5IMJORE7TTICRPM4LPUX6WNANCNFSM4HKGL2SQ
.

SidharthBansal on 11 May 2019

Deployment is complete! All live now.

I did find one odd thing. When logging in via GitHub OAuth from MapKnitter, i keep getting directed to Spectral Workbench's dashboard. The sequence is this:

go to MK.org and press github oauth button
log into github
directed to (logged in) PL.org/dashboard
went to SWB, logged in with github button
got back to SWB dash
tried to open MK.org but wasn't logged in there. Pressed GitHub again
directed to SWB dashboard
repeated 6 and 7 with same result

This was a complex path so I'll try again with a simpler sequence.

jywarren on 15 May 2019

Repeated just using GH OAuth to log in from MapKnitter. I got sent to PL.org/dashboard.

Good news is that Log in with PL.org is working for both MK and SWB. That's perfect. Looks like there are some oddities with Oauth login however.

jywarren on 15 May 2019

The Github OAuth redirect from MK took a while and I was able to copy this URL while it was loading (after the logging into GH step):

https://github.com/login/oauth/authorize?client_id=0ea2dd560523729ee8d8&redirect_uri=https%3A%2F%2Fpubliclab.org%2Fauth%2Fgithub%2Fcallback&response_type=code&scope=user%3Aemail&state=de44729c9fc96127586a05a33950d825d11e29a3a49a7d29

I just did the same with SWB (pressed GH OAuth login button from SWB) and also got directed to PL.org/dashboard. The URL while being redirected was:

https://github.com/login/oauth/authorize?client_id=0ea2dd560523729ee8d8&redirect_uri=https%3A%2F%2Fpubliclab.org%2Fauth%2Fgithub%2Fcallback&response_type=code&scope=user%3Aemail&state=f4351d908bb2607b3c466e74b5dacb6299a30ed8953eeb2b

jywarren on 15 May 2019

On Twitter OAuth from SWB, I got to this page and got stuck in an infinite refresh loop:

jywarren on 15 May 2019

Same for MK using Twitter OAuth:

I don't have a FB account so didn't test that.

jywarren on 15 May 2019

Using Google Oauth login from SWB got me to PL.org/dashboard. Interestingly it redirected through this URL: https://accounts.youtube.com/accounts/SetSID.

Google OAuth from MK.org also got me logged into PL.org and left at the dashboard.

jywarren on 15 May 2019

OK, so to sum up:

both OpenID logins to PL.org work fine!
Google Oauth logs into PL.org but leaves you at the dashboard
GitHub OAuth logs into PL.org but leaves you at the dashboard
Twitter OAuth gets into an infinite refresh loop

This is pretty good, @SidharthBansal ! It looks like just some minor things to work out, but overall I'm really happy with how this went!

jywarren on 15 May 2019

Oh, and also clicking PL button at SWB and /creating/ a new account worked fine too! Same with MK.

jywarren on 15 May 2019

I also tried the same. It is really weird behaviour.
I will check whats going on after my NLP exam today.
Apologies for the errors. Will fix it in the evening.

SidharthBansal on 15 May 2019

Please check the fields which are completed.
Thanks

SidharthBansal on 15 May 2019

OK, collecting all leftover checklist items:

[x] MK production recaptcha key
[x] testing all three for OAuth login (GH/Google for leaving you at the dashboard)
[x] figure out Twitter OAuth refresh loop
[x] need to display validation errors in MK/SWB (https://github.com/publiclab/plots2/issues/5734)

Thanks, and good luck with your exam!!!!

jywarren on 15 May 2019

@jywarren the following are also working fine.

[x] User is logged in at PL and tries to login at MK through Twitter(Twitter is logged in)
[x] User is logged in at PL and tries to login at MK through Facebook(Facebook is logged in)
[x] User is logged in at PL and tries to login at MK through Github(Github is logged in)
[x] User is logged in at PL and tries to login at MK through Google(Google is logged in)

[x] User is logged in at PL and tries to login at MK through Login Via PL
[x] User is logged in at PL and tries to login at SWB through Twitter(Twitter is logged in)
[x] User is logged in at PL and tries to login at SWB through Google(Google is logged in)
[x] User is logged in at PL and tries to login at SWB through Facebook(Facebook is logged in)
[x] User is logged in at PL and tries to login at SWB through Github(Github is logged in)
[x] User is logged in at PL and tries to login at MK through Twitter(Twitter is not logged in)
[x] User is logged in at PL and tries to login at MK through Facebook(Facebook is not logged in)
[x] User is logged in at PL and tries to login at MK through Github(Github is not logged in)
[x] User is logged in at PL and tries to login at MK through Google(Google is not logged in)
[x] User is logged in at PL and tries to login at MK through Login Via PL
[x] User is logged in at PL and tries to login at SWB through Twitter(Twitter is not logged in)
[x] User is logged in at PL and tries to login at SWB through Google(Google is not logged in)
[x] User is logged in at PL and tries to login at SWB through Facebook(Facebook is not logged in)
[x] User is logged in at PL and tries to login at SWB through Github(Github is not logged in)

Other flows for MK

[x] User is not logged in at PL and tries to use Login Via PL. Uses Login form at PL for logging in
[x] User is not logged in at PL and tries to use Login Via PL. Uses Signup form at PL for logging in
[x] User is not logged in at PL and tries to use Login Via PL. Uses Google Provider at PL for logging in.
Google Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use Login Via PL. Uses Google Provider at PL for logging in.
Google Provider is not linked to any account. Check for email linked with Google Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use Login Via PL. Uses Google Provider at PL for logging in.
Google Provider is not linked to any account. Check for email linked with Google Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email. Login MK user.
[x] User is not logged in at PL and tries to use Login Via PL. Uses Twitter Provider at PL for logging in.
Twitter Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use Login Via PL. Uses Twitter Provider at PL for logging in.
Twitter Provider is not linked to any account. Check for email linked with Twitter Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use Login Via PL. Uses Twitter Provider at PL for logging in.
Twitter Provider is not linked to any account. Check for email linked with Twitter Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email. Login MK user.
[x] User is not logged in at PL and tries to use Login Via PL. Uses GH Provider at PL for logging in.
GH Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use Login Via PL. Uses GH Provider at PL for logging in.
GH Provider is not linked to any account. Check for email linked with GH Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use Login Via PL. Uses GH Provider at PL for logging in.
GH Provider is not linked to any account. Check for email linked with GH Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email. Login MK user.
[x] User is not logged in at PL and tries to use Login Via PL. Uses FB Provider at PL for logging in.
FB Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use Login Via PL. Uses FB Provider at PL for logging in.
FB Provider is not linked to any account. Check for email linked with FB Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use Login Via PL. Uses FB Provider at PL for logging in.
FB Provider is not linked to any account. Check for email linked with FB Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email. Login MK user.

Other flows for SWB

[x] User is not logged in at PL and tries to use Login Via PL. Uses Login form at PL for logging in
[x] User is not logged in at PL and tries to use Login Via PL. Uses Signup form at PL for logging in
[x] User is not logged in at PL and tries to use Login Via PL. Uses Google Provider at PL for logging in.
Google Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use Login Via PL. Uses Google Provider at PL for logging in.
Google Provider is not linked to any account. Check for email linked with Google Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use Login Via PL. Uses Google Provider at PL for logging in.
Google Provider is not linked to any account. Check for email linked with Google Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email.
[x] User is not logged in at PL and tries to use Login Via PL. Uses Twitter Provider at PL for logging in.
Twitter Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use Login Via PL. Uses Twitter Provider at PL for logging in.
Twitter Provider is not linked to any account. Check for email linked with Twitter Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use Login Via PL. Uses Twitter Provider at PL for logging in.
Twitter Provider is not linked to any account. Check for email linked with Twitter Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email.
[x] User is not logged in at PL and tries to use Login Via PL. Uses GH Provider at PL for logging in.
GH Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use Login Via PL. Uses GH Provider at PL for logging in.
GH Provider is not linked to any account. Check for email linked with GH Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use Login Via PL. Uses GH Provider at PL for logging in.
GH Provider is not linked to any account. Check for email linked with GH Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email.
[x] User is not logged in at PL and tries to use Login Via PL. Uses FB Provider at PL for logging in.
FB Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use Login Via PL. Uses FB Provider at PL for logging in.
FB Provider is not linked to any account. Check for email linked with FB Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use Login Via PL. Uses FB Provider at PL for logging in.
FB Provider is not linked to any account. Check for email linked with FB Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email.

As per https://github.com/publiclab/plots2/issues/5652#issuecomment-492480385, shall I assume that all the above flows are tested by you?
I hope these all are working properly.
I have ticked those which I checked are working correctly.

SidharthBansal on 16 May 2019

Other flows to be tested OR giving errors for MK

Consider PL is not logged in for sure. User will use provider buttons at MK directly

[x] User is not logged in at PL and tries to use Google Provider
Google Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use Google Provider
Google Provider is not linked to any account. Check for email linked with Google Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use Google Provider
Google Provider is not linked to any account. Check for email linked with Google Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email.
[x] User is not logged in at PL and tries to use Twitter Provider
Twitter Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use Twitter Provider.
Twitter Provider is not linked to any account. Check for email linked with Twitter Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use Twitter Provider
Twitter Provider is not linked to any account. Check for email linked with Twitter Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email.
[x] User is not logged in at PL and tries to use GH Provider
GH Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use GH Provider
GH Provider is not linked to any account. Check for email linked with GH Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use GH Provider
GH Provider is not linked to any account. Check for email linked with GH Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email.
[x] User is not logged in at PL and tries to use FB Provider
FB Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use FB Provider
FB Provider is not linked to any account. Check for email linked with FB Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use FB Provider
FB Provider is not linked to any account. Check for email linked with FB Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email.

I am considering whether or not provider(GH,FB etc) is logged in can be taken as single test case. Their sign up/login should depend on their architecture so we can't play with that.

SidharthBansal on 16 May 2019

Other flows to be tested OR giving errors for SWB

Consider PL is not logged in for sure. User will use provider buttons at SWB directly

[x] User is not logged in at PL and tries to use Google Provider
Google Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use Google Provider
Google Provider is not linked to any account. Check for email linked with Google Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use Google Provider
Google Provider is not linked to any account. Check for email linked with Google Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email.
[x] User is not logged in at PL and tries to use Twitter Provider
Twitter Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use Twitter Provider.
Twitter Provider is not linked to any account. Check for email linked with Twitter Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use Twitter Provider
Twitter Provider is not linked to any account. Check for email linked with Twitter Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email.
[x] User is not logged in at PL and tries to use GH Provider
GH Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use GH Provider
GH Provider is not linked to any account. Check for email linked with GH Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use GH Provider
GH Provider is not linked to any account. Check for email linked with GH Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email.
[x] User is not logged in at PL and tries to use FB Provider
FB Provider already linked to some account. Then that account is logged in at PL and MK
[x] User is not logged in at PL and tries to use FB Provider
FB Provider is not linked to any account. Check for email linked with FB Provider. Link with the PL account having same email if email exists in the PL database
[x] User is not logged in at PL and tries to use FB Provider
FB Provider is not linked to any account. Check for email linked with FB Provider. Email doesn't exist in the PL database then create an account with that email and link provider with that email.

I am considering whether or not provider(GH,FB etc) is logged in can be taken as single test case. Their sign up/login should depend on their architecture so we can't play with that.

SidharthBansal on 16 May 2019

In above 70+ test cases, all should land to the MK or SWB.
They all are similar but little different. So, we need to ensure all of them.

SidharthBansal on 16 May 2019

On clicking on the icons on the SWB the following error comes SOMETIMES

I really don't know why it is SOMETIMES.
I have invested 7 hours trying to debug it but in vain.

SidharthBansal on 16 May 2019

@jywarren please see https://github.com/publiclab/plots2/pull/5640. I don't know why there are 45 files .

SidharthBansal on 16 May 2019

I checked there is no overlapping area in that PR with the changes you introduce. Trying to get what happened and where.

SidharthBansal on 16 May 2019

When we will deploy the code of PL next time? I think my little PR will work out many test cases.

SidharthBansal on 16 May 2019

@jywarren found this error in SWB

google ask me to refer
a2bd770f6d3b402dfb70ae24460014bf44c41abe

Trying to debug it. It has been 8-9 hours today. Very tough to know what has happened. But I hope we will find what's behind the scenes.
One strange thing

but

please explain this to me. I can't understand many things are going on simultaneously. Sorry for the inconvinience.

SidharthBansal on 16 May 2019

Oh goodness, first - thank you! I think given that people can log in with
PL.org and the other methods are new, we are no longer at extreme urgency -
one thing we could do is to hide the OAuth logins that /don't/ work, for
now. So, you can focus on your exams too.

Second, the validation - it's a small thing, not high priority. I think it
was that when logging in with OAuth and when it uses your email from your
OAuth provider, if the email already exists as an account in MK/SWB, it
doesn't validate because the email is not unique. But we don't display any
validation errors when redirecting back to MK. Honestly i don't really know
how this would be the case unless someone somehow managed to have 2 PL
accounts, but switched their PL email or something? I think it's small
enough that we could probably ignore it, i just thought displaying
validation messages in flash might be nice.

Not sure what sequence brought you to the SWB error but indeed it also
looks like a validation error due to an already-used email address. I think
this is unlikely to happen in production since most people have only one
account.

Hmm. On session type missing, i really don't know!

jywarren on 16 May 2019

Please review it tomorrow or next week in your free time. No hurry.

hide the OAuth logins that /don't/ work,

I have merged #5731 in plots2. I think we will not require to hide the icons. Major issues will be resolved by this PR, I guess. Localhost is working for major portions now.