Store password reset requests in database #610

Fusionnée
rfwatson a fusionné 4 révision(s) à partir de feature/persist_password_reset vers master 2019-06-04 18:55:18 +00:00
rfwatson a commenté 2019-06-02 13:59:38 +00:00. (Migré de github.com)

Had a go at closing #600.

New to lots of stuff here (Rust, Diesel, Rocket, Plume...) so all feedback appreciated.

TODO

  • SQLite migrations
  • resolve inline TODOs
  • integration tests?
  • fix CI issues
Had a go at closing #600. New to lots of stuff here (Rust, Diesel, Rocket, Plume...) so all feedback appreciated. ## TODO * [x] SQLite migrations * [x] resolve inline TODOs * [ ] integration tests? * [x] fix CI issues
rfwatson (Migré de github.com) a évalué cette demande d’ajout 2019-06-02 14:01:14 +00:00
@ -243,4 +223,4 @@
form: Form<NewPasswordForm>,
rockets: PlumeRocket,
) -> Result<Flash<Redirect>, Ructe> {
form.validate()
rfwatson (Migré de github.com) a commenté 2019-06-02 14:01:14 +00:00.

The logic to detect an expired token is private to the model now. Not sure how important having a specific error is. We could maybe return a custom Diesel error?

The logic to detect an expired token is private to the model now. Not sure how important having a specific error is. We could maybe return a custom Diesel error?
rfwatson (Migré de github.com) a évalué cette demande d’ajout 2019-06-02 14:10:29 +00:00
rfwatson (Migré de github.com) a commenté 2019-06-02 14:10:29 +00:00.

It would be good to test that old password reset requests are ignored.

The simplest way I can think of would be to insert a record, and then update its creation_date with some raw SQL in the test. It feels pretty dirty though. Better ideas appreciated.

It would be good to test that old password reset requests are [ignored](https://github.com/Plume-org/Plume/pull/610/files#diff-a248869f1983a92462c39c8b2ea8a66eR44). The simplest way I can think of would be to insert a record, and then update its `creation_date` with some raw SQL in the test. It feels pretty dirty though. Better ideas appreciated.
elegaanz (Migré de github.com) a évalué cette demande d’ajout 2019-06-02 16:33:19 +00:00
elegaanz (Migré de github.com) a laissé un commentaire

What you did so far looks great! Thank you for helping with that.

What you did so far looks great! Thank you for helping with that.
elegaanz (Migré de github.com) a commenté 2019-06-02 16:27:10 +00:00.

thread::sleep(2 * 60 * 60 * 1000)

More seriously, I think you can pass tuples to diesel::insert_into(...).values(), like:

diesel::insert_into(password_reset_requests::table)
    .values((
        password_reset_requests::email.eq("foo@bar.org"),
        password_reset_requests::token.eq("aaaaaaaaa"),
        password_reset_requests::creation_date.eq(now - 3.hours()),
    ))

You can probably use it instead of PasswordResetRequest::insert in this test.

~~`thread::sleep(2 * 60 * 60 * 1000)`~~ More seriously, I think you can pass tuples to `diesel::insert_into(...).values()`, like: ```rust diesel::insert_into(password_reset_requests::table) .values(( password_reset_requests::email.eq("foo@bar.org"), password_reset_requests::token.eq("aaaaaaaaa"), password_reset_requests::creation_date.eq(now - 3.hours()), )) ``` You can probably use it instead of `PasswordResetRequest::insert` in this test.
@ -243,4 +223,4 @@
form: Form<NewPasswordForm>,
rockets: PlumeRocket,
) -> Result<Flash<Redirect>, Ructe> {
form.validate()
elegaanz (Migré de github.com) a commenté 2019-06-02 16:32:01 +00:00.

Maybe, instead of mapping the error to to_validation, you could have a closure that returns the specific error that was previously returned?

Maybe, instead of mapping the error to `to_validation`, you could have a closure that returns the specific error that was previously returned?
codecov[bot] a commenté 2019-06-04 07:59:46 +00:00. (Migré de github.com)

Codecov Report

No coverage uploaded for pull request base (master@b2312d7). Click here to learn what that means.
The diff coverage is 66.94%.

@@            Coverage Diff            @@
##             master     #610   +/-   ##
=========================================
  Coverage          ?   35.31%           
=========================================
  Files             ?       68           
  Lines             ?     7907           
  Branches          ?     1893           
=========================================
  Hits              ?     2792           
  Misses            ?     4345           
  Partials          ?      770
# [Codecov](https://codecov.io/gh/Plume-org/Plume/pull/610?src=pr&el=h1) Report > :exclamation: No coverage uploaded for pull request base (`master@b2312d7`). [Click here to learn what that means](https://docs.codecov.io/docs/error-reference#section-missing-base-commit). > The diff coverage is `66.94%`. ```diff @@ Coverage Diff @@ ## master #610 +/- ## ========================================= Coverage ? 35.31% ========================================= Files ? 68 Lines ? 7907 Branches ? 1893 ========================================= Hits ? 2792 Misses ? 4345 Partials ? 770 ```
rfwatson (Migré de github.com) a évalué cette demande d’ajout 2019-06-04 17:55:40 +00:00
@ -243,4 +223,4 @@
form: Form<NewPasswordForm>,
rockets: PlumeRocket,
) -> Result<Flash<Redirect>, Ructe> {
form.validate()
rfwatson (Migré de github.com) a commenté 2019-06-04 17:55:40 +00:00.

I ended up adding a new Expired error variant - it felt generic enough to be useful elsewhere in the app too. Happy to iterate on this more if it doesn't feel like a good approach though.

I ended up adding a new `Expired` error variant - it felt generic enough to be useful elsewhere in the app too. Happy to iterate on this more if it doesn't feel like a good approach though.
rfwatson a commenté 2019-06-04 18:06:09 +00:00. (Migré de github.com)

Pretty happy with this now. Still a couple of CI failures but they look random to me

Pretty happy with this now. Still a couple of CI failures but they look random to me
elegaanz a commenté 2019-06-04 18:13:11 +00:00. (Migré de github.com)

Yes the CI fails sometimes because it doesn't have enough memory to build…

Yes the CI fails sometimes because it doesn't have enough memory to build…
elegaanz (Migré de github.com) a approuvé ces modifications 2019-06-04 18:55:05 +00:00
elegaanz (Migré de github.com) a laissé un commentaire

Everything seems to work fine. Thank you! 😊

Everything seems to work fine. Thank you! :blush:
Connectez-vous pour rejoindre cette conversation.
Sans évaluateur
Sans jalon
Sans projet
Sans assignation
1 participant
Notifications
Échéance
La date d’échéance est invalide ou hors plage. Veuillez utiliser le format "aaaa-mm-dd".

Aucune échéance n'a été définie.

Dépendances

Aucune dépendance définie.

Référence : Plume/Plume#610
Sans contenu.