Not getting a lot of takers, I think it was wise to start small just in
case, but there doesn't seem to be a floodgate problem, so let's remove
the limitations and increase the ask! (But still not a full launch yet,
because I want to funnel people through the feedback process first.)
Got the icon and background style from Neopets.com! I didn't quite copy
the whole button style, both because getting it to play nice with our
existing styles didn't *immediately* work, but also because I think
this works out as a really good compromise between our two styles
anyway!
Yay, we got the API endpoint for this! The `linkage` scope is the key.
Rather than pulling back the specific fallback behavior we had wrote
for usernames before, which was slightly different and involved
appending `neopass` in there too (e.g. `matchu-neopass-1234`), I
figured let's just use a lot of the same logic, and just use the
preferred name as the base name. (I figure the `neopass` suffix isn't
that useful anyway, `matchu-1234` kinda looks better tbh! And it's all
fallback stuff that I expect serious users to replace, anyway.)
Note: I validated this was working by temporarily changing the URI to
`https://echo.free.beeceptor.com`, which echoes the headers back, then
called `OwlsValueGuide.load_itemdata` directly.
Note: I validated this was working by temporarily changing the URI to
`https://echo.free.beeceptor.com`, which echoes the headers back, then
called `NeopetsMediaArchive.load_file_from_origin` directly.
Oh right, I never did catch this when setting up User-Agent in the app!
(I noticed this because I'm making a new request now, and went to look
how we set it in previous stuff, and was like. Oh. We don't anywhere
right now. Interesting LOL)
Oh right, if you can remove your email, there's a way to fully lock out
your account:
1. Create account via NeoPass, so no password is set.
2. Ensure you have an email saved, then disconnect NeoPass.
3. Remove the email.
4. Now you have no NeoPass, no email, and no password!
In this change, we add a validation that requires an account to always
have at least one login method. This works well for the case described
above, and also helps offer server-side validation to the "can't
disconnect NeoPass until you have an email and password" stuff that
previously was only enforced by disabling the button.
That is, the following procedure could also lock you out before,
whereas now it raises the "Whoops, there was an error disconnecting
your NeoPass from your account, sorry." message:
1. Create account via NeoPass, so no password is set.
2. Ensure you have an email saved, so "Disconnect" button is enabled.
3. Open a new browser tab, and remove the email.
4. In the original browser tab, click "Disconnect".
This is gonna help me in development, to stop having to add stuff to
the URL all the time!! I also considered just always making it
available in development, but I wanted to match production behavior to
help us ensure the hiding behavior is working, to avoid leaking NeoPass
without realizing.
Ahh okay tricky lil thing: if you show the settings page with a partial
change to `AuthUser` that didn't get saved, it can throw off the state
of some stuff. For example, if you don't have a password yet, then
enter a new password but leave the confirmation box blank, then you'll
correctly see "Password confirmation can't be blank", but you'll *also*
then be prompted for your "Current password", even though you don't
have one yet, because `@auth_user.uses_password?` is true now.
In this change, we extend the Settings form to use two copies of the
`AuthUser`. One is the copy with changes on it, and the other is the
"persisted" copy, which we check for parts of the UI that care about
what's actually saved, vs form state.
Ah okay, if you leave the password field blank but don't have one set,
our simple `update` method gets annoyed that you left it blank.
In this change, we simplify the model API by just overriding
`update_with_password` with our own special behavior for the
no-password case.
Simplified this a bit into a helper. It's kinda odd to me, but
convenient for this moment, that Rails allows views to read `params`! I
guess it's for escape hatches exactly like this! lol
including validation logic to make sure it's not already connected to
another one!
The `intent` param on the NeoPass form is part of the key! Thanks
OmniAuth for making it easy to pass that data through!
Ahh I see, if you do a no-op update, it still clears the
`previously_new_record?` state, so our NeoPass controller thinks this
account already existed. Instead, let's only do this update if it's an
account that already exists, instead of depending on the no-op-iness!
I'm getting ready to add handling for "what if you don't *have* a
current password*??", so it seems like the right way to do that is to
just eject the controller and start customizing!
That is, you're required to add a password *or* an email before
disconnecting your NeoPass, but idk, I think it's rude to demand an
email from someone for the sake of *disconnection*. Email is no longer
required for accounts that already exist!
This is more consistent with the `uses_omniauth?` we already have, and
it also will help for the next change, where I want a `uses_password?`
method (and using the name `password?` breaks some of Devise's
validation code).
Ahh right, in development `User` and `AuthUser` will have the same ID,
but that got messed up early on for us in production DTI 😅
Here, we switch the form to reference the `User` instead of the
`AuthUser` (to get the ID right), then we also change how we compare
the IDs, because `User#to_param` appends extra text onto the ID after
the number!
Motivation is that I wanna add NeoPass stuff to here! But also like,
it's looked bad for a long time, let's clean it up!! (I just used the
Devise default without any styling at all lol)
Huh, I was writing up an API inventory doc to send to TNT, and was
gonna explain why we proxy these assets… but turns out we don't need to
anymore! Nice!
This is a bit fragile if they ever change their headers, so I'll
mention that in the doc, but for now, yeah sure let's save the planet
some computational effort!
Previously, the way we loaded the image hash for a given pet was to
navigate to `https://pets.neopets.com/cpn/<pet_name>/1/1.png`, but
*not* follow the redirect, and extract the image hash from the URL
where it redirected us to.
In this change, we refactor to use the AMFPHP RPC `PetService.getPet`
instead. I don't think it had this data last time I looked at it, but
now it does! Much prefer to use an actual RPC than our weird hacky
thing!
(We might also be able to use this call for other stuff, like
auto-labeling gender & mood for pet states, maybe?? That's in this data
too! We used to load petlookups for this, long long ago, before the
petlookup captchas got added.)
I guess this was like, we had some call site that was handling loading
the viewer data itself, and didn't want to have to reload it?
But whatever, not used now, let's simplify! We can rebuild this easily
if we need it again.
Locale is the big one that's not really relevant anymore (I don't want
to be loading non-English item names anymore, now that we've simplified
to only support English like TNT has!), but there was also `item_scope`
and stuff.
The timeout option is technically not used in any call sites, but I
think that one's useful to leave around; timeout stuff is important,
and I don't want to rewrite it sometime if we need it again!
Just a small thing, I guess when I was a kid I did a weird thing where
I attached `origin_pet` to `PetType`, then upon saving `PetType` I
loaded the image hash for the pet to save as the pet type's new image
hash.
I guess this does have the nice property of not bothering to load that
stuff until we need it? But whatever, I'm moving this into `Pet` both
to simplify the relationship between the models, and to prepare for
another potential refactor: using `PetService.getPet` for this instead!
Ahh, I had assumed the `uid` provided by NeoPass would be the user's
Neopets username, but in hindsight that was never gonna work out since
NeoPass doesn't think of things in terms of usernames at all!
For now, we create 100% random NeoPass usernames, of the form
"neopass-shoyru-5812" or similar. This will be an important fallback
anyway, because it's possible to have a NeoPass with *no* Neopets.com
account attached.
But hopefully we'll be able to work with TNT to request the user's main
Neopets account's username somehow, to use that as the default when
possible!
Hacky and inconvenient, but it works!
I want this primarily to enable me to live-debug what info we're
getting back in the auth token. In production right now, the flow with
NeoPass succeeds, but we fail to create the account, and my production
error logs say it's because the username field is too long. I had hoped
it would just be the Neopets username, but now that I've poked at
NeoPass itself a bit, I'm realizing it won't be that simple.
So, we'll use this to investigate!
Ahh okay, the NeoPass spec says to pass the `client_id` and
`client_secret` as POST form data when exchanging the code for the
access token, but the default behavior of our client is to pass it as
an `Authorization` header instead.
In this change, we set an option to change that behavior, and also add
a lot of comments about this and the other options!
Wowie, it's starting to happen! :3
When you run this in production, though, you get back the auth failure
message, and the OmniAuth logs say the server returned the following:
> invalid_client: Client authentication failed (e.g., unknown client,
> no client authentication included, or unsupported authentication
> method). The OAuth 2.0 Client supports client authentication method
> 'client_secret_post', but method 'client_secret_basic' was requested.
> You must configure the OAuth 2.0 client's
> 'token_endpoint_auth_method' value to accept 'client_secret_basic'.
I'll add a fix for this in the next commit, with some explanations as
to why!
Ah beans, I didn't notice when doing my Turbo fixes in
40804c1543, that I had accidentally
prevented Chakra from applying some of its usual global styles! This
caused some minor visual regressions in various parts of the app, e.g.,
the default border color for the search field in the wardrobe app
became way darker.
Here, instead of copy-pasting the styles and missing some parts, we
scope the global styles a bit more carefully: we first use
`extendTheme` with the same code as Impress 2020 to get a good
`globalStyles` function that includes Chakra's defaults, *then* delete
the key from the theme.
Then, in `ScopedCSSReset`, we use code similar to Chakra's own global
style application code: call the `globalStyles` function with the
current theme and color mode, use Chakra's `css` function to convert
values like `green.800` to CSS values, prepend our scoping rule onto
the selectors, and drop it into our Emotion CSS.
Tbh I was worried because when I first noticed this issue while on my
trip, I saw the black item search box border, and was like "ah dang,
did I destroy all the color in the app by breaking the part where
Chakra defines its CSS color variables??" And no, that's not actually
what happened, a lot of the app still had its colors!
So this was less pressing than I had thought, but still good to get
fixed!
