I used the new profiler tools on this page, and noticed a lot of
allocations in the Globalize library, which we use for translating
database records. I realized that we were loading all of the fields of
not just all of the items on the page, but all of their translation
records in all locales! We used to scrape data for lots of languages, so
that can be quite a lot!
Unfortunately, Rails's `includes` method to efficiently preload related
records always loads all fields, and simply can't be overridden.
So, in this change we write manual preloading code, to identify the
records we need, load them in big bulk queries, and assign them back to
the appropriate associations. Basically just what `includes` does, but
written out a bit more, to give us the chance to specify SELECT and
WHERE clauses!
It shows up in development always, and if you're logged in as Me
Specifically in production!
I'm using this to poke at memory usage for pages that seem suspicious.
I don't know why our app reliably grows so large in RAM, but my hunch is
that maybe there are some pages that just use a truly large amount to
begin with - and I've learned Ruby doesn't release memory back after
it's GC'd, it just grows the process and keeps the free space to itself
in its own heap!
So I'm just eyeing pages that I know *can* have a lot going on, and
seeing what I find!
We used to do this for weird clever caching tricks that I don't think
were actually very effective. We stopped using this a few months ago,
and now I'm finally cleaning up this supporting code!
Huh, Arel can *sometimes* handle just having an attribute stand in as
"X is true" in a condition, but sometimes gets upset about it. I guess
this changed in Rails since we recently wrote this?
Specifically, item search would crash on "is:nc" (but *not* "is:np"),
saying:
```
undefined method `fetch_attribute' for #<struct Arel::Attributes::Attribute relation=#<Arel::Table:0x0000000109a67110 @name="items", @klass=Item(…), @type_caster=#<ActiveRecord::TypeCaster::Map:0x0000000109a66e90 @klass=Item(…)>, @table_alias=nil>, name="is_manually_nc">
```
The traceback was a bit misleading (it happened at the part where we
merge all the scopes together), but that hinted to me that it working
with an attribute in a place where it expected a conditional. So I
converted the attribute in the `is_nc` scope to a conditional, and made
the matching change in `is_np`, and that fixed it! Ok phew!
The URL anchors were getting like. double-encoded? The `closet[]` part
was encoding as `closet%255B%255D`. Maybe a thing in Rails, where you
need to mark them `html_safe` to insert them in a URL like that?
Well anyway, those URLs are redundant now, I just have it link straight
to the same outfit page as the big link!
Now, like in DTI 2020, opening an outfit will go straight to the editor.
I'm not 100% on whether this is actually like. the superior behavior?
But I think it's good enough, and it's what the wardrobe-2020 code
expects, so let's just roll with it for now!
Ohh I see, I made a mistake converting this from Next.js routing. It's
not that we had a URL search parameter named `outfitId`; it's that if
you were coming from the `/outfits/:outfitId` route, it would use that!
I still haven't gotten the rest of the site to point that route to this
page, but I'll do that in a later change.
Notable things:
- We used to have the parameters in the hash (`#`) part of the URL.
- We used to use the key `outfit=123` instead of `outfitId=123`.
In this change, we add backwards-compatibility for these things, while
still keeping the latest behavior too, with no change to the URLs we
generate!
Looks like the version of Prettier I just installed is v3, whereas our
last run in the impress-2020 repo was with v2. I don't think we had any
special config in that project, I think these are just changes to
Prettier's defaults, and I'm comfortable accepting them! (Mostly seems
like a lot of trailing commas.)
Idk if this used to be different or what, but it looks like the current
behavior is: if you delete a closet list, it'll leave the hangers
present, but Classic DTI would not show them anywhere; but Impress 2020
(until recently) would crash about it.
Now, we use `dependent: :destroy` to delete the hangers when you delete
the list (which I think makes sense, and is different than what I
decided in the past but that's ok, and is what the current behavior
*looks* like to people!), and we add a migration that deletes orphaned
hangers.
The migration also outputs the deleted hangers as JSON, for us to hold
onto in case we made a mistake! I'm also backing up the database in
advance of running this migration, just in case we gotta roll back HARD!
This is an important workflow for people doing art stuff, I'm told! They used to use the Classic DTI broken image UI for this, but now that that's uhh Fully Gone, let's add this more explicitly!
Ah right, the CSS reset only applies in the ScopedCSSReset container, which doesn't work for elements portaled out with the <Portal> component (which a LOT of Chakra components use for things like tooltips etc).
Here, we take advantage of <Portal> having a hardcoded classname .chakra-portal, and applying it to them too!
This was used by the Neopia server to send us the modeling data it requested out-of-band. But now we do all our modeling requests back in-app again, so we don't need this!
Okay, this is a process that idk if it's even been working for a while anyway, I don't think Neopets translates item names anymore?
And it's crashing when I try to model stuff now, so like. yeah ok I'm fine with just skipping this, it's a shame to lose out on potential data going forward but *I think there just isn't data to get anyway*
I think we used this for both conversion to image, and also for CORS stuff when rendering Flash-based previews… let's trash it, I don't want to be growing our hard drive with files I don't think we use anymore!
If I'm wrong and it turns out we do use them for something, then like. hey I'm sure we'll find out soon enough, and it's very recoverable operation.
I hope this doesn't cause problems! But yeah, with Puma doing threading, and maybe switching to Falcon someday to get even better concurrency properties, I feel like this will probably be fine?
And it makes the UX a loootttt better, to be back in the world where all these forms just work, whew.
Oh okay, I was misinterpreting the error: it was that our NEOPETS_URL_ORIGIN secret value isn't the real Neopets.com IP address anymore, so amfphp requests were just plain *always* failing in production. Oops!
I've remove that environment variable from our production config, and now modeling is working in the bulk thing!
Also I'm noticing that we're using puma these days, which does good threading stuff. I think there might be merit to switching over to Falcon because of just how async-y our stuff is, but having 5 threads going is honestly probably good enough that I don't need to worry too much about mutual blocking, and could probably just write stuff to get Neopia out of the picture like *right now*. Neat!
Okay so… I'm worried about this because of Rails whole single-threaded situation, which doesn't really let it handle blocking on external network requests very well.
Ultimately I think we're gonna have to do a clever thing but idk quite what?
I should look into whether like, puma + the new async stuff can enable Rails to be more tolerable about this, and handle a few requests at once, instead of having to have the Neopia server doing it. (Right now, the Neopia server isn't really doing its job quite right, because it depends on the Rails app being *local* to send stuff to it.)
But for now, let's just extend the timeout, cuz it's basically always getting hit in production—because there's currently no other way to do modeling, oops lol
Just find_all_by's that I never cleaned up
Oddly enough, I still got a "neopets seems down" message out of this, idk if that's an actual bug or just sluggishness rn
Okay, right, if we're just using www.neopets.com (like we are for now), it fails on http://www.neopets.com because it triggers a redirect that we don't follow.
So here I 1) change the default to HTTPS, and 2) add HTTPS support to our little RocketAMF lib
Just cleaning up a bit! I'm sure there's more to remove, these were just some clear candidates: old wardrobe code, and stuff in `public` that I just fully don't recognize and don't think is doing anything? (We'll find out if something crashes though lol!)
Oops, this was causing the page to render in a weird zoomed-out way on mobile!
Note that, for most of the site, we intentionally haven't added this tag yet because most of our pages aren't especially responsively-designed; so we _want_ the device's best attempt to work with that, rather than trying to enforce something.
This required a buncha fixes to how SASS scoping works! Needed to add a bunch of imports for stuff that previously would get read from the global scope by being imported *after* the constants and mixins etc.
There's clearly a lot of refactor opportunity here, but I'm not gonna worry about it!!
I wasn't sure what we were actually using it for, turns out it was mostly polyfills for CSS features that are very standard now!
I didn't audit these changes very carefully tbqh because they seemed pretty simple? Fingers crossed!
Eyyy tasty! There were some issues with conflicting styles with the main app, but I think we got it!
Scoping Chakra's CSS reset was a big deal to not accidentally overwrite the app's own styles lol, and we had to solve a specificity problem for that, thanks Aria for the :where tip!! <3
We never had a specific reason why we didn't use the router for this I don't think? Not that I wrote down anyway. Let's just switch it over and see what happens!
I mainly did this as a misdiagnosis of the page reload problem fixed in c162864, but it seems like a good idea to try out anyway!
This I think is why the page was reloading when you try to item search? The failed import was triggering our "hey maybe this is an old module URL that got deleted" code?
We add jsbuilding-rails to get esbuild running in the app, and then we copy-paste the files we need from impress-2020 into here!
I stopped at the point where it was building successfully, but it's not running correctly: it's not sure about `process.env` in `next`, and I think the right next step is to delete the NextJS deps altogether and use React Router instead.
Nice, just turning it on seemed to do all we need for now!
Fair questions to be asked about like, should you be able to look up by username instead of email? But like idk, this feels simpler *and* more solid, to give you feedback on if it's the right email.
In the login case, we save the `return_to` parameter in the session, because login can be a multi-step process.
In the logout case, we just read it directly from the form params.
Note that you *could* end up in a weird scenario where an old return_to value sticks around for a bit? But we have the sense to delete it when we use it on a successful sign-in, and most links to the login page come with a `return_to` param which should reset it. So, you'd have to 1) have started but not finished a sign-in, 2) during the same session, and 3) get to the login page by an unusual means.
Probably fine!
This is a bit more standard, and has the bonus of being compatible with Devise, which is using `flash[:notice]` and so its flashes were coming out unstyled, oops!
Hey nice!!
Note that I removed an account delete button from the settings page. You can still send a DELETE request to the right endpoint to do it, but it's not gonna delete all the associated records, and I wanna think a bit about how to handle that better before exposing that button.
I noticed this was stopping changing your default list visibility bc contact neopets connection can't be empty, so I fixed that!
And then I just decided to scroll through every `belongs_to` relationship and add optional to the ones that jumped out at me lol
A lot of rough edges here (e.g. no styles on the flash messages), but it's working and that's good!!
I tested this by temporarily switching to the production database and logging in as matchu!
Still missing a lot of big features too, like registration, password resets, settings page, etc.
This removes login/logout/session logic for integrating with OpenNeo ID, replacing them with stubs that just redirect to `/?TODO` when you click login, and helpers that act as if you're not logged in.
This gives us a clean slate to plug in new Devise logic to integrate with the `openneo_id` database directly!
No user-facing functionality here yet, just configuring the database connection to work with openneo_id records.
This is a first step in integrating Devise stuff into this app instead of connecting with a weird second app.
My basic testing for this was to temporarily connect to production `openneo_id`, and see `AuthUser.first` correctly return a user!
I had added this many Rails versions ago during the recent upgrade process, because it was in latest Rails but not in the version of Rails I was using when replacing Elasticsearch with MySQL queries. We can remove it now!
lmao I keep forgetting things! note that the negative case of this filter, like the negative case of `fits`, is currently broken because Rails changed the default SQL mode and I didn't notice! We'll need to add a `database.yml` file and set `sql_mode: TRADITIONAL`.
Whew! Seems like a pretty clean one? Ran `rails app:upgrade` and stuff, and made some corrections to keyword arguments for `translate` calls. There might be more such problems elsewhere? But that's hard to search for, and we'll have to see.
This one was pretty straightforward yaay! Main thing was the change from `render file` to `render template` in a couple places, oh and a thing with complex `order()` clauses.
I ran `rails zeitwerk:check`, which eager-loads the app, and it found two problems: `closet_group.rb` doesn't define `ClosetGroup` (cuz it's empty), and I left in a reference to a cache sweeper observer oops. Goodbye!
Rails 5 added new validation on `belongs_to` to ensure the corresponding record exists. In the case of moving to the null list, this shouldn't trigger!
I wish we could flag that specifically `nil` is okay, but other values should be validated? But oh well, this is fine!
Ok so weird little situation, usually Arel will accept an attribute as a param to `order()`, but not when it's in a very specific situation of all of the following:
`Item.joins(:translations).includes(:translations).limit(30).order(Item::Translation.arel_table[:name])`
For some reason, it's all like "hey I can't call `to_sql` on an attribute!", but only in the scenario where all 3 of those other things are present. Weird!
Anyway, explicitly saying `.asc` fixes this. Ok!
Some important little upgrades but mostly straightforward!
Note that there's still a known issue where item searches crash, I was hoping that this was a bug in Rails 4.2 that would be fixed on upgading to 5, but nope, oh well!
Also uhh I just got a bit silly and didn't actually mean to go all the way to 5.2 in one go, I had meant to start at 5.0… but tbh the 5.1 and 5.2 changes seem small, and this seems to be working, so. Yeah ok let's roll!
Some tricks required here to get the dependencies to work out, but we got it!!
Oh also, we move away from the rbenv in Ubuntu's package manager, because it doesn't support more recent Rubies like 2.4.10.
This labeling technique hasn't worked in a long time bc it requires being logged in. These days we just manually label them with the 2020 support tools I think!
Clearing out the Neopets gem should help us manage some gem dep conflicts in the 4.2 upgrade too (I think the nokogiri one gets tricky?)
At one point we piloted a "Camo" service to proxy HTTPS image urls for us, but it doesn't exist anymore.
We already have proxies and stuff for this, so I left `Image` as a placeholder for this, but it's not working yet!
This also deletes our final reference to the Addressable gem, so we can remove it!
I don't think these work anymore, and our volunteers get new items into the db fast anyway, Impress 2020 is doing better spidering these days. And then we get to remove the cron job `whenever` gem!
Using `s3_path` and stuff made it sound like we were still referencing the original Amazon S3 images - but actually our new asset proxy just uses the same path structure, and we didn't change anything about it.
Oh also I deleted an after_conversion method that isn't used anymore, forgot about that!
We've already swapped out the backend for this stuff to Impress 2020, so the resque task and the broken image report UI aren't actually relevant anymore. Delete them!
This helps us delete Resque soon too.
Idk this one might actually be a bit of a pain to load? But I'd want to optimize it differently anyway, and there's overhauls we're already planning to do here.
Huh! This cache key seemed to only be referenced in checks and expirations, but was never actually used! So I guess we've been loading the modeling predictions every time for a while huh??
We'll get smarter about that someday, but anyway, that lets us delete our Item resque tasks and ItemObserver!
Again I'm just not convinced of the perf on this, and it enables us to delete some whole infra over it, we can improve it another time if it's useful to!
Just removing some caching and the expiration of it! There's still more superfluous(?) caching on the item page to audit, but these seem a bit more sensible about avoiding loading extra data.
In the interest of clearing out Resque, I'm just gonna remove a lot of our more complex caching stuff, and we can do a perf pass for things like big item list pages once everything's upgraded. (I'm hopeful that the upgrades themselves improve perf; and if not, that some improved sensibilities 10 years later can find simpler approaches.)
We uninstalled Flex, our Elasticsearch gem, to replace item search with direct DB queries; but I forgot these calls, oops!
I also kinda want to see about deleting the resque tasks altogether, since I'm not sure how to get Resque installed on latest rails bc there seems to be a conflict over the version of Rack? And it'd be nice to get rid of the complexity if we can.