Our DataSource came in very iteratively, when we first started using it we specifically tried not to use it for things that would require portions of the @src="" attribute to be URL encoded (so things like service names couldn't be used, but dc etc would be fine). We then gradually added an easy way to url encode the @src="" attributes with a uri helper and began to use the DataSource component more and more. This meant that some DataSource usage continued to be used without our uri helper.
Recently we hit #10901 which was a direct result of us not encoding @src values/URIs (I didn't realise this was one of the places that required URL encoding) and not going back over things to finish things off once we had implemented our uri helper, resulting in ~half of the codebase using it and ~half of it not.
Now that almost all of the UI uses our DataSource component, this PR makes it even harder to not use the uri helper, by wrapping the string that it requires in a private URI class/object, that is then expected/asserted within the DataSource component/service. This means that as a result of this PR you cannot pass a plain string to the DataSource component without seeing an error in your JS console, which in turn means you have to use the uri helper, and it's very very hard to not URL encode any dynamic/user provided values, which otherwise could lead to bugs/errors similar to the one mentioned above.
The error that you see when you don't use the uri helper is currently a 'soft' dev time only error, but like our other functionality that produces a soft error when you mistakenly pass an undefined value to a uri, at some point soon we will make these hard failing "do not do this" errors.
Both of these 'soft error' DX features have been used this to great effect to implement our Admin Partition feature and these kind of things will minimize the amount of these types of bugs moving forwards in a preventative rather than curative manner. Hopefully these are the some of the kinds of things that get added to our codebase that prevent a multitude of problems and therefore are often never noticed/appreciated.
Additionally here we moved the remaining non-uri using DataSources to use uri (that were now super easy to find), and also fixed up a place where I noticed (due to the soft errors) where we were sometimes passing undefined values to a uri call.
The work here also led me to find another couple of non-important 'bugs' that I've PRed already separately, one of which is yet to be merged (#11105), hence the currently failing tests here. I'll rebase that once that PR is in and the tests here should then pass 🤞
Lastly, I didn't go the whole hog here to make DataSink also be this strict with its uri usage, there is a tiny bit more work on DataSink as a result of recently work, so I may (or may not) make DataSink equally as strict as part of that work in a separate PR.
This PR adds a check to policy, role and namespace list pages to make sure the user has can write those things before offering to create them via a button. (The create page/form would then be a read-only form)
* ui: Don't show the CRD menu for read-only intentions
The UI bug here manifests itself only when a user/token is configured to have read-only access to intentions. Instead of only letting folks click to see a read only page of the intention, we would show an additional message saying that the intention was read-only due to it being 'Managed by [a kubernetes] CRD'. Whilst the intention was still read only, this extra message was still confusing for users.
This PR fixes up the conditional logic and further moves the logic to use ember-can - looking at the history of the files in question, this bug snuck itself in partly due to it being 'permission-y type stuff' previous to using ember-can and when something being editable or not was nothing to do with ACLs. Then we moved to start using ember-can without completely realising what IsEditable previously meant. So overall the code here is a tiny bit clearer/cleaner by adding a proper can view CRD intention instead of overloading the idea of 'editability'.
* ui: Gracefully recover from non-existent DC errors
This PR fixes what happens in the UI if you try to navigate to a non-existing DC.
When we received a 500 error from an API response due to a non-existent DC, previously we would show a 404 error, which is what we were trying to convey. But in the spirit of the UI being a 'thin client', its probably best to just show the 500 error from the API response, which may help folks to debug any issues better.
* Automatically set the CONSUL_DATACENTER_LOCAL env var for testing
* ui: Ignore response from API for KV permissions
Currently there is no way for us to use our HTTP authorization API
endpoint to tell us whether a user has access to any KVs (including the
case where a user may not have access to the root KV store, but do have
access to a sub item)
This is a little weird still as in the above case the user would click
on this link and still get a 403 for the root, and then have to manually
type in the URL for the KV they do have access to.
Despite this we think this change makes sense as at least something about KV is
visible in the main navigation.
Once we have the ability to know if any KVs are accessible, we can add
this guard back in.
We'd initially just removed the logic around the button, but then
noticed there may be further related KV issues due to the nested nature
of KVs so we finally decided on simply ignoring the responses from the
HTTP API, essentially reverting the KV area back to being a thin client.
This means when things are revisited in the backend we can undo this
easily change in one place.
* Move acceptance tests to use ACLs perms instead of KV ones
This PR supersedes #10706 and fixes#10686 whilst making sure that saving intentions continues to work.
The original fix in #10706 ignored the change action configured for the change event on the menus, meaning that the selected source/destination namespace could not be set by the user when editing/creating intentions. This, coupled with the fact that using the later intention exact endpoint for API requests endpoint means that you could not use wildcard namespaces for saving intentions.
All in all this meant that intentions could no longer be saved using the UI (whilst using ENT)
This PR reverts #10706 to fix the intention saving issue, and adds a fix for the original visual issue of nspaces doubling up in the menu once clicked. This meant repeating the existing functionality for nspaces aswell as services. It did seem strange to me that the original issue was only apparent for the nspace menus and not the service menus which should all function exactly the same way.
There is potentially more to come here partly related to what the exact functionality should be, but I'm working with other folks to figure out what the best way forwards is longer term. In the meantime this brings us back to the original functionality with the visual issue fixed.
Squashed commits:
* Revert "ui: Fix dropdown option duplications (#10706)"
This reverts commit eb5512fb74.
* ui: Ensure additional nspaces are added to the unique list of nspaces
* Add some acceptance tests
* Add Partition to all our models
* Add partitions into our serializers/fingerprinting
* Make some amends to a few adapters ready for partitions
* Amend blueprints to avoid linting error
* Update all our repositories to include partitions, also
Remove enabled/disable nspace repo and just use a nspace with
conditionals
* Ensure nspace and parition parameters always return '' no matter what
* Ensure data-sink finds the model properly
This will later be replaced by a @dataSink decorator but we are find
kicking that can down the road a little more
* Add all the new partition data layer
* Add a way to set the title of the page from inside the route
and make it accessibile via a route announcer
* Make the Consul Route the default/basic one
* Tweak nspace and partition abilities not to check the length
* Thread partition through all the components that need it
* Some ACL tweaks
* Move the entire app to use partitions
* Delete all the tests we no longer need
* Update some Unit tests to use partition
* Fix up KV title tests
* Fix up a few more acceptance tests
* Fixup and temporarily ignore some acceptance tests
* Stop using ember-cli-page-objects fillable as it doesn't seem to work
* Fix lint error
* Remove old ACL related test
* Add a tick after filling out forms
* Fix token warning modal
* Found some more places where we need a partition var
* Fixup some more acceptance tests
* Tokens still needs a repo service for CRUD
* Remove acceptance tests we no longer need
* Fixup and "FIXME ignore" a few tests
* Remove an s
* Disable blocking queries for KV to revert to previous release for now
* Fixup adapter tests to follow async/function resolving interface
* Fixup all the serializer integration tests
* Fixup service/repo integration tests
* Fixup deleting acceptance test
* Fixup some ent tests
* Make sure nspaces passes the dc through for when thats important
* ...aaaand acceptance nspaces with the extra dc param
This PR mainly adds partition to our HTTP adapter. Additionally and perhaps most importantly, we've also taken the opportunity to move our 'conditional namespaces' deeper into the app.
The reason for doing this was, we like that namespaces should be thought of as required instead of conditional, 'special' things and would like the same thinking to be applied to partitions.
Now, instead of using code throughout the app throughout the adapters to add/remove namespaces or partitions depending on whether they are enabled or not. As a UI engineer you just pretend that namespaces and partitions are always enabled, and we remove them for you deeper in the app, out of the way of you forgetting to treat these properties as a special case.
Notes:
Added a PartitionAbility while we were there (not used as yet)
Started to remove the CONSTANT variables we had just for property names. I prefer that our adapters are as readable and straightforwards as possible, it just looks like HTTP.
We'll probably remove our formatDatacenter method we use also at some point, it was mainly too make it look the same as our previous formatNspace, but now we don't have that, it instead now looks different!
We enable parsing of partition in the UIs URL, but this is feature flagged so still does nothing just yet.
All of the test changes were related to the fact that we were treating client.url as a function rather than a method, and now that we reference this in client.url (etc) it needs binding to client.
This commit fixes a problem where parent Failovers where not showing (subset children were fine).
Seems to have been introduced with a move/glimmer upgrade here #9154 so I'm adding a 1.9.x backport.
This commit fixes 2 problems with our OIDC flow in the UI, the first is straightforwards, the second is relatively more in depth:
1: A typo (1.10.1 only)
During #10503 we injected our settings service into the our oidc-provider service, there are some comments in the PR as to the whys and wherefores for this change (https://github.com/hashicorp/consul/pull/10503/files#diff-aa2ffda6d0a966ba631c079fa3a5f60a2a1bdc7eed5b3a98ee7b5b682f1cb4c3R28)
Fixing the typo so it was no longer looking for an unknown service (repository/settings > settings)
fixed this.
2: URL encoding (1.9.x, 1.10.x)
TL;DR: /oidc/authorize/provider/with/slashes/code/with/slashes/status/with/slashes should be /oidc/authorize/provider%2Fwith%2Fslashes/code%2Fwith%2Fslashes/status%2Fwith%2Fslashes
When we receive our authorization response back from the OIDC 3rd party, we POST the code and status data from that response back to consul via acallback as part of the OIDC flow. From what I remember back when this feature was originally added, the method is a POST request to avoid folks putting secret-like things into API requests/URLs/query params that are more likely to be visible to the human eye, and POSTing is expected behaviour.
Additionally, in the UI we identify all external resources using unique resource identifiers. Our OIDC flow uses these resources and their identifiers to perform the OIDC flow using a declarative state machine. If any information in these identifiers uses non-URL-safe characters then these characters require URL encoding and we added a helper a while back to specifically help us to do this once we started using this for things that required URL encoding.
The final fix here make sure that we URL encode code and status before using them with one of our unique resource identifiers, just like we do with the majority of other places where we use these identifiers.
Previously when namespaces were enabled, we weren't requesting permission for the actively selected namespace, and instead always checking the permissions for the default namespace.
This commit ensures we request permissions for the actively selected namespace.
This commit adds a bit of string wrangling to avoid the keys in our javascript source file also being transformed. Additionally, whilst looking at this we decided that Maps are a better dictionary than javascript objects, so we moved to use those here also (but this doesn't affect the issue)
Adds 'can access ACLs' which means one of two things
1. When ACLs are disabled I can access the 'please enable ACLs' page
2. When ACLs are enabled, its the same as canRead
When clicking to create a KV within folder name, would would be viewing a form that was a form for creating a KV in the root, which when the user clicked to save, saved the KV in the root.
For the moment at least I've removed the code that strips double slashes, and whilst this isn't ideal, it looks like we've picked up one of those bugs that turns into a 'feature', and completely reworking KV to not rely on the double slashes is not really an option right now.
The default namespace, and the tokens default namespace (or its origin namespace) is slightly more complicated than other things we deal with in the UI, there's plenty of info/docs on this that I've added in this PR.
Previously:
When a namespace was not specified in the URL, we used to default to the default namespace. When you logged in using a token we automatically forward you the namespace URL that your token originates from, so you are then using the namespace for your token by default. You can of course then edit the URL to remove the namespace portion, or perhaps revisit the UI at the root path with you token already set. In these latter cases we would show you information from the default namespace. So if you had no namespace segment/portion in the URL, we would assume default, perform actions against the default namespace and highlight the default namespace in the namespace selector menu. If you wanted to perform actions in your tokens origin namespace you would have to manually select it from the namespace selector menu.
This PR:
Now, when you have no namespace segment/portion in the URL, we use the token's origin namespace instead (and if you don't have a token, we then use the default namespace like it was previously)
Notes/thoughts:
I originally thought we were showing an incorrectly selected namespace in the namespace selector, but it also matched up with what we were doing with the API, so it was in fact correct. The issue was more that we weren't selecting the origin namespace of the token for the user when a namespace segment was omitted from the URL. Seeing as we automatically forward you to the tokens origin namespace when you log in, and we were correctly showing the namespace we were acting on when you had no namespace segment in the URL (in the previous case default), I'm not entirely sure how much of an issue this actually was.
This characteristic of namespace+token+namespace is a little weird and its easy to miss a subtlety or two so I tried to add some documentation in here for future me/someone else (including some in depth code comment around one of the API endpoints where this is very subtle and very hard to miss). I'm not the greatest at words, so would be great to get some edits there if it doesn't seem clear to folks.
The fact that we used to save your previous datacenter and namespace into local storage for reasons also meant the interaction here was slightly more complicated than it needed to be, so whilst we were here we rejigged things slightly to satisfy said reasons still but not use local storage (we try and grab the info from higher up). A lot of the related code here is from before we had our Routlets which I think could probably make all of this a lot less complicated, but I didn't want to do a wholesale replacement in this PR, we can save that for a separate PR on its own at some point.
This commit adds a couple of debug utilities to help us to continue slowly adding i18n support:
- We've added a CONSUL_INTL_DEBUG env/cookie variable to turn off variable interpolation within the t helper so you can see which variables are being interpolated.
- We've added a CONSUL_INTL_LOCALE env/cookie which currently supports two 'pseudo-locales' - la-fk (fake latin) and - (just dashes) either of which will make it easier to see what has not been localized until we can add prettier rules to prevent adding any copy into templates at all. I would guess if we ever translated the app we would use this for looking at things whilst developing also - but as yet I've not adding anything for that here seeing as we don't translate anything.
Both variables are dev-time only and all code for this is removed from the production build.
* ui: Standardize logo naming
According to structure it should always be logo-name not name-logo
* Make sure all our logos use logo-name format
* Upgrade to @hashicorp/structure-icons 1.9.0
* Add `-color` to be consistent with other logos
* Add ms logo back in
* Remove all the old `*-color` icons from before when we got masks
* Add missing files
* Missed glimmer extend name change
* ui: Move all our icons to use CSS custom properties
The good thing about SASS vars is, if you don't use them they get removed from the final build. Whereas with CSS we have no tree-shaking to get rid of unused CSS custom properties. We can mostly work around this and for some things like colors its no big deal if we have some hex-codes in the build that we don't use as hex-codes are relatively small.
We've been slowly but surely moving all of our colors (and other things) to use CSS custom properties instead of SASS vars now that we have them available.
This commit makes use of the 'tree-shaking' abilities of @extend to ensure that we only compile in the icons that we use.
This commit is mostly churn-less as we already use @extend for the majority of our icons, so generally there is zero change here for working on the UI, but I did spot one single place where we were using SASS vars instead of @extend. This now uses the new form (second commit)
Interestingly this reduces our CSS payload by ~2kb to ~53kb (around 25kb of that is these icons)
This commit uses docfy to isolate the individual parts and options and investigates the why you might use certain options and document how you might use certain options.
Originally we used a single %icon-definition CSS component to represent this, but seeing as some of them don't have icons, it didn't seem like the best name. So this PR splits this component into various different ones and then uses the new ones to continue to provide a now deprecated %icon-definition.
The component is currently a CSS only component that assumes a single (or multiple) description lists for its markup component, and provides for multiple different options (including a reversed mode which I'm still not totally sure about, but we don't use this right now anyway).
- %icon-definition
- %horizontal-kv-list
- %csv-list
- %tag-list
- %badge
Kenia
John Cowen
John Cowen
jkirschner-hashicorp
Jared Kirschner