* ci: npm apparently needs a tag flag now for some reason...? like what the hell is a dist-tag even
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* revert frontend styling
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* fix autosubmit duplicate label
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
---------
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* unrelated: dont show nested for user
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* unrelated: fix error when no extents in. map
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* events: use pending_user when possible
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* fix for identification stage "fake" user
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* better username rendering
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
---------
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* web: Add InvalidationFlow to Radius Provider dialogues
## What
- Bugfix: adds the InvalidationFlow to the Radius Provider dialogues
- Repairs: `{"invalidation_flow":["This field is required."]}` message, which was *not* propagated
to the Notification.
- Nitpick: Pretties `?foo=${true}` expressions: `s/\?([^=]+)=\$\{true\}/\1/`
## Note
Yes, I know I'm going to have to do more magic when we harmonize the forms, and no, I didn't add the
Property Mappings to the wizard, and yes, I know I'm going to have pain with the *new* version of
the wizard. But this is a serious bug; you can't make Radius servers with *either* of the current
dialogues at the moment.
* This (temporary) change is needed to prevent the unit tests from failing.
\# What
\# Why
\# How
\# Designs
\# Test Steps
\# Other Notes
* Revert "This (temporary) change is needed to prevent the unit tests from failing."
This reverts commit dddde09be5.
* web/element: empty-state should not have a default label when used as a loading indicator
* .
* web/bug/empty-state: Fix issues with EmptyState and Loading Overlay
- Add a method, `hasSlotted()`, to the Base component.
- Revise `EmptyState` to use `hasSlotted()`.
- Revise `LoadingOverlay` to use `hasSlotted()`.
- Provide (hopefully complete) Storybook stories for both
- Revise use of these components throughout the codebase.
The essential problem here was mine: I misunderstood what the Patternfly `SlotController` does (and,
yikes, how it does it). Slots aren't magical; they're just named containers, in which lightDOM
elements that appear between the opening and closing tags of a web component can be strategically
placed, shown or hidden, and to some extent styled, within the rendered and visible results of the
shadowDOM component that will fill the browser's RECT allocated to that component.
SlotController tries to associate the template with slots by creating the shadowDOM *first*, then
working backwards to see if there are lightDOM components to put into those slots. That's not what
we want; we want to see if there are lightDOM components that meet our slot requirements and, if
there are, create corresponding slots for them.
That's what `hasSlotted()` does: it returns true or false to the question, "Is there currently in
the lightDOM for this component an entry requesting a known slot name?" Components are free to do
what they want with that knowledge.
`<ak-empty-state>` now has several modes, all well-documented in the Storybook story. But in short,
the Title is now a default slot; any HTML Element not sent to one of the named slots are put into
the Title. The two named slots are `body` and `primary`. The header is bold and large; body is
just text, and primary is boxed to indicate that one or more buttons should be placed there, to
allow interaction.
The extra modes are controlled by boolean attributes:
- `loading`: Shows the loading spinner, overriding the `icon` attribute
- `default`: Shows the loading spinner *and* the word "Loading" (i18n-aware).
The priority for all of these is:
- Has something in the default (header) slot: That text will be shown. Overrides both
- `default` overrides `loading`
- `loading`
q`<ak-loading-overlay>` is a specialized variant of `<ak-empty-state>` over what will become
`<ak-backdrop>`, but for now is just internal. It allows only for the heading and primary slots,
forwarding them `<ak-empty-state>`. Since this is literally the *Loading*Overlay, showing the
`loading` spinner is the default; to prevent it, pass `no-spinner` as an attribute.
* Grammatical error.
* Prettier had opinions that shouldn't have been aired in public.
* Prettier had opinions that shouldn't have been aired in public.
* Collapsing unnecessary boolean nest.
* fix typo
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* always render icon
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* missing default in flow exec
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* unrelated: fix loading interface
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* rename default attr
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* fix jsdoc
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
---------
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
Co-authored-by: Jens Langhammer <jens@goauthentik.io>
* web: fix esbuild issue with style sheets
Getting ESBuild, Lit, and Storybook to all agree on how to read and parse stylesheets is a serious
pain. This fix better identifies the value types (instances) being passed from various sources in
the repo to the three *different* kinds of style processors we're using (the native one, the
polyfill one, and whatever the heck Storybook does internally).
Falling back to using older CSS instantiating techniques one era at a time seems to do the trick.
It's ugly, but in the face of the aggressive styling we use to avoid Flashes of Unstyled Content
(FLoUC), it's the logic with which we're left.
In standard mode, the following warning appears on the console when running a Flow:
```
Autofocus processing was blocked because a document already has a focused element.
```
In compatibility mode, the following **error** appears on the console when running a Flow:
```
crawler-inject.js:1106 Uncaught TypeError: Failed to execute 'observe' on 'MutationObserver': parameter 1 is not of type 'Node'.
at initDomMutationObservers (crawler-inject.js:1106:18)
at crawler-inject.js:1114:24
at Array.forEach (<anonymous>)
at initDomMutationObservers (crawler-inject.js:1114:10)
at crawler-inject.js:1549:1
initDomMutationObservers @ crawler-inject.js:1106
(anonymous) @ crawler-inject.js:1114
initDomMutationObservers @ crawler-inject.js:1114
(anonymous) @ crawler-inject.js:1549
```
Despite this error, nothing seems to be broken and flows work as anticipated.
* initial OSM for events
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* remove card title
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* split with volume
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* add pin
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* basic map selection
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* update pin
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* rewrite map points to be more imperative
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* zoom to fit
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
---------
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
Co-authored-by: Ken Sternberg <ken@goauthentik.io>
* web: Add InvalidationFlow to Radius Provider dialogues
## What
- Bugfix: adds the InvalidationFlow to the Radius Provider dialogues
- Repairs: `{"invalidation_flow":["This field is required."]}` message, which was *not* propagated
to the Notification.
- Nitpick: Pretties `?foo=${true}` expressions: `s/\?([^=]+)=\$\{true\}/\1/`
## Note
Yes, I know I'm going to have to do more magic when we harmonize the forms, and no, I didn't add the
Property Mappings to the wizard, and yes, I know I'm going to have pain with the *new* version of
the wizard. But this is a serious bug; you can't make Radius servers with *either* of the current
dialogues at the moment.
* This (temporary) change is needed to prevent the unit tests from failing.
\# What
\# Why
\# How
\# Designs
\# Test Steps
\# Other Notes
* Revert "This (temporary) change is needed to prevent the unit tests from failing."
This reverts commit dddde09be5.
* web/components: Remove all special cases of slug handling, replace with a "smart slug" component
This commit removes all special handling for the `slug` attribute in our text. A variant of the text
input control that can handle formatting-as-slugs has replaced all the slugificiation code; simply
drop it onto a page and tell it the (must be unique) selector from which to get the data to be
slugified. It only looks up one tier of the DOM so be careful that both the text input and its slug
accessory occupy the same DOM context.
## Details
### The Component
Now that we know a (lot) more about Lit, this component has been slightly updated to meet our
current standards.
- web/src/components/ak-slug-input.ts
Changes made:
- The "listen for the source object" has been moved to the `firstUpdated`, so that it no longer has
to wait for the end of a render.
- The `dirtyFlag` handler now uses the `@input` syntax.
- Updated the slug formatter to permit trailing dashes.
- Uses the `@bound` decorator, eliminating the need to do binding in the constructor (and so
eliminating the constructor completely).
### Component uses:
The following components were revised to use `ak-slug-input` instead of a plain text input with the
slug-handling added by our forms manager.
- applications/ApplicationForm.ts
- flows/FlowForm.ts
- sources/kerberos/KerberosSourceForm.ts
- sources/ldap/LDAPSourceForm.ts
- sources/oauth/OAuthSourceForm.ts
- sources/plex/PlexSourceForm.ts
- sources/saml/SAMLSourceForm.ts
- sources/scim/SCIMSourceForm.ts
### Remove the redundant special slug handling code
- web/src/elements/forms/Form.ts
- web/src/elements/forms/HorizontalFormElement.ts
### A special case among special cases
- web/src/admin/stages/invitation/InvitationForm.ts
This form is our one case where we have a slug input field with no corresponding text source. Adding
a simple event handler to validate the value whenever it changed and write back a "clean" slug was
the most straightforward solution. I added a help line; it seemed "surprising" to ask someone for a
name and not follow the same rules as "names" everywhere else in our UI without explanation.
* After writing the commit message, I realized some of the comments I made MUST be added to the component.
* The `source` attribute needed its own comment to indicate that a `query()` compatible selector is expected.
* Added public/private/protected/# indicators to all fields. Trying to balance between getting it 'right' and leaving an opening for harmonizing style-sharing and state-sharing between (text / textarea), slug, password and (visible / hidden / secret).
* Removed the ids as requested; the default "look for this" matches the original behavior without requiring it be hard-coded and unchangable.
* web: Add InvalidationFlow to Radius Provider dialogues
## What
- Bugfix: adds the InvalidationFlow to the Radius Provider dialogues
- Repairs: `{"invalidation_flow":["This field is required."]}` message, which was *not* propagated
to the Notification.
- Nitpick: Pretties `?foo=${true}` expressions: `s/\?([^=]+)=\$\{true\}/\1/`
## Note
Yes, I know I'm going to have to do more magic when we harmonize the forms, and no, I didn't add the
Property Mappings to the wizard, and yes, I know I'm going to have pain with the *new* version of
the wizard. But this is a serious bug; you can't make Radius servers with *either* of the current
dialogues at the moment.
* This (temporary) change is needed to prevent the unit tests from failing.
\# What
\# Why
\# How
\# Designs
\# Test Steps
\# Other Notes
* Revert "This (temporary) change is needed to prevent the unit tests from failing."
This reverts commit dddde09be5.
* web: remove Lit syntax from always true attributes
## What
Replaces instances of `?loading=${true}` and `?loading="${true}"` with `loading`
## Why
The Lit syntax is completely unnecessary when the attribute's state is constant, and it's a few
(just a few) extra CPU cycles for Lit to process that.
More to the point, it annoys me.
## How
```
$ perl -pi.bak -e 's/\?loading=\$\{true\}/loading/' $(rg -l '\?loading=\$\{true\}')
$ find . -name '*.bak' -exec rm {} \;
$ perl -pi.bak -e 's/\?loading="\$\{true\}"/loading/' $(rg -l '\?loading="\$\{true\}"')
$ find . -name '*.bak' -exec rm {} \;
```
* Prettier had opinions
* web: move optional textual information out of attributes and into slots
## What
Replaces instances of:
```
<ak-empty-state header=${msg(...)}></ak-empty-state>
```
with
```
<ak-empty-state><span slot="header">${msg(...)}</span></ak-empty-state>
```
## Why
1. It's correct.
2. It lets us elide the decorations for any slots we aren't using.
3. It's preparation for moving to Patternfly 5
4. It annoyed me.
## How
Since we already have Patternfly Elements installed, we have access to the PFE-Core, which has the
unbelievable useful `SlotsController`. Using it, I created a conditional render template that will
only put in the header, body, and primary slots if there is something in the lightDOM requesting
those slots. The conditional template will still put the spinner in if the header is not provided
but the loading state is true.
I then had to edit all the places where this is used. For about 30 of them, this script sufficed:
```
perl -pi.bak -e 's/header="?(\$\{msg\([^)]+\)\})"?>/><span slot="header">\1<\/span>/' \
$(rg -l `<ak-empty-state[^>]header')
```
The other six had to be done by hand. I have tested a handful of the automatic ones, and all of the
ones that were edited manually. I'm pleasantly surprised that the textual rules [are inherited by
the slots as expected](https://htmlwithsuperpowers.netlify.app/styling/inheritable.htm).
* web: Add InvalidationFlow to Radius Provider dialogues
## What
- Bugfix: adds the InvalidationFlow to the Radius Provider dialogues
- Repairs: `{"invalidation_flow":["This field is required."]}` message, which was *not* propagated
to the Notification.
- Nitpick: Pretties `?foo=${true}` expressions: `s/\?([^=]+)=\$\{true\}/\1/`
## Note
Yes, I know I'm going to have to do more magic when we harmonize the forms, and no, I didn't add the
Property Mappings to the wizard, and yes, I know I'm going to have pain with the *new* version of
the wizard. But this is a serious bug; you can't make Radius servers with *either* of the current
dialogues at the moment.
* This (temporary) change is needed to prevent the unit tests from failing.
\# What
\# Why
\# How
\# Designs
\# Test Steps
\# Other Notes
* Revert "This (temporary) change is needed to prevent the unit tests from failing."
This reverts commit dddde09be5.
* web/admin: Provide `hidden` text and textarea components
## Details
This commit provides two new elements (technically, since they're API-unaware), one for `<input
type="text">`, and one for `<textarea>`, that provide for the ability to create fields that are (or
can be) hidden. A new boolean attribute, `revealed`, shows the state of the component (the content
is therefore *not* revealed by default).
It also includes a third new element, `ak-visibility-toggle`, that creates a hide/show toggle with
all the right icons, styling, and eventing. It's straightforward, and isolating it improved the
DX of everything that uses that feature by quite a bit.
Storybook stories (with autodoc documentation) have been provided for `ak-hidden-text-input`,
`ak-hidden-textarea-input`, and `ak-visibility-toggle`.
## Maintenance Notice
As a maintenance detail, the field `ak-private-text` has been renamed `ak-secret-text` to reflect
its usage, and the places where it was used have all been changed to reflect that update.
* web/component: embed styling (for now) to handle the lightDom/shadowDom/slot conflicts in HorizontalLightComponent and HorizontalFormElement
* Comments and Types. I really shouldn't have to catch this stuff with my eyeballs.
* fix typo
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
---------
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
Co-authored-by: Jens Langhammer <jens@goauthentik.io>
* web: Add InvalidationFlow to Radius Provider dialogues
## What
- Bugfix: adds the InvalidationFlow to the Radius Provider dialogues
- Repairs: `{"invalidation_flow":["This field is required."]}` message, which was *not* propagated
to the Notification.
- Nitpick: Pretties `?foo=${true}` expressions: `s/\?([^=]+)=\$\{true\}/\1/`
## Note
Yes, I know I'm going to have to do more magic when we harmonize the forms, and no, I didn't add the
Property Mappings to the wizard, and yes, I know I'm going to have pain with the *new* version of
the wizard. But this is a serious bug; you can't make Radius servers with *either* of the current
dialogues at the moment.
* This (temporary) change is needed to prevent the unit tests from failing.
\# What
\# Why
\# How
\# Designs
\# Test Steps
\# Other Notes
* Revert "This (temporary) change is needed to prevent the unit tests from failing."
This reverts commit dddde09be5.
* web/admin: fix language in certificate import modal
There are two buttons on the CertificateKeyPairList: "Create" and "Generate". This language is confusing, as it is unclear what the difference between those two options could be. The
second generates and stores a new key pair; the first only creates an entry in our database for an existing pair generated externally. After a short conversation, changing it to
"Import" seemed like a safe option.
\# What
\# Why
\# How
\# Designs
\# Test Steps
\# Other Notes
---------
Co-authored-by: Jens Langhammer <jens@goauthentik.io>
* Only attach logo to email if used
* Fix MIME logo attachment to adhere to standard
* format, fix web
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
* not tuple
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
---------
Signed-off-by: Jens Langhammer <jens@goauthentik.io>
Co-authored-by: Jens Langhammer <jens@goauthentik.io>
Teffen Ellis
Jens L.
Ken Sternberg
Leandro4002
CodeMax IT Solutions Pvt. Ltd.
Marc 'risson' Schmitt
Christian Elsen