/* ═══════════════════════════════════════════════════════════════════════════
   Third-party CSS, quarantined in a cascade layer.

   Bootstrap and Radzen used to be plain <link>s, and Radzen's sat LAST in
   index.html — after app.css and after the scoped-CSS bundle. Every tie on
   specificity therefore went to Radzen, which is why app styles had grown a
   crop of `!important` and hard-coded hex values to shout over it.

   Anything inside a layer loses to unlayered CSS regardless of specificity or
   order, so importing the vendors here makes every one of our own rules win by
   default. Load order stops being load-bearing and the overrides can go back to
   being ordinary declarations.

   Paths resolve relative to this file (/css/), hence the `../`.
   ═══════════════════════════════════════════════════════════════════════════ */
@layer vendor;

/* Bootstrap was removed. An audit of every class attribute in every .razor file
   found zero Bootstrap classes in use — no grid, no utilities, no components. It
   was 230KB of CSS the app had stopped using, plus a second opinion about what
   `.btn` and `.card` mean, competing with Radzen for the same selectors.
   pt-tokens.css owns the design system; Radzen owns the components. */
@import url("../_content/Radzen.Blazor/css/default.css") layer(vendor);

/* Leaflet belongs in the layer for the same reason. It was a plain <link>, which
   left it unlayered and competing on specificity: `.leaflet-pane > canvas` (0,1,1)
   beat our `.pt-flow-canvas` (0,1,0) and silently overrode the flow overlay's
   z-index. Inside the layer it loses to every app rule regardless. */
@import url("../lib/leaflet/leaflet.css") layer(vendor);