msabitov / effcss / #17

100%

push

v6.0.0:

Added:
- `getContexts()` has been added, returning all active isolated contexts keyed by their key
- each `context` now have lifecycle controls, and `Symbol.dispose` is wired to `dispose` on every context for `using` support
- selector modifiers (postfixes) on `className`/`id`: pseudo-classes, at-rules, arbitrary attributes
- `id` and `className` now accept a plain object (without `EffCSS` tokens inside) or a function (lazy, may reference tokens)
- `style` has been added, it accepts both a function `() => ({...})` and a plain object

Changed:
- fully lazy model: styles materialize only when a resolver is used (called or stringified); `serialize()` returns only already-created styles and does not materialize
- `configure` has been replaced with `setConfig`/`getConfig` (fixes options before the first use; returns `false` if the provider already exists)
- token scheme unified to `key_type-counter`; `key` is required for `context`/`setConfig`
- `sideEffects: false` added to `package.json` for safe tree-shaking
- `variable.set(value)` uses closure-based updates instead of CSS rule parsing

Removed:
- `classNames`, `attributes`, `customStyles` and their `lazy*` variants → use `className`/`id` modifiers and `style`
- `update` → use `varResolver.set(value)` directly
- `stylesheet` and all `*Stylesheet` public accessors → use `serialize()` / `clear()`
- `subscribe` (and the event system) → no replacement
- `serializeMeta`, `serverMeta` and meta reading from `<script type="application/json">` → no replacement
- `serialize(resolver)` → use `serialize()` (resolvers share stylesheets)
- `configure` → use `setConfig`/`getConfig` (must be called before the first styles usage)
- `lazy`, `minify`, `emulate` options → no replacement (lazy is the default)
- multiple at-rules (`variables`, `animations`, `containers`, `fonts`) → removed; `layers([...])` remains as the only batch at-rule

139 of 141 branches covered (98.58%)

Branch coverage included in aggregate %.

295 of 295 new or added lines in 1 file covered. (100.0%)

346 of 346 relevant lines covered (100.0%)

67.93 hits per line