Rich Harris
Rich Harris
### Describe the bug in: ```svelte $: mobile = width < 640; let width = 0; {mobile ? 'mobile' : 'desktop'} ``` out: ```svelte let mobile = $derived(width < 640);...
### Describe the bug This... ```svelte const { foo } = x(); {#each foo as f}{/each} ``` ...gets turned into this by the migrator: ```svelte const { foo } =...
### Describe the bug At a couple of places during parsing, we need to skip ahead to figure out what to feed Acorn — inside `read_pattern`, and as of #12070...
### Describe the problem We replaced the Svelte 4 `legacy` option (which added IE compatibility at the cost of bloatier output) with a completely different `legacy` option. Per #12100, this...
I have a hunch that we can simplify reactivity a bit. `check_dirtiness` seems unduly complex, and it doesn't make sense to me that we call `mark_reactions` inside `update_derived` (the reactions...
Trying to salvage parts of #12073, since the memory leak thing is proving finicky. We have some slightly elaborate dependency tracking logic in the codebase. It's not explained anywhere but...
#11989 got me thinking — aside from the question of keeping `then`/`catch` blocks mounted rather than recreating them each time (as though they contained `{#key ...}` blocks), why do we...
Follow-up to #11947. Right now our HTML validation (for checking which elements are allowed inside which other elements) isn't great — we have special handling for `` (why?) and interactive...
stumbled on this while working on #11730
### Describe the bug Not exactly sure where to report this but this feels like as good a place as any — `vitePreprocess` takes input like this... ```svelte a b...