Conversation
|
Thanks for quick turnaround on this -- but I worry for lines 179-180 & 186-187 A possible fix is to read from whichever property exists: There is also a concern with Vite 5–7: semicolons present (Rollup output), but with Vite 8 non-native path: polyfill serialised via .toString() from TS source (no semicolons) .. so the |
Tests say otherwise actually, still works fine AFAICT. In fact, as you can see in the diff here, Vite 8 actually threw an error upon attempting to use my dummy
I have absolutely no idea what you're trying to say here. |
|
On rollupOptions: Thanks for testing that: Good to see that Vite 8 still honours On the polyfill: Sorry, that was unclear. In plain terms: the plugin patches Maybe seeing the changes in this draft will help? #35 |
I can't imagine they'll break that any time soon, even if it's discouraged; the entire ecosystem relies on that pretty freaking heavily.
All you're doing there is taking the 3 branches and merging them into a single check; it's not actually more flexible in any way and is just a heck of a lot slower to execute & debug when something inevitably changes. |
|
Ok - fair enough ... I'll close. |
Closes #33
Bumped the deps locally & ran the suite, all pass (after a couple tweaks) except the source map errors test. For whatever reason, with Vite 8, line numbers are off by 1. Not a blocking issue though.