You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In vite, in order to give users a better development experience, we didn't convert all JS into ast for lexical analysis. There are some specific syntaxes that are implemented through string matching.
We will encounter the following boundary problems.
I think we may need to see it in a PR to check how good the abstraction is at cleaning up the code. One detail to take into account is that in some cases we want to first do a clean up and then re-use the cleaned code through several operations. Maybe bundling the two stages in a single operation isn't the best. But it may be good for some cases.
I was also thinking on some basic utils like withEmptyComments(code) and withEmptyStrings(code)
Clear and concise description of the problem
In vite, in order to give users a better development experience, we didn't convert all JS into ast for lexical analysis. There are some specific syntaxes that are implemented through string matching.
We will encounter the following boundary problems.
For example:
workerImportMetaUrl
pluginassetImportMetaUrl
pluginoptimizer/scan
but ignored data is also needed (
/* @vite-ignore */
is needed.)html
pluginit will got error in this case (#6525)
The text was updated successfully, but these errors were encountered: