Source code could not be parsed (Unexpected UTF-8 character)

[kefniark]

Description

I have a language detection library tinyld which works in node (CJS) and browser (ESM) without trouble.

But I run into a strange issue when I try to use it in deno.

error: The module's source code could not be parsed: Unexpected character 'ำ'

The issue come from a big object where I store n-gram for different languages (as utf-8 strings)
I have thousands of other keys in there which do not cause any problem (japanese, chinese, arabic, ...)

 //...
 อ: 9,
 ฮ: 9,
 ะ: 9,
 า: 9,
  ำ: 9,
 //...

But I don't now why, there is one specific thai character, ำ which cause that Unexpected character error
If I remove that line, it works fine.

So my question, is there a reason why this specific utf-8 character cause trouble?

P.S. I noticed that the same problem seem to appear in VSCode.
It kinda feel like this character is detected as a single quote, maybe more a typescript issue than deno.

[andreubotella]

This is not a bug with Deno's UTF-8 decoding, but an upstream bug with swc's parsing of JS/TS identifiers (variable names, etc., including non-string property names in object literals, as in this case). I filed swc-project/swc#3341.

[kefniark]

Thanks for the quick feedback 👍

[andreubotella]

This is now fixed upstream: swc-project/swc#3867

[bartlomieju]

@dsherret can we close this issue now?

[dsherret]

Yes, it seems fixed in 1.21