Decode error when encountering accents chars

Good evening,

In some of the scripts i’m executing, people used accents chars in comments (not in vars at least! 😌 )

I created a small case that is easier to debug. The following code crashes :

duk_peval_string(ctx, "/* àéè */");

The error is : SyntaxError: decode failed (line 0)

If i remove the accents, everything works fine

Thansk in advance Best

R. Bloch

[svaarala]

The other school of thought is that UTF-8 decoding should in general be strict and always reject invalid sequences so that they can be detected and fixed. Pros and cons in each I guess.

I’ll improve the error message.

[rbloch]

Yes usually it’s way more developer friendly to tolerate invalid char values and replace them like you said.

I agree on the replacement char solution, that would trivially solve 99% of the problems as usually the non proper chars are set in the comments.

Maybe another simpler improvement would be to just slightly change the error message to SyntaxError: char decode failed

Which makes you understand immediately that you’ve some weird chars -> encoding. It’ll still be annoying as the dev will have to remove them somehow (src or on the fly) but at least he’ll be 100% sure of what is the problem. I think that just a decode failed is not verbose enough.

Best 👍