CSS files with UTF-8 BOM triggers CSS errors

[alitokmen]

I noticed that if a CSS file has a UTF-8 BOM then HtmlUnit seems to not like it - It throws a CSS error:

com.gargoylesoftware.htmlunit.DefaultCssErrorHandler error
WARNING: CSS error: 'http://localhost:59425/cargo-daemon-webapp/cargo.css' [38:1] Error in style rule. (Invalid token "body". Was expecting one of: <S>, <LBRACE>, <COMMA>.)

The UTF-8 BOM doesn't seem to "bother" Web browsers, could HtmlUnit also be updated to behave similarly?

[rbri]

can you please provide the whole css file.

[alitokmen]

Thanks for your quick reply!

The CSS file is on: https://github.com/codehaus-cargo/cargo/blob/d60fb4e650557ff8754960c6a4ea608d29478db7/extensions/daemon/webapp/src/main/webapp/cargo.css

[rbri]

Thanks, will have a look during the next days.

[rbri]

The parsing fails because the file has a bom header. Do you face the problem while using HtmlUnit with some webpage?

[alitokmen]

@rbri , you know very well your code !! :)

Yes, removing the UTF-8 BOM fixed the issue. I think it could be worthwhile to enable HtmlUnit to work with UTF-8 BOMed files, as Windows tends to save with a BOM and browsers seem to be fine with it.

[rbri]

I think it could be worthwhile to enable HtmlUnit to work with UTF-8 BOMed files, as Windows tends to save with a BOM and browsers seem to be fine with it.

Yes this is the plan - removing the BOM is not so tricky (i already did it at some other places) but it will require some time to think about test cases to ensure the same behavior as real browsers.

[rbri]

Ok, hopefully i have covered the most important cases. Please check twitter for a new snapshot build.

Many thanks for reporting this.