Character encoding in script files

[jfcorbett]

Is there any particular requirement for character encoding in WEng (and WAsP) script files? If so, does it depend on the operating system? What else could be the cause of the problem described below?

This one took me a while to debug: a comment was causing a "Object not set or With block variable not set" error! Turns out the compiler was tripping on the "special" character "ç" in my commented-out name:

' Author: Jean-François Corbett

Error went away when comment changed to:

' Author: Jean-Francois Corbett

This was on a 64-bit Windows 7 workstation. The exact same script gave no error on my Windows XP laptop.

[Duncan]

Hi Jean-François,

I'm actually surprised you got this problem, and I'm sorry you had to diagnose it yourself. As you can imagine, we usually flush these issues out early because of the 'ø' in 'Risø'!

I think we've seen something similar with the '°' degree character in report text on Chinese systems, and so have sometimes needed to replace these with character codes. But getting caught out with a code comment just seems unfair...

The VBScript code in a script is in a CDATA Block, so the XML parser should be blindly taking what's there without parsing it. The XML is tagged as UTF-8, so it should handle these characters without problem. I guess that when we load that VBScript into the scripting runtime, passing it as a string, that's where things go wrong.

I have logged this as a bug to investigate. Thanks for the report.