MSConvert outputs wrong encoding for Waters raw files

[t0mdavid-m]

We are using MSConvert to convert Waters raw files to mzML. Unfortunately we have been experiencing issues with downstream processing. The issues seem to be caused by the encoding of the mzML files.

The mzML files are encoded with Windows-1252 but in the header UTF-8 is reported:

<?xml version="1.0" encoding="utf-8"?> 

This causes our XML-Parser to assume UTF-8 and fail when running into non-ASCII characters (in our case µ/b5 was problematic).

I would suggest to either convert the files to UTF-8 or report the correct encoding in the header (I could not find any requirements for UTF-8 in the mzML specification).

[chambm]

What element/attribute had the non-ASCII character? Filepaths should be UTF-8 encoded and XML ids and idrefs should be xHHHH encoded.

[t0mdavid-m]

This actually only occurs in the id and idrefs for the file I am looking at.

For instance:
id="µBSM System Pressure"

It looks like no xHHHH encoding occurrs. Instead, it looks like the character is included as is using the original encoding. The only character reference I could find in the document was ".

[chambm]

Interesting. Looks like Visual C++'s isalpha() function says µ is an alphabetic character. But their example code implies it's not:

https://learn.microsoft.com/en-us/cpp/c-runtime-library/character-classification?view=msvc-170

Generally these routines execute faster than tests you might write and should be favored over. For example, the following code executes slower than a call to isalpha(c):

if ((c >= 'A') && (c <= 'Z')) || ((c >= 'a') && (c <= 'z'))
return TRUE;

Looks like I need to go for the simpler approach.