Symbol file format?

Poking around a qsym file in a text editor, there are leading undiscernible characters?
At first I thought it was a BOM marker, but Notepad++ does not recognize the sequence and claims the file is ANSI.
I am looking to convert prior symbols programmatically and wanted to understand the format.

Ex. → C:\QSPICE\Qorvo\SiC FET\650V\UF3C065030B3.qsym
ÿØÿÛ«symbol UF3C065030B3
«type: MN»
«description: Cascode SiC FET»
«library file: Level2010.txt»

They are just regular characters ÿ Ø ÿ Û using CP1252 encoding¹ (basically Windows’ proprietary legacy ANSI² encoding for the western world). You can also leave them off (like in my additional symbols), they are just skipped, as far as I can tell. I guess they are intended as marker but unfortunately they mark a file as a corrupted JPEG outside of QSPICE (like if you want to upload the file as *.txt in this forum³).

¹ The blue text are links and readers are supposed to click them.

² I don’t know how well it works with Japanese Windows where ANSI is a variant of Shift JIS.

³ This rejection and more could probably be configured but it looks like they are still getting familiar with the forum software.

I put those charactors there so certain e-mail servers don’t convert the contents of the file. QSPICE symbols and schematics look like ASCII, and can be edited in notepad or any other eight-bit clean otherwise “ASCII” editor. But they are not ASCII. So I give the files the same majic number as JPEG. E-mail servers don’t touch the contents of JPEG.

I guess currently the easiest way to figure out the parameters is to select a (modified) object (click or select with mouse rectangle), copy (Ctrl+C), and paste (Ctrl+V) into Notepad++. Possibly modify, select and copy in Notepad++ and paste back into QSPICE. In the symbol editor click background and copy or use right mouse button and select Copy Symbol to Clipboard. Then paste into Notepad++ or QSPICE schematic or symbol editor.

I simply put all the symbols into a single ANSI file and then used a script to split it into individual files.

UTF-8 encoding probably wouldn’t have that problem nowadays (already when XVII came out; nobody would use a system today where an emoji wouldn’t pass :person_facepalming:).