Error parsing input file: Unexpected value

[a1rb0rn3]

When creating input files with a normal text editor in Linux and then trying to use your tool with those files doesn't work.
The tool exits with the error: Error parsing input file: Unexpected value: [...some garbage].

The problem is that normal text files get encoded with UTF-8 and your tool expects UTF-16LE encoding for the input files.

Solution:
A solution for the above mentioned issue is to simply convert those files to UTF-16LE.

With iconv:

iconv -f UTF-8 -t UTF-16LE input.txt > output.txt
printf '\xFF\xFE' > bom.txt
cat bom.txt output.txt > final_output.txt

With vim:

vim input.txt
:set bomb
:set fileencoding=utf-16le
:wq! output.txt

Note

The Byte Order Mark (BOM) is not needed for the input file to work as setup_var.efi strips it away anyway. However, without it, displaying the input file in the UEFI shell would be a bit ugly.

Expectations:

I don't know if you want to make your tool compatible with other file encodings, but I would somehow expect this to be mentioned in the documentation somewhere. Maybe this will help other people in future facing the same issue. It took me a while to figure out what was causing the problem :D

[datasone]

Thanks for detailing the encoding problem, the origin design that assuming UTF-16LE is for recognizing UEFI shell redirected output, and I think adding the whole libiconv or encoding-rs for encoding conversion is a little too much for a UEFI application.

Some other not that important details about the design:

• When the read from file feature was originally designed and coded, there was no fs support in uefi-rs, so reading from stdin (piped) is chosen.
• UEFI shell will interpret file content as UTF-16, so we must use codec conversion to interprete UTF-8 bytes that are wrongly read as UTF-16.
• Reading from file using fs seems to be a better idea, but it increases confusion (if we don't want to add codec conversion), as you must use UTF-8 for file specified in arguments (which will be read using fs api, and intepreted as UTF-8) and use UTF-16 for file specified by feeding into stdin.

I will first fix document to add text encoding information and link to this issue. And see if there will be a better design for handling both UTF-8 and UTF-16LE files.