Bug: On Windows, Unicode file name not supported by tex binaries

John Collins jcc8 at psu.edu
Sat Nov 11 03:44:49 CET 2023

A few further observations:

I find the identical bug on macOS.  When I run

      xelatex Banana🍌.tex

I get the same errors as on Windows. The messages in the texput.log file 
contain incorrect UTF-8 encodings: Instead of the 4-byte coding for 🍌 
(Banana), which is 0xF0 0x9F 0x8D 0x8C, I see a 3-byte coding 0xEF 0x8d 0x8c. 
The relevant lines in the texput.log file are identical between the Windows and 
macOS versions.

So I don't think this is a Windows-specific bug in xelatex, but is OS-independent.

I did verify that pdflatex, lualatex, and uplatex have no problem on both 
Windows and macOS with the same file, just as Akira found on Windows.

On the other hand, there is no problem with xelatex when the filename contains 
Cyrillic characters

     xelatex Renormalization-Перенормовка.tex

even though on Windows, these characters are outside the system code page with 
my default settings.

Relevant to the original report of the bug, is that on Windows, Perl scripts 
(e.g., pdfcrop, latexmk) fail when applied to a file whose name contains 
characters outside the system code page.  Those can be made to work correctly 
with such filenames by changing the system code page appropriately. (That's 
done by Settings -> Language & region -> Administrative language settings -> 
Change system locale -> "Beta: Use Unicode UTF-8 for worldwide language support".)

However, the results of running

     xelatex Banana🍌.tex

are unchanged after that change of system code page.

John Collins

On 11/9/23 5:01 PM, Akira Kakuto wrote:
> I can confirm that
>   xelatex.exe --interaction=errorstopmode --synctex=-1 "Banana🍌.tex"
> does not work, but
>   pdflatex.exe --interaction=errorstopmode --synctex=-1 "Banana🍌.tex"
> works on Windows. (Same for uplatex and lualatex).
> Takuji Tanaka-san supported unicode file names for pdftex, xetex, euptex and 
> luatex.
> It seems incomplete for xetex.
> Best,
> Akira

More information about the tex-live mailing list.