Tkinter show different symbol for the same unicode character depending on which character is used first

[patrik-gustavsson]

Depending on which character is used first in a font instance, that will set the look for the symbols for other unicode characters. To demonstrate this behavior see the following code (I have tested on windows 10 and 11, python 3.11.1):

import tkinter as tk
from tkinter.font import Font

app = tk.Tk()

arial1 = Font(family="Arial", size=32)
arial2 = Font(family="Arial", size=32)

tk.Label(app, text="\u25b6 \u23EE \u270E", font=arial1).pack()
tk.Label(app, text="\u23EE \u25b6 \u270E", font=arial1).pack()
tk.Label(app, text="\u23EE \u25b6 \u270E", font=arial2).pack()
tk.Label(app, text="\u25b6 \u23EE \u270E", font=arial2).pack()

app.mainloop()

This is my understanding of the problem: If a font instance does not support a unicode character, then the font instance will select a fallback font-family. However, depending on which unicode character that is used, I believe different fallback font-families are selected and therefore the look of the characters may differ.

I do not know if this is a bug in tkinter because I noticed this behavior in VSCode, IDLE, Google Chrome, Edge and Slack.

I asked this question on stackoverflow if you want to read more on the comments we had, see https://stackoverflow.com/questions/75026676/why-does-some-unicode-characters-change-shape-depending-on-which-character-is-fi/.

[serhiy-storchaka]

Tkinter is a thin wrapper around the Tcl/Tk library. It is not responsible of rendering characters, can not and should not interfere with it. If it is a bug, file a report on the Tk bug tracker. Although, I think that they are not responsible of this too and just use system libraries to output characters. If you see this behavior in programs which do not use Tk, it is most likely not related to Tk. Also, it may not be a bug at all.