Relating to the following question I asked:
Does vref of varioref work with page numbers which are not built-in arabic?
I would like to know the consequences I might encounter if I do the following as a stopgap solution:
I will replace 0, 1, 2, ..., 9 with ၀, ၁, ၂, ၃, ..., ၉ as shown below using FontForge. And then I'll just use LaTex Arabic numbering as it is. My intention for doing this is I expect that LaTex and its packages will use some sort of internal calculation base on normal Arabic numbers. If I inject other unicodes characters in place of the original Arabic numbers, those calculation, most likely, will go wrong. For instance, vref does not behave as expected and has the problem I described in the above question. So, I am making changes at the font level.
In this pic,I've edited my unicode font, Padauk, to clarify what I means.

%%% Auto numbering in Myanmar %%% %This macro is to produce myanmar numbering by adopting the thai numbering method \makeatletter \def\@mmnum#1{\expandafter\@@mmnum\number#1\@nil} \def\@@mmnum#1{% \ifx#1\@nil \else \char\numexpr#1+"1040\relax % 1040 is zero character for Burmese. \expandafter\@@mmnum\fi } \renewcommand\@arabic{\@mmnum} % to reset number in \arabic to \mmnum. – Pyi Soe Jul 17 '22 at 08:08\XeTeXinterwordspaceshaping=2option in XeLaTex, the word is not rendered properly. – Pyi Soe Jul 17 '22 at 16:22