Dear Pandoc community,
I have encountered two issues regarding Pandoc's handling of quotation marks in cases where Japanese and English texts are mixed.
I am using Pandoc version 3.1.2 on macOS 12.6.3, and I can reproduce these issues. If these are indeed bugs, I am planning to submit them as issues on GitHub. However, I would appreciate any guidance if these issues arise from my incorrect usage.
Issue 1: Conversion of English phrases within Japanese text
I have observed the following issue. "input.md" is the input file, and "input.tex" is the output file.
$ pandoc input.md -o input.tex
input.md:
その人は"Hello, world!"と言いました。
input.tex:
その人は''Hello, world!{}``と言いました。
However, the conversion is correct when spaces are added before and after the double quotation marks.
input.md:
その人は "Hello, world!" と言いました。
input.tex:
その人は ``Hello, world!'' と言いました。
Issue 2: The quotation marks are treated as Japanese text
When converting with Pandoc, the quotation marks are treated as Japanese text, resulting in an unnaturally wide gap. I have confirmed this using two files, "preamble.tex" and "input.md," and specifying as follows:
$ pandoc input.md -o input.pdf --pdf-engine=xelatex -H preamble.tex.
preamble.tex:
\usepackage{fontspec}
\setmainfont{Georgia}
\setjamainfont{BIZ UDMincho Medium}
input.md:
---
documentclass: bxjsarticle
classoption: pandoc
papersize: a4
fontsize: 10pt
---
# はじめに
その人は "Hello, world!" と言いました。
That person said, "Hello, world!"
In contrast, when I directly write the content in TeX and output it using $ xelatex test.tex, the quotation marks are treated as English text, and the expected output is obtained.
test.tex:
\documentclass[a4paper,xelatex,ja=standard]{bxjsarticle}
\usepackage{fontspec}
\setmainfont{Georgia}
\setjamainfont{BIZ UDMincho Medium}
\title{テスト}
\begin{document}
\maketitle
\section{はじめに}
その人は ``Hello, world!'' と言いました。
That person said, ``Hello, world!''
\end{document}
Shigeru Kobayashi