RT#7802: Bug: Convert JP text file to PDF errors

[rakunavi]

Hello all,

I have noticed that character corruption occurs when converting Japanese text files with JIS (ISO-2022-JP) encoding via the "Convert Text Files to PDF" dialog, even when "Japanese (JIS)" is selected in the "Text Encoding" options.

• figure1.png

I have attached a video demonstrating this behavior, comparing the conversion process between PDF-XChange Editor and Adobe Acrobat X using the same sample text file. The attached image highlights the corrupted parts with yellow markers.

• figure2.png

• SampleFiles1.zip

• Video1.webm

During my investigation, I noticed certain patterns in how the corruption occurs. Specifically, I generated 350 text files using the following AutoHotkey v1 script:

Code: Select all

Loop 350 {
	count := A_Index
	Loop 6 {
		if (A_Index == 1) {
			paragraph := ""
			Loop, % count
				paragraph .= "あ"
			output := paragraph
		} else
			output .= "`n`n" . paragraph
	}
	FileAppend, %output%, %count%.txt
}

To put it in words, the key points for generating text files are as follows:

1. Each text file consists of 6 paragraphs.
2. Each paragraph consists of the Hiragana character "あ" (JIS code: 0x2422), with the character count increasing by one for each file.
3. Paragraphs are separated by a blank line.
4. The encoding is Japanese JIS (ISO-2022-JP) with LF line endings.

• count8.png

• SampleFiles2.zip

• Video2.webm

Observed Symptoms:

1. Within the paragraphs, each character あ (JIS code: 0x2422) is incorrectly rendered as the string $". This results in long sequences of corruption such as $"$"$"$".
2. The specific corruption string ←(B (likely related to the ESC (B escape sequence) appears at the very end of the text.
3. Notably, this ←(B corruption only occurs when the text file ends without a trailing newline (e.g., ...あ[EOF]). If the file ends with a newline (e.g., ...あ[LF][EOF]), this specific trailing corruption does not occur.
4. On pages 1 through 81, as well as on specific pages (pages 124 and 252), no character corruption occurs at all within the paragraphs.

I hope these specific patterns provide the developers with a clue to identifying the root cause.

Hoping that the above information will be of some help to you.
Thank you so much for your continued support.

Best regards,
rakunavi

- PDF-XChange Editor PRO Version: 10.8.4 build 409
- OS Version: Windows 11 Pro / Home 25H2 Build 26200.8037
- Adobe Acrobat X Standard Version 10.1.16
- PC Model: GMKtec Nucbox M7 Pro with HUION Kamvas Pro 19 / Lenovo IdeaPad C340-15IWL

[EDIT] Since a ticket has been officially issued for this request, I have changed the topic title as follows. 2026-3-27 6:07 JST (UTC+9)

• Previous Title: Character corruption when converting Japanese JIS (ISO-2022-JP) text files to PDF
• New Title: RT#7802: Bug: Convert JP text file to PDF errors

[Daniel - PDF-XChange]

Hello, rakunavi

Thank you for the detailed report. I have reproduced the issue and raised this with the Dev team for review and hopefully, resolution.

RT#7802: Bug: Convert JP text file to PDF errors

Kind regards,

[rakunavi]

Thank you, Daniel, for creating the ticket. I really appreciate you taking the time out of your busy schedule.

By the way, if you convert a Japanese text file encoded in Japanese EUC to PDF, a blank PDF file containing no objects at all is generated.

• Animation.gif

• SampleFile.zip

This is exactly the same behavior reported in the following topic just the other day. I am providing this additional report for your reference.

• RT#7799: Issue with Converting Text Files to PDF
  viewtopic.php?t=49328

Please give my best regards to developers.

Best regards,
rakunavi

- PDF-XChange Editor PRO Version: 10.8.4 build 409
- OS Version: Windows 11 Pro / Home 25H2 Build 26200.8037
- PC Model: GMKtec Nucbox M7 Pro with HUION Kamvas Pro 19 / Lenovo IdeaPad C340-15IWL

[Daniel - PDF-XChange]

Hello, rakunavi

Yes, that one was also discussed at the same time. In my testing I stumbled on it just by leaving the encoding on Japanese "Auto". The Dev team is aware and working on that as well.

Kind regards,