demo project to illustrate the per-character font rendering problem

By default, jasperreport will use itext-2.1.7 for PDF rendering

This project will switch to itext-5 rendering due to client requirement The following source code are adopted to use itext5 instead (for package under com.jaspersoft.*) https://github.com/Jaspersoft/jasperreports-pdf-lib5

Prequisite

• java 1.8

Usage

use gradlew test to test the report output the test will read the file demo.txt and render to the PDF using the jasper design file remarks.jrxml

Motivation

For Chinese Character, the character set has been expanded to various CJK Extension in the unicode plane. It is impossible to have a single font to have cover all the character.

In CSS, we can define a list of font to be used, which the font appear in the later sequence can be served as fallback We want the same behavior for jasperreport PDF output.

Expected output

the text should be rendered by these fonts, by the following preferring order:

• Noto Serif (for all latin characters)
• Noto Serif CJK TC (for all normal CJK characters)
• HanaMinA (for CJK characters cannot be found in Noto Serif CJK TC)
• HanaMinB (for CJK extension-B/C/D/...)
• HanaMinPlus (for CJK extension-B/C/D/...)

which roughly translate to this css font-family: 'Noto Serif', 'Noto Serif CJK TC', 'HanaMinA', 'HanaMinB', 'HanaMinPlus', serif

actual output

Inside the docs/, it contains the output and diagnosis You can use PDF-XChange Viewer to view the actual font specified https://www.tracker-software.com/product/pdf-xchange-viewer

jr-6-12-2-patched.pdf

you can see that the text are rendered by 3 different fonts

• abc -> Noto Serif
• 𨋢 -> HanaMinA
• 所有 -> Noto Serif CJK TC

jr-6-12-2-orig.pdf

• abc -> Noto Serif
• 𨋢 所有 -> HanaMinA

Fonts copyrights

Noto fonts https://www.google.com/get/noto/

HanaMinA / HanaMinB / HanaMinPlus fonts https://fonts.jp/hanazono/