Fix JQL Japanese character encoding issue

[kentoyoshida]

Summary

• Fix an error that occurs when JQL queries contain Japanese characters
• The issue was caused by StringEntity defaulting to ISO-8859-1 encoding instead of UTF-8
• Explicitly specify UTF-8 encoding to properly handle Japanese and other multibyte characters

Problem

When using Japanese characters in JQL queries (e.g., labels="テスト"), the following error occurred:

Error: Could not guess schema due to empty data set

However, the same query works correctly when using curl with --data-urlencode option.

Root Cause

In JiraClient.java:228, the StringEntity constructor was called without specifying character encoding:

request.setEntity(new StringEntity(body));

Apache HttpClient 4.x defaults to ISO-8859-1 encoding when no encoding is specified, which causes Japanese characters to be incorrectly encoded when sent to the JIRA API.

This issue was introduced in commit 2692818 "use new jql search API (#6)" when the implementation was changed to use POST requests with JSON body instead of GET requests with query parameters.

Changes

Modified Files

• src/main/java/org/embulk/input/jira/client/JiraClient.java

  • Added StandardCharsets import
  • Changed new StringEntity(body) to new StringEntity(body, StandardCharsets.UTF_8)

• src/test/java/org/embulk/input/jira/client/JiraClientTest.java

  • Added test case test_searchIssues_withJapaneseJql() to verify Japanese JQL queries work correctly

• src/test/resources/jira_client.json

  • Added test data for Japanese JQL query test case

Testing

• ✅ All existing tests pass
• ✅ New Japanese JQL query test case passes
• ✅ Verified UTF-8 encoding in compiled bytecode using javap

🤖 Generated with Claude Code

[yas-okadatech]

LGTM!