[GROOVY-9954] JsonSlurper LAX parser skips the first character after comments

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 4.0.0-alpha-2
• Fix Version/s: 3.0.8, 4.0.0-alpha-3
• Component/s: JSON
• Assignee: Unassigned
• Reporter: Yuki Kawata
• Created: 25/Feb/21 05:11
• Updated: 02/May/21 09:16
• Resolved: 02/Mar/21 14:57

Description

Problem

JsonSlurper with JsonParseType.LAX skips the first character after comments.
This can cause unexpected parse results or even failures in some cases.

Example 1

Input

println new JsonSlurper().setType(JsonParserType.LAX).parseText("""
// comment
{
  "flag": true,
  "flag2": false
}
""")

Expected result

[flag2:false, flag:true]

Actual result

flag

Example 2

Input

println new JsonSlurper().setType(JsonParserType.LAX).parseText("""
{
  /* comment */"flag": true,
  "flag2": false
}
""")

Expected result

[flag2:false, flag:true]

Actual result

The current character read is 'f' with an int value of 102
expecting current character to be ':' but got 'f' with an int value of 102

line number 4
index number 35
  "flag2": false
...^
groovy.json.JsonException: expecting current character to be ':' but got 'f' with an int value of 102

Root cause

JsonParserLax.handleComment() (and handleBashComment()) sets the index of the parsed string to the first character after the comment.
This index is incremented by the for-loop before checking the next character.
This can cause some unexpected behaviors if the skipped character is not a whitespace.