[SPARK-27193] CodeFormatter should format multi comment lines correctly

• Type: Improvement
• Status: Resolved
• Priority: Trivial
• Resolution: Fixed
• Affects Version/s: 2.4.0
• Fix Version/s: 3.0.0
• Component/s: SQL
• Assignee: wuyi
• Reporter: wuyi
• Due: 18/Mar/19
• Created: 18/Mar/19 15:24
• Updated: 19/Mar/19 07:15
• Resolved: 19/Mar/19 06:54

Description

when enable `spark.sql.codegen.comments`,  there will be multiple comment lines. However, CodeFormatter can not handle multi comment lines currently:

 
Generated code:
/* 001 */ public Object generate(Object[] references)

{ /* 002 */   return new GeneratedIteratorForCodegenStage1(references); /* 003 */ }

/* 004 */
/* 005 / /*
* Codegend pipeline for stage (id=1)
* *(1) Project (id#0L + 1) AS (id + 1)#3L
* +- *(1) Filter (id#0L = 1)
*    +- *(1) Range (0, 10, step=1, splits=4)
*/
/* 006 */ // codegenStageId=1
/* 007 */ final class GeneratedIteratorForCodegenStage1 extends org.apache.spark.sql.execution.BufferedRowIterator {

Issue Links

• links to: GitHub Pull Request #24133