Uncaught TypeError: str_repeat() expects parameter 2 to be integery, null given

[ibennetch]

Highlighting a query causes an error about str_repeat with current master.

Fatal error: Uncaught TypeError: str_repeat() expects parameter 2 to be integer, null given in /Users/isaacbennetch/Documents/svn/pma/sql-parser/src/Utils/Formatter.php:497
Stack trace:
#0 /Users/isaacbennetch/Documents/svn/pma/sql-parser/src/Utils/Formatter.php(497): str_repeat('    ', NULL)
#1 /Users/isaacbennetch/Documents/svn/pma/sql-parser/src/Utils/Formatter.php(671): PhpMyAdmin\SqlParser\Utils\Formatter->formatList(Object(PhpMyAdmin\SqlParser\TokensList))
#2 /Users/isaacbennetch/Documents/svn/pma/sql-parser/src/Utils/CLI.php(89): PhpMyAdmin\SqlParser\Utils\Formatter::format('select\ncase whe...', Array)
#3 /Users/isaacbennetch/Documents/svn/pma/sql-parser/bin/highlight-query(30): PhpMyAdmin\SqlParser\Utils\CLI->runHighlight()
#4 {main}
  thrown in /Users/isaacbennetch/Documents/svn/pma/sql-parser/src/Utils/Formatter.php on line 497

This is the content of case.sql

case when count(id)>0
  then 'Yes'
  else 'No'
end
from products where 1;

This works correctly with the QA branch, but git bisect doesn't help me. I've tested with both PHP 7.1.23 and PHP 7.4.0.

[MauricioFauth]

The commit 1a99634 fixes the uncaught TypeError, but I don't think it is a proper fix.

This issue started to happen because we enabled the PHP strict mode. Before null was being implicit converted to int.
Looks like sql-parser only expects an END after a BEGIN. It doesn't handle the CASE statement well enough.

array_pop() returns null if $blocksIndentation is empty or is not an array.

sql-parser/src/Utils/Formatter.php

Lines 441 to 449 in 1a99634

	// Indenting BEGIN ... END blocks.
	if ($prev->type === Token::TYPE_KEYWORD && $prev->keyword === 'BEGIN') {
	$lineEnded = true;
	$blocksIndentation[] = $indent;
	++$indent;
	} elseif ($curr->type === Token::TYPE_KEYWORD && $curr->keyword === 'END') {
	$lineEnded = true;
	$indent = array_pop($blocksIndentation);
	}

I think we should close this issue and open a new one related to the CASE statement issue. What do you think @ibennetch ?

[ibennetch]

I tend to agree; it seems that your commit suppresses the actual error but we should still make a new issue to address the underlying shortcomings with CASE statements.

Good research and conclusion. @MauricioFauth do you mind creating that issue?

[MauricioFauth]

Thank you. No problem, I can open the issue.

[williamdes]

what about QA branch?

[MauricioFauth]

This TypeError only affects master.

[williamdes]

Okay, because of the php version?

[MauricioFauth]

No, because of declare(strict_types=1);.