Problem with UTF8 charset

[kukihn]

After update from WP 4.1.4 to 4.1.2 at PHP 5.5.0. I got this error:

Warning: preg_replace(): Compilation failed: repeated subpattern is too long at offset 592 in /home/www/xxxxxx/html/www.xxxxxx.com/blog/wp-includes/wp-db.php on line 2640

I found a workaround here:
​http://forum.wpde.org/installation/142182-nach-update-auf-4-2-1-keine-umlaute-mehr-speicherbar-2.html#post614130

After disabling this line in wp-db.php it is working again.

// $value['value'] = preg_replace( $regex, '$1', $value['value'] );
continue;
}

Kind regards,
Marcus

[pento]

Thank you for the bug report, @kukihn!

You'll notice just above that line, there's a bit of code reading {1,50}.

Could you please try reducing the 50, and seeing what happens? What's the largest value it can be before you get warnings?

[kukihn]

Hi,
I changed this from 1,50 to 1,40 and it seems to work on my blog.

[pento]

In 32375:

The UTF-8 regex can occasionally fail on very low memory machines. Reduce the amount of memory it uses.

See #32204.

[pento]

In 32376:

The UTF-8 regex can occasionally fail on very low memory machines. Reduce the amount of memory it uses.

Merges [32375] to the 4.2 branch.

See #32204.

[pento]

The wp-db.php change needs to be backported to 3.7-4.1, after #32165 is done.

[helen]

In 32396:

The UTF-8 regex can occasionally fail on very low memory machines. Reduce the amount of memory it uses.

Merges [32375] to the 4.1 branch.

props pento.
See #32204.

[helen]

In 32397:

The UTF-8 regex can occasionally fail on very low memory machines. Reduce the amount of memory it uses.

Merges [32375] to the 4.0 branch.

props pento.
See #32204.

[helen]

In 32398:

The UTF-8 regex can occasionally fail on very low memory machines. Reduce the amount of memory it uses.

Merges [32375] to the 3.9 branch.

props pento.
See #32204.

[helen]

In 32399:

The UTF-8 regex can occasionally fail on very low memory machines. Reduce the amount of memory it uses.

Merges [32375] to the 3.8 branch.

props pento.
See #32204.

[helen]

In 32400:

The UTF-8 regex can occasionally fail on very low memory machines. Reduce the amount of memory it uses.

Merges [32375] to the 3.7 branch.

props pento.
fixes #32204.