HTML API: Fix possible infinite loop parsing incomplete script tags

[sirreal]

Trac ticket: https://core.trac.wordpress.org/ticket/61810

When the Tag Processor (or HTML Processor) attempts to parse certain incomplete script tags, the parser enters an infinite loop and will hang indefinitely. The conditions to reach this situation are:

• Input HTML ends with an open script tag.
• The final character of input is - or <.

If these conditions are satisfied, the parser will enter an infinite loop and hang when it attempts to parse the script tag.

Example problematic inputs:

• <script>-
• <script><

Creating a processor and calling next_tag() will hang. In both cases, next_tag() should return false with the processor in incomplete token state.

Diagnosis

This bug was caused by essential code for advancing the parser position ($at++) never being reached by a short-circuit condition: if ( $too_short || '<' !== $html[ $at++ ] ) {}. If the left-hand side is true, the RHS is never evaluated and the processor may not advance.

Other code in the block would advance the parser under most conditions, it was only when - or < were in the final position in the document that the LHS would evaluate to true and the parser would enter an infinite loop.

This was difficult to spot because of its similarity to many other blocks in the parser like this:

if (
  $at + 2 < $doc_length &&
  '-' === $html[ $at ] &&
  '-' === $html[ $at + 1 ] &&
  '>' === $html[ $at + 2 ]
) {}

---

This Pull Request is for code review only. Please keep all other discussion in the Trac ticket. Do not merge this Pull Request. See GitHub Pull Requests for Code Review in the Core Handbook for more details.

[github-actions]

Test using WordPress Playground

The changes in this pull request can previewed and tested using a WordPress Playground instance.

WordPress Playground is an experimental project that creates a full WordPress instance entirely within the browser.

Some things to be aware of

• The Plugin and Theme Directories cannot be accessed within Playground.
• All changes will be lost when closing a tab with a Playground instance.
• All changes will be lost when refreshing the page.
• A fresh instance is created each time the link below is clicked.
• Every time this pull request is updated, a new ZIP file containing all changes is created. If changes are not reflected in the Playground instance,
  it's possible that the most recent build failed, or has not completed. Check the list of workflow runs to be sure.

For more details about these limitations and more, check out the Limitations page in the WordPress Playground documentation.

Test this pull request with WordPress Playground.

[sirreal]

I've chosen to break here because there's no way to find the script closer. The document is incomplete, so breaking and returning false below enters the correct code path. This could also return false directly instead of breaking out of the loop.

This condition could likely be increased. We need to match at least <!-- so we could apply this with at least $at + 3.

Suggested change

	if ( $at + 1 >= $doc_length ) {
	// The parser needs to match at least `<!--` from here.
	if ( $at + 3 >= $doc_length ) {

---

I also think this could apply more aggressive length checks earlier, then skip all of the subsequent length checks. Ultimately, we need to find a script closer in order to exit true. It doesn't matter how much of the script contents are parsed if it can't be closed the parser will stop on incomplete input. After this line:

wordpress-develop/src/wp-includes/html-api/class-wp-html-tag-processor.php

Line 1416 in a799101

$at += strcspn( $html, '-<', $at );

We could include this to eagerly reach incomplete input if we know a closer cannot be found.

if ( $at + 8 > $doc_length ) { return false; }

Here's my thinking:

/* Things could happen in the loop, but they're irrelevant if a `</script>` cannot be found  */

// $at:  V
<script> ----->␃
// $at+8         ^

/* A script closer could fit here and is ultimately found */

// $at:          V
<script> bla bla </script>␃
// $at+8                 ^

/* A script closer would fit so processing continues */
// $at:          V
<script> bla bla -            more document…␃
// $at+8                 ^

Maybe the proposed bugfix could be landed and included in WordPress 6.6, with more optimizations included landed in another change to trunk.

@dmsnell curious to hear your thoughts.

[dmsnell]

these are valid points. what I like about advancing one character at a time is that in some cases, I leaned on that to ensure we didn't count wrong. a targeting fix sounds nice now, and then a follow-up can be tested in isolation, particularly for performance, for skipping ahead.

[dmsnell]

I think I also like returning false directly - no need to further confuse control flow

[sirreal]

I've started work on this in #9230

[github-actions]

The following accounts have interacted with this PR and/or linked issues. I will continue to update these lists as activity occurs. You can also manually ask me to refresh this list by adding the props-bot label.

Core Committers: Use this line as a base for the props when committing in SVN:

Props jonsurrell, dmsnell.

To understand the WordPress project's expectations around crediting contributors, please review the Contributor Attribution page in the Core Handbook.

[dmsnell]

Merged in [58845]
bdef9de