After updating to 2.2.3, my site experienced the same issue:
An inline JavaScript stopped working because backslashes have been removed.
Comparing version 2.2.2 and 2.2.3 showed that it happens in
"/src/class-ss-url-extractor.php"
where in two functions a new stripslashes() has been inserted in 2.2.3
function extract_and_replace_urls_in_html()
$html_string = $this->get_body();
$html_string = stripslashes( $html_string ); // 2.2.3
function extract_and_replace_urls_in_script( $text )
$text = stripslashes( $text ); // 2.2.3
Because parameters $html_string and $text contained following (truncated) JS code, in the generated static pages, this JS code failed.
... "<span class=\"ast-icon icon-close\" ><\/span >" ...
By removing these stripslashes()calls, Simply Static 2.2.3 again worked fine as version 2.2.2.
But what about replacing URLs with rare/unusual encoding (in JS/CSS) ?
Can we avoid stripslashes()calls and still replace all kind of encoded URLs?
One possibility would be, by using arrays containing all (most) possible encoded URL pairs for the replacement:
// prepare encoded URL pairs for replacement:
$dyn[0..n] = DYNAMIC_URLs; // encoded source URLs
$stat[0..n] = STATIC_URLs; // corresponding replacements
str_replace( $string, $dyn, $stat );
It would also help to provide extra filters, that allow replacing special cases of encoded URLs by using the provided URL strings to perform its own str_replace() to cover special cases.
For example:
apply_filters( 'ss_fix_urls', $text, $url_dyn, $url_static)
Last not least:
Thank you for providing Simply Static, it works great!
/GeHo
