#25707 [Ver->Asn]: html_entity_decode over-decodes <

ID: 25707 Updated by: moriyoshi@php.net Reported By: Bjorn dot Victor at it dot uu dot se -Status: Verified +Status: Assigned -Bug Type: Feature/Change Request +Bug Type: Strings related Operating System: Solaris 8 PHP Version: 4.3.3 -Assigned To: +Assigned To: moriyoshi Previous Comments: ------------------------------------------------------------------------ [2003-10-01 17:31:22] elmicha@php.net html_entity_decode(htmlentities("&lt;")) returns "<", but IMHO it should return the original "&lt;". The unhtmlentities() function given on http://www.php.net/html_entity_decode works like it should (in my eyes). ------------------------------------------------------------------------ [2003-10-01 03:31:42] Bjorn dot Victor at it dot uu dot se Sorry, this is not an RTFM error, and has nothing to do with the optional parameters of the function. I have changed the summary to refer to "lt", to avoid confusion with ENT_QUOTES etc - believe me, I tried this before looking at the source and figuring out what the error really was. The current code works like this: iterate over the 6 "basic_entities", replace the entity with its character in the string. "&amp;" is the first item in basic_entities, which is good when you're doing htmlentities (the reverse operation). Given a string "&amp;lt;", it will first become "&lt;", and then (because "&lt;" is handled after "&amp;"), "<". Consider doing "&amp;" last, e.g. by traversing basic_entities backwards: "&amp;lt;" becomes "&lt;", which is the expected. ------------------------------------------------------------------------ [2003-09-30 15:00:59] sniper@php.net RTFM: http://www.php.net/html_entity_decode (the 2nd optional parameter..) ------------------------------------------------------------------------ [2003-09-30 14:52:20] Bjorn dot Victor at it dot uu dot se Description: ------------ Symptom: html_entity_decode("&amp;quot;") returns '"', while the expected value would be "&quot;". Corresponding (wrong) behaviour for &amp; followed by "lt;", "gt;" etc. Another example is html_entity_decode(htmlentities("&lt;")) which returns "<" rather than "&lt;" as expected. As a result, html_entity_decode can not be used as the inverse of htmlentities. Diagnosis: The function (php_unescape_html_entities in ext/standard/html.c) replaces each entity in basic_entities with its corresponding character, but starts by replacing "&amp;" with "&", the resulting string being "&quot;", which is then replaced by '"'. Solution: php_unescape_html_entities in ext/standard/html.c traverses the basic_entities from the wrong end; it must replace "&amp;" *last*, not *first*. Reproduce code: --------------- print html_entity_decode("&amp;quot;&amp;lt;&amp;gt;"); Expected result: ---------------- &quot;&lt;&gt; Actual result: -------------- "<> ------------------------------------------------------------------------ -- Edit this bug report at http://bugs.php.net/?id=25707&edit=1