Bug #46745 [Com]: Multipart/form-data field names truncated at semicolon in certain cases

Edit report at https://bugs.php.net/bug.php?id=46745&edit=1 ID: 46745 Comment by: sji at sj-i dot dev Reported by: david at sickmiller dot com Summary: Multipart/form-data field names truncated at semicolon in certain cases Status: Open Type: Bug Package: *General Issues Operating System: Linux PHP Version: 5.2.6 Block user comment: N Private report: N New Comment: Couldn't reproduce this on PHP 7.4. The result on PHP 7.4 ------------------------- Array ( [before_;_after] => [%22_before_;_after] => ['_before_;_after] => ['_'_before_;_after] => ['_%22_before_;_after] => [%22_'_before_;_after] => [%22_%22_before_;_after] => [%22_'_'_before_;_after] => [%22_'_%22_before_;_after] => [%22_%22_'_before_;_after] => [%22_%22_%22_before_;_after] => ['_'_'_before_;_after] => ['_'_%22_before_;_after] => ['_%22_'_before_;_after] => ['_%22_%22_before_;_after] => ) Previous Comments: ------------------------------------------------------------------------ [2018-03-11 23:56:09] cmb@php.net Regardless of whether this is a niche case: the described behavior would be a bug. ------------------------------------------------------------------------ [2008-12-09 03:49:45] david at sickmiller dot com For a web application, I have written a PHP file that is a general form handler. Various different forms all post to this same file, which does a foreach through $_POST and essentially produces a table with each field question and answer. Some of the forms include file uploads, so I have to use multipart/form-data. The forms are generated dynamically in a moderately expensive operation and result in HTML like this: <form> What's your first name? <input name="What's your name?"> </form> The questions actually come from different systems, so I don't have the flexibility to change them to remove the non-alphanumeric characters. And it shouldn't be necessary according to the HTML spec. It's generally been working, except for the peculiar combinations of apostophes, quotation marks, and semicolons that I inadvertently ran into recently and have documented in this bug. I checked the header, and it looks like Firefox is correctly submitting the data. I realize this is a niche case. How difficult does the fix look? ------------------------------------------------------------------------ [2008-12-03 21:54:47] david at sickmiller dot com Description: ------------ When a form is submitted using multipart-/form-data enctype, PHP will truncate field names if they contain semicolons and certain other characters. It seems to happen when there is an uneven number of quotation marks that precede the semicolon. Reproduce code: --------------- <html><form action="semicolon.php" method="POST" enctype="multipart/form-data"> <input type="hidden" name="before ; after"> <input type="hidden" name="&quot; before ; after"> <input type="hidden" name="' before ; after"> <input type="hidden" name="' ' before ; after"> <input type="hidden" name="' &quot; before ; after"> <input type="hidden" name="&quot; ' before ; after"> <input type="hidden" name="&quot; &quot; before ; after"> <input type="hidden" name="&quot; ' ' before ; after"> <input type="hidden" name="&quot; ' &quot; before ; after"> <input type="hidden" name="&quot; &quot; ' before ; after"> <input type="hidden" name="&quot; &quot; &quot; before ; after"> <input type="hidden" name="' ' ' before ; after"> <input type="hidden" name="' ' &quot; before ; after"> <input type="hidden" name="' &quot; ' before ; after"> <input type="hidden" name="' &quot; &quot; before ; after"> <input type="submit"> </form><pre><?php print_r($_POST); ?></pre> </html> Expected result: ---------------- Array ( [before_;_after] => ["_before_;_after] => ['_before_;_after] => ['_'_before_;_after] => ['_"_before_;_after] => ["_'_before_;_after] => ["_"_before_;_after] => ["_'_'_before_;_after] => ["_'_"_before_;_after] => ["_"_'_before_;_after] => ["_"_"_before_;_after] => ['_'_'_before_;_after] => ['_'_"_before_;_after] => ['_"_'_before_;_after] => ['_"_"_before_;_after] => ) Actual result: -------------- Array ( [before_;_after] => ["_before_] => ['_before_;_after] => ['_'_before_;_after] => ['_"_before_] => ["_'_before_;_after] => ["_"_before_;_after] => ["_'_'_before_] => ["_'_"_before_;_after] => ["_"_'_before_;_after] => ["_"_"_before_] => ['_'_'_before_;_after] => ['_'_"_before_] => ['_"_'_before_;_after] => ['_"_"_before_;_after] => ) ------------------------------------------------------------------------ -- Edit this bug report at https://bugs.php.net/bug.php?id=46745&edit=1