Bug #74230 [Opn->Csd]: iconv fails to fail on surrogates

Edit report at https://bugs.php.net/bug.php?id=74230&edit=1

 ID:                 74230
 Updated by:         [email protected]
 Reported by:        paul dot crovella at gmail dot com
 Summary:            iconv fails to fail on surrogates
-Status:             Open
+Status:             Closed
 Type:               Bug
 Package:            ICONV related
 Operating System:   Windows
 PHP Version:        7.1.2
-Assigned To:        
+Assigned To:        ab
 Block user comment: N
 Private report:     N

 New Comment:

The deps was updated with libiconv-1.15 for all 7.x branches. Currently you can fetch any latest master snap for the tests and OFC teh upcoming RCs this week.

Thanks.


Previous Comments:
------------------------------------------------------------------------
[2017-03-13 20:29:57] paul dot crovella at gmail dot com

Wow, and their very first bullet point for that release is:

 - The UTF-8 converter now rejects surrogates and out-of-range code points.

http://savannah.gnu.org/forum/forum.php?forum_id=8790

Talk about timing :D

------------------------------------------------------------------------
[2017-03-13 12:41:26] [email protected]

Thanks for the report. This seems to be the libiconv issue. The library used is quite old, but surprisingly libiconv-1.15 was released a month ago after 6 years of silence :) I'm goint to check with it. Otherwise, in 7.1 you might get an acceptable result with sapi_windows_cp_conv().

Thanks.

------------------------------------------------------------------------
[2017-03-09 10:48:45] paul dot crovella at gmail dot com

Description:
------------
Surrogate codepoints cannot be validly encoded in UTF-8. Iconv should return false when given a byte sequence that would encode them and an in_charset of 'UTF-8'.

This works as expected on Linux https://3v4l.org/Ff6cr but doesn't on Windows (tested with PHP 7.0.16 and 7.1.2.)

Test script:
---------------
$high = "\xED\xa1\x92"; // codepoint D852
$low = "\xED\xBD\xA2"; // codepoint DF62
$pair = $high.$low;
var_dump(
    @\iconv('UTF-8', 'UTF-8', $high) === false,
    @\iconv('UTF-8', 'UTF-8', $low) === false,
    @\iconv('UTF-8', 'UTF-8', $pair) === false
);

Expected result:
----------------
bool(true)
bool(true)
bool(true)

Actual result:
--------------
bool(false)
bool(false)
bool(false)


------------------------------------------------------------------------



--
Edit this bug report at https://bugs.php.net/bug.php?id=74230&edit=1