php.net |  support |  documentation |  report a bug |  advanced search |  search howto |  statistics |  random bug |  login

go to bug id or search bugs for

Doc Bug #64739 Invalid Title and Author data returned
Submitted: 2013-04-30 07:06 UTC Modified: 2017-07-22 21:08 UTC
Votes:1
Avg. Score:3.0 ± 0.0
Reproduced:1 of 1 (100.0%)
Same Version:0 (0.0%)
Same OS:1 (100.0%)
From: php-qa at sebastianmendel dot de Assigned: kalle (profile)
Status: Closed Package: EXIF related
PHP Version: 5.4.14 OS: Linux
Private report: No CVE-ID: None
View Developer Edit
Welcome back! If you're the original bug submitter, here's where you can edit the bug or add additional notes.
If you forgot your password, you can retrieve your password here.
Password:
Status:
Package:
Bug Type:
Summary:
From: php-qa at sebastianmendel dot de
New email:
PHP Version: OS:

 

 [2013-04-30 07:06 UTC] php-qa at sebastianmendel dot de 
Description:
------------
exif_read_data returns invalid Titel and Author data.

ExifTool cli returns valid data.

Test script:
---------------
php-5.4.14 -r "var_dump(exif_read_data('http://sebastianmendel.de/php/exif_segmentation_fault/test.jpg'));"


Expected result:
----------------
  ....
  ["Title"]=>
  string(8) "55845364"
  ["Author"]=>
  string(13) "100420.000000"
  ....

Actual result:
--------------
  ....
  ["Title"]=>
  string(8) "????????"
  ["Author"]=>
  string(13) "?????????????"
  ....

Patches

Pull Requests

History

AllCommentsChangesGit/SVN commitsRelated reports
 [2013-04-30 15:35 UTC] cweiske@php.net 
Reason is probably that the exif data were stored on a PowerPC Mac and have big endianess.
 [2016-08-05 06:29 UTC] kalle@php.net 
@cweiske, you are right. This picture linked is stored in Motorola byte order which qualifies as big endian. This image has no makernote tag so I guess it was made from an editor (Paint.NET) and it also correctly identifies it as a motorola byte order in the COMPUTED section.

Of all the data returned, it indeed seems to only be the title and author that is somehow corrupted.
 [2017-07-07 10:28 UTC] kalle@php.net
-Status: Open +Status: Feedback
 [2017-07-07 10:28 UTC] kalle@php.net 
Hi, could you please re-upload the image somewhere so I can use it for testing?

Thanks
 [2017-07-13 12:35 UTC] php-qa at sebastianmendel dot de
-Status: Feedback +Status: Open
 [2017-07-13 12:35 UTC] php-qa at sebastianmendel dot de 
link http://sebastianmendel.de/php/exif_segmentation_fault/test.jpg restored
 [2017-07-13 23:15 UTC] kalle@php.net 
Thank you Sebastian, from some quick debugging it seems that the Motorola/Intel offset is poorly detected, causing it to read it as such
 [2017-07-13 23:15 UTC] kalle@php.net
-Status: Open +Status: Assigned -Assigned To: +Assigned To: kalle
 [2017-07-14 01:37 UTC] kalle@php.net 
I debugged it a little further, it seems that internally ext/exif detects the image byte order as Motorola, which causes it to use a wrong encoding to process the tags (from the WINXP section).

A temporary hack can be done by modifying the exif.decode_unicode_motorola ini directive:

[exif]
exif.decode_unicode_motorola=UCS-2LE

(It defaults to UCS-2BE)

I'm gonna look into possible approaches for solving this
 [2017-07-21 19:41 UTC] kalle@php.net
-Status: Assigned +Status: Closed
 [2017-07-21 19:41 UTC] kalle@php.net 
I have given this some thought and attempts and trying to solve this in an automatic behavior, but its almost impossible with the current implementation of EXIF in ext/exif. Therefore I recommend to change the ini setting if the image contains another byte order than that of the default charset used to decode it with.

TL;DR for your image you would therefore do:
exif.decode_unicode_motorola=UCS-2LE

Apologies again, but I don't see any other clear way to go about this
 [2017-07-22 13:51 UTC] php-qa at sebastianmendel dot de 
The problem is you never know which encoding is used and if you process a lot of pictures which are always changing - in a CMS - you will always loose.

Ok, so at least the manual page should have a hint about this.
 [2017-07-22 14:52 UTC] kalle@php.net
-Status: Closed +Status: Assigned -Type: Bug +Type: Documentation Problem
 [2017-07-22 14:52 UTC] kalle@php.net 
I totally agree with you on this, possibly having to iterate charsets just to make it decode is non sense to me as well.

I will update the docs accordingly

Side note: I'm working on a new extension to see if it can replace the exif one at one point if you are interested, keep in mind its very basic: http://github.com/KalleZ/pecl-exifkit
 [2017-07-22 21:07 UTC] kalle@php.net 
I have updated the docs for this
 [2017-07-22 21:08 UTC] kalle@php.net
-Status: Assigned +Status: Closed
 
PHP Copyright © 2001-2024 The PHP Group
All rights reserved. 		Last updated: Sun Dec 22 03:01:28 2024 UTC