Christian Vigh - 2016-10-30 10:43:55 - In reply to message 1 from Ammar Qasem
Thanks for your feedback.
Regarding the backtrack_limit error, you will have to change this setting in your php.ini file (pcre.backtrack_limit). You may also want to change the pcre.recursion_limit one.
On Unix systems, those two settings are usually ok for most PDF files. On Windows systems, they are surprisingly lower, so you may want to change them.
Arabic language (and, more generally, languages written from right-to-left) are not yet correctly rendered. In fact, they are rendered from left-to-right, which also changes the ligatures between two letters in Arabic. I'm still investigating this issue, and the feature will be available in a future release.
Regarding PDF files having lots of images, I have to my knowledge no issue reported about that. I suspect this could need setting memory_limit in PHP.INI to a higher value. If you like, you can send me a sample file at the following address, so that I can test what happens :
Christian Vigh - 2016-10-30 12:39:50 - In reply to message 3 from Ammar Qasem
Thanks for sending me the files.
I have fixed two issues with new version 1.3.1 :
a) Author data was not correctly recognized in some cases (this is the case of the two PDF files you sent to me)
b) Numerous warning messages were issued for an undefined index 'page' (file The_Way_To_GO.pdf)
You can download the new version on this site ; please let me know if you have any other problem.
Regarding the "backtrack_limit" error, you will have no choice ; you must set the "pcre.backtrack_limit" setting of your php.ini file to at least 2000000 (I tried with this value and all went fine).
A little warning about the second file ("The_Way_To_GO.pdf") : once loaded, it occupies 126MB, so make sure that your memory_limit setting of your php.ini file is greater than that. In the next version of the class, I will try to reduce the memory footprint.