Exception not handling on "Allowed memory size exhausted"

[bastienuh]

Hi,

I have the following exception :

Symfony\Component\ErrorHandler\Error\FatalError

  Allowed memory size of 536870912 bytes exhausted (tried to allocate 167772160 bytes)

  at vendor/smalot/pdfparser/src/Smalot/PdfParser/Font.php:221
    217▕                         $char_to = hexdec($matches['to'][$key]);
    218▕                         $offset = hexdec($matches['offset'][$key]);
    219▕
    220▕                         for ($char = $char_from; $char <= $char_to; ++$char) {
  ➜ 221▕                             $this->table[$char] = self::uchr($char - $char_from + $offset);
    222▕                         }
    223▕                     }
    224▕
    225▕                     // Support for : <srcCode1> <srcCodeN> [<dstString1> <dstString2> ... <dstStringN>]

I tried to handle it by a try / catch, but it's not working:

try {
    echo 'A';
    $pdf = $pdfParser->parseFile($pathname.'.pdf');
}
catch (\Whoops\Exception\ErrorException $e) {
    echo ' - Error in $pdf->parseFile() : '.$e->getMessage().' - ';
    return $new_document;
}
catch (\Symfony\Component\ErrorHandler\Error\FatalError $e) {
    echo ' - Error in $pdf->parseFile() : '.$e->getMessage().' - ';
    return $new_document;
}
catch (\Exception $e) {
    echo ' - Error in $pdf->parseFile() : '.$e->getMessage().' - ';
    return $new_document;
}
echo 'B';

The error is in parseFile() because the "A" is echoing, but not the "B".

Do you know why the exception is not handling?
Maybe it can be a solution to add a try / catch around the self::uchr ?

Thanks for your help and advice :-)

(If it could help, the parsed PDF is here : https://www.assemblee-nationale.fr/dyn/opendata/PIONANR5L15TAP0528.pdf)

[j0k3r]

What PHP version are you using? Which version of PDFParser are you using?

[bastienuh]

I'm running PHP 8.0.12 and using PDFParser 2.1.0 :-)

[dsuurlant]

I ran into a similar problem which might have the same cause.

Debugging the code I found that the $offset had a really strange number:

char: 65018, char_from: 65012, offset: 2.1254268986858E+96

Which led to the TypeError for self::uchr.

It's possible to add a check so that if the result of $char - $char_from + $offset is not an integer type, it is skipped. (Trying to cast 2.1254268986858E+96 to an integer definitely doesn't work.)

But I'm unsure what effect doing that will have on the parsing of the PDF.

However I was able to catch the error using Error (PHP native, not Symfony). Ideally though the PDF would be parsed without errors.

[GreyWyvern]

This is probably resolved by #623. The sample PDF is still accessible, and it's working in my copy.

[k00ni]

@bastienuh please try again and get back to us. Thanks

[k00ni]

Probably solved. If not, comment here.

[UnnitMetaliya]

Probably solved. If not, comment here.

@k00ni I think it's still an issue.

I ran the command with 3GB memory

php -d memory_limit=3G artisan pdf:process

I am being thrown this:

I am on "smalot/pdfparser": "^2.11"

Fatal error: Allowed memory size of 3221225472 bytes exhausted (tried to allocate 1342177280 bytes) in ../vendor/smalot/pdfparser/src/Smalot/PdfParser/Font.php on line 150

Symfony\Component\ErrorHandler\Error\FatalError

Allowed memory size of 3221225472 bytes exhausted (tried to allocate 1342177280 bytes)

at vendor/smalot/pdfparser/src/Smalot/PdfParser/Font.php:150
146▕
147▕ if (!isset(self::$uchrCache[$code])) {
148▕ // html_entity_decode() will not work with UTF-16 or UTF-32 char entities,
149▕ // therefore, we use mb_convert_encoding() instead
➜ 150▕ self::$uchrCache[$code] = mb_convert_encoding("&#{$code};", 'UTF-8', 'HTML-ENTITIES');
151▕ }
152▕
153▕ return self::$uchrCache[$code];
154▕ }

Whoops\Exception\ErrorException

Allowed memory size of 3221225472 bytes exhausted (tried to allocate 1342177280 bytes)

at vendor/smalot/pdfparser/src/Smalot/PdfParser/Font.php:150
146▕
147▕ if (!isset(self::$uchrCache[$code])) {
148▕ // html_entity_decode() will not work with UTF-16 or UTF-32 char entities,
149▕ // therefore, we use mb_convert_encoding() instead
➜ 150▕ self::$uchrCache[$code] = mb_convert_encoding("&#{$code};", 'UTF-8', 'HTML-ENTITIES');
151▕ }
152▕
153▕ return self::$uchrCache[$code];
154▕ }

  +1 vendor frames

2 [internal]:0
Whoops\Run::handleShutdown()`

I opened this issue -- with more details.