Name length error in Lexer_getName

[griffinmyers]

I'm loading up From ADV Part 1a from the SEC and encountering the following error when rendering Page 34:

Warning: name token is longer than allowed by the spec: 139

console.loging the name buffer that is throwing the error yields:

D,o,e,s, ,a,n,y, ,r,e,p,o,r,t, ,p,r,e,p,a,r,e,d, ,b,y, ,t,h,e, ,i,n,d,e,p,e,n,d,e,n,t, ,p,u,b,l,i,c, ,a,c,c,o,u,n,t,a,n,t, ,t,h,a,t, ,a,u,d,i,t,e,d, ,t,h,e, ,p,o,o,l,e,d, ,i,n,v,e,s,t,m,e,n,t, ,v,e,h,i,c,l,e, ,o,r, ,t,h,a,t, ,e,x,a,m,i,n,e,d, ,i,n,t,e,r,n,a,l, ,c,o,n,t,r,o,l,s

which I can see present on Page 34 of the pdf.

I'd like to solicit your opinion re: the severity of the error, if it's technically and issue with PDFJS or the PDF (Adobe Reader sluggishly can open this file and save it) and options for recovering from the error.

• PDFJS Version: 1.0.1130
• Browser: Chrome 43.0.2357.130 on Mac OSX 10.7.5 and PhantomJS 2.0 on Max OSX 10.7.5

[timvandermeij]

Adobe Reader/Acrobat does not complain about this PDF for me, so either they silently ignore the error (and the PDF is corrupted), or there is a bug in our parsing logic. We need to apply further triage here.

[griffinmyers]

👍 Thanks for the quick look @timvandermeij. Looking forward to seeing what we can come up with.

[timvandermeij]

Probably we need to change error to warn here:

pdf.js/src/core/parser.js

Line 842 in d105734

error('Warning: name token is longer than allowed by the spec: ' +

This change is motivated by the fact that it also says 'warning' in the message and I don't think it's a fatal error. I'm not sure why we say it's a warning and then use a error call. I would like to hear what others think about this.

Edit: Poppler seems to do something similar (https://github.com/danigm/poppler/blob/master/poppler/Lexer.cc#L434-L444), which is confusing me.

[Snuffleupagus]

According to the specification, names are limited to a length (in bytes) of 127; see http://www.adobe.com/content/dam/Adobe/en/devnet/acrobat/pdfs/PDF32000_2008.pdf#G16.1010218.

I'm wondering if we're perhaps failing to parse the file correctly, based on the message above, hence I'm wondering if the actual issue is somewhere else in the parser.js.
In any case, just removing the error might move a potential error elsewhere in the code, so I think we should be careful with that kind of "solution".

[timvandermeij]

Good find, and you're right. I was confused because the error has 'warning' in the message.

[Rob--W]

@Snuffleupagus
FYI: Our parser is correct, the PDF is bad. I have identified the ID of the faulty object, extracted it from the PDF using qpdf (qpdf page34.pdf --show-object=5144,0), and then formatted the output. Clearly, the name is too long.

5142 0 obj
<<
  /AP <<
    /D <<
      /Does#20any#20report#20prepared#20by#20the#20independent#20public#20accountant#20that#20audited#20the#20pooled#20investment#20vehicle#20or#20that#20examined#20internal#20controls 5144 0 R
      /Off 5145 0 R
    >>
    /N <<
      /Does#20any#20report#20prepared#20by#20the#20independent#20public#20accountant#20that#20audited#20the#20pooled#20investment#20vehicle#20or#20that#20examined#20internal#20controls 5143 0 R
    >>
  >>
  /AS /Off
  /BS <<
    /W 0
  >>
  /DA (/ZaDb 0 Tf 0 g)
  /F 4
  /FT /Btn
  /Ff 49152
  /MK <<
    /CA (n)
  >>
  /P 4 0 R
  /Rect [ 378.84000 579.96000 387.60000 588.72000 ]
  /Subtype /Widget
  /T (Does any report prepared by the independent public accountant that audited the pooled investment vehicle or that examined internal controls contain an unqualified opinion? yes check box)
  /TU (Does any report prepared by the independent public accountant that audited the pooled investment vehicle or that examined internal controls contain an unqualified opinion?yes check box)
  /Type /Annot
>>
endobj