Black Rectangular boxes seen in pages instead of actual content

[SushmaRaghu]

Attach (recommended) or Link to PDF file here: Sample_PDF1_page1.pdf

Configuration:

• Web browser and its version: Chrome, Firefox, MS Edge (latest versions)
• Operating system and its version: Windows, Mac
• PDF.js version: latest released
• Is a browser extension: NA

Steps to reproduce the problem:

1. Open the attached sample document using pdf.js viewer.
2. Instead of actual content , we can see black rectangular boxes

What is the expected behavior? (add screenshot)
Actual content has to be seen instead of black boxes

Link to a viewer (if hosted on a site other than mozilla.github.io/pdf.js or as Firefox/Chrome extension):

[Snuffleupagus]

• Web browser and its version: Chrome, Firefox, MS Edge (latest versions)
• Operating system and its version: Windows, Mac
• PDF.js version: latest released
• Is a browser extension: NA

For future reference: Please remember to always provide actual version numbers, since "latest" isn't really helpful in general.

[SushmaRaghu]

@Snuffleupagus
Can I please know if there is any update on this ?

[THausherr]

Your PDF is broken, several "stop reading corrupt stream due to a DataFormatException" errors. I get a terrible output on edge and on pdfbox.

[SushmaRaghu]

Hi ,

The document works fine in Adobe reader and also native Chrome Viewer. PFA screenshots of the same and the one in Firefox(pdf.js)

The expectation is the output similar to Chrome.

[SushmaRaghu]

Any update on this ?

[calixteman]

The object 46 0 is compressed stream and it contains some errors, qpdf outputs:

PDF Version: 1.7
File is not encrypted
File is not linearized
WARNING: Sample_PDF1_page1.pdf (offset 15631): error decoding stream data for object 46 0: stream inflate: inflate: data: invalid distance too far back
page 1: content stream (content stream object 46 0): errors while decoding content stream

If I return an empty stream for this obj at

pdf.js/src/core/xref.js

Line 786 in b18a166

:

    if (num === 46) {
      return new StringStream("");
    }

the rendering is "correct".

[SushmaRaghu]

@calixteman

Can you please elaborate on the above comment ? Are you suggesting the file is correct ?

[SushmaRaghu]

Hi,

Is there any update on the progress of this issue ?

Thanks
Sushma Raghupathi

[calixteman]

@calixteman

Can you please elaborate on the above comment ? Are you suggesting the file is correct ?

The file isn't correct: the compressed stream has some errors.

Is there any update on the progress of this issue ?

No there isn't.

[SushmaRaghu]

Hi @calixteman

The same document renders fine with pdf.js version as 2.8.92. With version 3.8 is where I see the issue.
Please share your thoughts on the same.

[SushmaRaghu]

Is there any update on this ? As mentioned , this was working in earlier versions.

[calixteman]

I just tried with Firefox nightly 87.0a1 (https://hg.mozilla.org/mozilla-central/rev/e0774d6f87cc) and the rendering was wrong on Windows 11.
@SushmaRaghu could you git bisect to try to figure out which patch could have cause this issue ?
As said previously, your pdf is bad because it contains some streams which have compressed with errors.
It works in some viewers because they just manage to ignore the objects which are associated with these erroneous streams.

[SushmaRaghu]

@calixteman
It is working as expected till v2.3.200. It has started to fail from v2.4.456.
I suspect below 3 commits:
https://github.com/mozilla/pdf.js/pull/11335/files
https://github.com/mozilla/pdf.js/pull/11301/files
https://github.com/mozilla/pdf.js/pull/11334/files

Can you share your thoughts?

Thanks
Sushma Raghupathi

[SushmaRaghu]

Hi team..

Is there any update on this ?

Thanks
Sushma Raghupathi