Highlight annotation text doesn't cover the selection for some files

Which product are you using?
PDF.js Express Plus 8.4.0

Detailed description of issue
Hi there!

For some PDF files higlight annotations are not working correctly. Selected text (as you can see in comments below on the print screens) is different than that we can see in the selected area on PDF. Besides, selection process doesn’t run so smoothly for such files. Sometimes it isn’t possible to select the whole word, it is selected with space (before or after) or it misses the first letter. The behaviour is different for pdftron, where I have never noticed any situation like that. PDF files attached below were created from text files, one of them on Linux, the second one on Mac. It is simple to reproduce on demo pages, print screens show the differences:

The Linux file is in the first row (please, notice: the further from the beginning of the line, the longer the shift is) and the Mac file is in the second row (behaviour is a little bit different here, but it’s also unexpected):

Iceland.pdf (42.5 KB)
lorem_mac.pdf (45.5 KB)

Does your issue happen with every document, or just one?
Not every document, but it isn’t a single case.

Is it possible to fix that?

Thanks,
Dorota

1 Like

Hi Dorota,

Are these files word files that are converted to PDF?

Best Regards,
Zach Serviss
Web Development Support Engineer
PDFTron Systems, Inc.


Share how you are using PDF.js Express in your organization you could win a $500 Amazon gift card. All participants will receive 6 months of PDF.js Express+ for free. Learn more here

Hi Zach,

The Linux file was exported from Libre Office Writer (7.0, but for the version 7.4 is the same).
The Mac file was created with Pages.

Thanks,
Dorota

Hi Dorota,

Thank you for the clarification.

We are seeing this problem with word document conversions and currently have this issue in our backlog. I will add this information to our ticket and will let you know when a fix is available for testing.

Best Regards,
Zach Serviss
Web Development Support Engineer
PDFTron Systems, Inc.


Share how you are using PDF.js Express in your organization you could win a $500 Amazon gift card. All participants will receive 6 months of PDF.js Express+ for free. Learn more here

2 Likes

Hi Zach,

when you are planning to release a fix for this?

Thank you.

1 Like

Hi there,

After conducting an internal review, this problem seems to stem from PDFJS. Unfortunately we are unable to fix this issue in PDFJS as its maintained by Mozilla.

Best Regards,
Zach Serviss
Web Development Support Engineer
PDFTron Systems, Inc.


Share how you are using PDF.js Express in your organization you could win a $500 Amazon gift card. All participants will receive 6 months of PDF.js Express+ for free. Learn more here

Hi there,

I’ve just found an issue that seems to be related to our problem - if I’m wrong, I apologize for the confusion. Nevertheless, this issue remains important to us. The related issue can be found [here](https://github.com/mozilla/pdf.js/issues/13083), along with the last comment posted by one of the repository’s contributors. Is it possible that the newer version of pdf.js has a solution implemented? Could you integrate it into your library and check it out? I would be grateful for any help on this matter.

Best regards,
Dorota

Hello dorota.k,

Thank you for pointing this pdf.js issue out, we are looking into updating the pdf.js version, however, since we are only providing product-breaking bug fixes at the moment this may be put at a lower priority.

Thank you,
Tyler

Hi there,

Any updates? One and a half year has already passed from my first note. I’ve found quite similar topic here - I’m not alone in it. This is a very serious problem in our application and probably, if you fix it, the current annotations on pdf won’t cover correct positions - technical debt is growing. Of course I am not sure if upgrading psf.js library may help, but we won’t know if we don’t check.

Best Regards,
Dorota

Am I the only one who feels that the authors of the solution have lost interest in whether it works well?

This forum is used by people who work with software, building, developing and maintaining it. We all realize that it is impossible to do everything at once. You have to queue topics, prioritize them, but…

When such a serious problem is not addressed in any way for a year and a half then one slowly loses faith that it will ever be taken care of.

The license is not one of the cheapest, $440 per month. For a year and a half that’s almost $8,000 in licensing costs for a tool that has serious shortcomings, and we don’t even get a glimpse from the authors that it will be addressed in some time, no hope at all.

I have the impression that every major problem is thrown into the backlog… The legendary black hole.

I’m sorry for the emotional uplift, but your hands fall when you want to do something right, but the partners you’ve chosen seriously under-deliver.

1 Like

Absolutely, looks like this product is not supported any more. No decent documentation, no fixes etc.