Internal Links are not working on merged PDF

[aarkue]

When merging multiple PDFs that contain references or links to different PDF locations, those no longer work in the merged PDF.

We already explore and collect all named destinations. We might somehow need to re-add them back to the merged PDF file.
Moreover, we will need to investigate further how references/links are implemented in PDFs.

[stefnotch]

Could this be related to Hopding/pdf-lib#341 ? I did notice that you're using a better maintained fork of pdf-lib

[aarkue]

Yeah, the issue you linked is the relevant upstream issue 👍. A simple API that allows manipulating Named Destinations in PDFs will likely not come soon to https://github.com/Hopding/pdf-lib or https://github.com/cantoo-scribe/pdf-lib.

I think the best way forward would be to implement this functionality ourselves, based on the lower-level APIs of pdf-lib. Initially, I started some work on this but did not get very far.
There are some functions related to the PDF Outlines Entries (like getNamedDestinations) already implemented, which can probably be used for this as well.

[stefnotch]

Sounds good then. I'm guessing that that is the only major pdf-lib issue?

From what I remember, there are other issues like Hopding/pdf-lib#169 and Hopding/pdf-lib#140 which do have workarounds.

If I want to help with implementing the named destinations API ourselves, where would I get started? I sadly know very little about the PDF file format.

[aarkue]

Great, yes, I think otherwise we can work pretty well with what pdf-lib provides!

But good that you mention the deletePage issue, I think the implementation in https://github.com/aarkue/pdf-tools-web/blob/main/components/ReorderPages.tsx could be improved to actually remove content correctly from the PDF file.

Regarding this issue, I think the best way to get started would be creating/gathering some sample PDF documents with internal links (Probably easy to do with a simple LaTeX PDF with a \ref or \cite ). Next, we would try to find where pdf-lib exposes this link information (likely named destinations + something else). After that, it would be great to try to create a PDF file using pdf-lib which contains internal links.

So basically:

1. Gather sample PDF files with internal links
2. Extract link information using pdf-lib
3. Create a PDF file containing internal links using pdf-lib

Finally, then combining all of this to actually preserve internal links in merged/reordered PDFs should not be too hard, hopefully. 😄