PDF processing bug

[skrul]

Here is a possible issue I've noticed while experimenting with the dropbox API. The repro is something like:

1. Find a PDF in your dropbox via the web UI

2. Ue the "Edit PDF" feature, make a change to the document (e.g. delete a page, rotate, etc), then choose Save -> Replace Original

3. Download the edited file using the dropbox API

The file that you download from the dropbox API is corrupted. For example, I have a file with ID "id:VNGR990xOeAAAAAAAAACSw" that originally had two pages and I deleted the first page. When viewing with MacOS Preview app, the API downloaded version still has two pages but the first page has corrupted text; the itext library I'm using to process it refuses to open it ("Error at file pointer 82,080").

However, if I download the file through the web UI the file is fine and looks right.

I've found in the past if I use the "Edit PDF" feature again and save the file again, it is sometimes fixed when I use the download API.

cheers,

-steve

[iNeil]

Hello @skrul,

Based on the information that you have provided, I was unable to replicate this issue. May I ask if you are still encountering the issue while editing and downloading new pdfs? If possible, could you attempt to download the affected PDF to cross-check if the issue still occurs?

If you are still encountering the issue, please provide the response headers when you are executing the 2/files/download endpoint. Also, please let me know if the issue is occurring for all PDFs or only particular ones.

[skrul]

Thanks iNeil, I'll try to gather more information if it happens again.

[DbxNate]

Hi @skrul, thank you for reporting this.

I'm the engineering lead on the team which owns PDF editing. Please do let me know if you encounter this again. I've made a mental note of this issue, and I've subscribed to this thread so that I don't miss any future updates.