FileReader can be used to save PDF file to desk but Chrome needs a.href = window.URL.createObjectURL(blob) while iOS needs a.href = reader.result.

However, if the PDF only has ASCII characters then you can use reader.result where reader = new FileReader().

Known issues

1. Chrome has a hard limit of 25MB, if the PDF is larger than that, then the download is aborted.
2. Chrome also has an issue with UTF8 encoded PDF. Chrome can only convert ASCII encoding via the FileReader API, to base64 encoding. Firefox does the right thing. IE11 and Edge uses a Microsoft API, that can handle UTF8.
3. iOS Safari (not desktop Safari) does not support the download attribute on a link, so you can not give the PDF a file name in iOS Safari. https://bugs.webkit.org/show_bug.cgi?id=167341 (fixed but not released).

• https://dotnetcarpenter.github.io/FileReader_Chrome/
• https://dotnetcarpenter.github.io/FileReader_Chrome/createObjectURL.html

To run locally

1. npm i
2. npm start
3. open web browser at localhost:9000

Output of reader.readAsDataURL(blob)

• utf8.pdf Firefox: firefox.base64
• utf8.pdf Chrome: chrome.base64

Chrome bug report

FileReader readAsDataURL can not base64 encode utf-8 characters in a PDF file

https://bugs.chromium.org/p/chromium/issues/detail?id=925096

Steps to reproduce the problem:

1. Download all attached files
2. Run index.html in a local web server
3. Make sure that "utf8.pdf" is checked
4. Click on "Save PDF" button

Does this feature work correctly in other browsers?

Yes - This is just a Chromium problem

What is the expected behavior?

Chrome should download a valid PDF file.

What went wrong?

Chrome reports: "Failed - Network error". That is however not true.

TLDR; Chrome can not base64 encode a file that is not in ASCII

I am using the FileReader readAsDataURL method to base64 encode the PDF file and add that to a hidden link element. This works with example.pdf which appears to be in ASCII. But utf8.pdf which is written in Norwegian fails with "Failed - Network error".

No errors are thrown.

I have also tried to create the blob myself with new Blob([text], {encoding:"UTF-8",type:"application/pdf;charset=UTF-8"}) where text is the return value of text method of the Response interface of the Fetch API. This result in a file download but the PDF file is corrupted.

Any other comments?

In Firefox, the base64 text starts with data:application/pdf; charset=utf-8;base64, and result in a valid PDF file download. But in Chrome the base64 text starts with data:application/pdf;base64, and fails.

The reduced test case can also be found at https://github.com/dotnetCarpenter/FileReader_Chrome