mara004

I think I should yet add a `pyproject.toml` buildsystem stub in order to use PEP 517, though.

> There isn't anything fancy to do in the current setup.py so making the move should be pretty easy. +1 > And the release/command workflows should already support it. Sorry,...

I haven't read all of this due to lack of time, but I ran into the same problem as described in the issue title while developing helpers for pypdfium2. Apparently...

pypdfium2's support model currently just prohibits moving page objects across pages, which seemed like the only reasonable thing I could do until this is made possible upstream. I don't know...

Thanks for the hints! Is approach 1 possible with pdfium's public API, or are you also using non-public APIs for that? I see that approach 2 will work, but as...

> especially if you have a way of reading an object dictionary! I would love to know. This is https://crbug.com/pdfium/1694. I'm sorely missing that, too, and hope they'd get a...

A further hack amending your second approach came to my mind: When the pageobject in question is isolated on a page in a temporary document, you could crop down the...

@ajrcarey FYI, I reported this as https://crbug.com/pdfium/2015. Very short description due to lack of time; I just saw related commits happening in pdfium lately and thus wanted to report it....

@ajrcarey Small hint on your code sample (https://github.com/ajrcarey/pdfium-render/issues/60#issuecomment-1349925791): `assert!(bindings.get_pdfium_last_error().is_none());`, which calls `FPDF_GetLastError()` under the hood, only makes sense after failed document loading. It doesn't work with the other APIs (see...

> then it is a serious problem in the Pdfium documentation because the documentation explicitly states that any failing SDK call, not just FPDF_LoadDocument(), can set an error value. IIRC,...