[TYPO3-core] RFC #13283: Thumbnail generation broken for PDF files

[Michael Stucki]

Hi Steffen,

> thanks for your work at T3BOARD, I have one question.
> We introduced the thing with pdf, since all other framed file types
> beside of pdf rendered without a frame-number, too. Multipage PDFs wont
> ever without a frame.

Point is that there may be more filetypes than just PDFs which can be
multi-page. The list of supported images is not hardcoded, it can be
configured easily within the Install Tool.

And the 2nd point is that obviously every file will render even when
called with "[0]" after the filename. So why should we drop it if we can
rely that it works with all filetypes (tested with IM4, IM6, GM)?

> You're patch, ignores the case again. So if you have a multipage PDF and
> select noFramePrepended no PDFs will be shown in frontend anymore...
> Other filetypes will just render "the whole picture".

True. We thought about this point yesterday. You are right that
multipage PDFs won't render if this option is set, but that's exactly
the point of the option. People who use it don't want to have the frame
appended!

The reason for using the option is a bug in some older IM5 versions (see
Install Tool). So yes, it won't render PDF files, but it would also not
render them if the frame would have been appended. However, at least it
will now render all other files which don't have the frame appended.

> So your patch fixes "the normal" rendering without "noFramePrepend". But
> as long this option is there, it has to work out with option set too.

Nope. Please proove me that it's possible to render a PDF with a version
that needs this option. I strongly believe that it would not work anyway...

(Well, don't waste too much time on it. We're respecting the value as
expected now, but as I already said, it should not be used at all since
nobody should be using this broken IM version any more...)

> So to me there are just two possibilities:
>  - care for pdf in special
>  - drop noFramePrepended

Right. It should be removed at some point, but that's not what this RFC
is about. It is about to fix the behaviour for all other versions...

> since dopping the option is not possible in already released version,
> there we have to care about. To me it would be OK depreciating this
> option and say that it may fail sum functionality.

We do care about it, see above. The fact that it won't work for PDFs is
not new at all, in fact it never worked for them.

Greetings, Michael