Posted by Philip Hallstrom on 09/27/31 11:27

> I have written a script which dynamically generates PDF documents (with
> PDFlib). The link to the PDF file is presented as a static link (thanks btw
> to Richard Lynch and his previous contributions to this list on the subject
> "force download"). It is redirected to the script via a htaccess file. If the
> name of the PDF document matches a certain pattern, than the script will
> output the document, if not, a 404 error page will pop up. Everything works
> very well with all kinds of browsers, except Microsoft Explorer (tested with
> Windows Explorer 6 SP1, Internet Explorer 5.2 for Macintosh). And I want to
> know why.
>
> Explorer shows the PDF document as plain text only. Trying to save the linked
> document to the computer (right click) gives something like "The server can
> not be reached or the document does not exist."
>
> I know that explorer is a real p.i.t.a. when it comes to all kinds of web
> standards. And my best guess is that this could have to do with headers, or?

A long time ago I helped out a buddy who had a problem with dynamically
generated PDFs, but only certain ones. I don't remember all of the
specifics, but this is what I do remember.

What I remember is that it ended up only being a problem on PDF's that
were larger than about 1mb -- but only for IE.

We finally figured out that IE was beginning the download of the PDF
itself, but if it downloaded more than about 1mb, it would pass it off to
a helper application to do the rest. The thing was the helper application
remade the request. But it wasn't passing along the form parameters used
to generate the PDF so he was always getting a blank PDF (blank in the
sense of generically empty).

My memory is that the PDF was the result of a form POST submission and
that changing it to GET fixed it, but he eventually kept the POST
submission, generated the PDF, stored it on disk, and redirected to that.

I don't know... maybe that helps, maybe it doesn't...

[Back to original message]