Dynamically generated PDF don't work

[mattiz]

I have a service generating PDF's on the fly and returning it with the correct Content-Type and Content-disposition, but the viewer won't show it. If I save the PDF as a file, the viewer works as expected.

This doesn't work:
http://mypage.com/viewer.js/#http://mypage.com/service/createpdf

This works:
http://mypage.com/viewer.js/#http://mypage.com/file.pdf

[thz]

the viewer will not fetch the whole document as one chunk but instead do multiple
ranged requests. if you do not want to regenerate the document (server-side) for
every partial request, then you need to cache the generated PDF somewhere.
if you do not care about the fact that you would regenerate the pdf on every request,
then you only need to support the Range-header correctly.

[mattiz]

I did a little digging, and I found the problem. The function loadDocument extracts the file extension to load the correct plugin. My URL doesen't have a file extension, thus no plugin is loaded.

https://github.com/kogmbh/Viewer.js/blob/master/PluginLoader.js#L57

[thz]

ahh - that too.
in your case it would probably best to pass an argument to ViewerJS which pre-determines the plugin-type.

[Mule52]

Thanks for posting this! This was a huge help.

In my situation, I have a PDF being returned from my server that does not contain a PDF extension. I added a default case to the loadDocument(documentUrl) function in PluginLoader.js to load the default PDFViewerPlugin and it all works. I tried appending a .pdf to my URL that did not have the extension, but that is not working. I get an error, "Warning: Unhandled rejection: Missing PDF."

I will stick with what I have working for now as I have a deadline. Thanks for the help by posting this.

[kossebau]

ViewerJS 0.5.5 got released today, and it improves situation around content type detection. It both tries to query to server for the mimetype initially and also can be passed the type as argument in the passed url, e.g. /ViewerJS/index.html?type=pdf#http://mypage.com/service/createpdf

That should solve this problem, so closing this issue.

[fukit0]

ViewerJS 0.5.5 got released today, and it improves situation around content type detection. It both tries to query to server for the mimetype initially and also can be passed the type as argument in the passed url, e.g. /ViewerJS/index.html?type=pdf#http://mypage.com/service/createpdf

That should solve this problem, so closing this issue.

?type parameter is never mentioned in official site. I was struggling to view a pdf from blob url. That parameter made my day! Huge thanks!!!