[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Is okular unable to open gzipped PDFs?

Alle giovedì 19 aprile 2012, Stanley Schade ha scritto:
> However, using the command "mimetype" returned the
> correct mimetype and so did the properties dialog which can be
> accessed by right-clicking on the file in dolphin:
> $ mimetype file.pdf.gz
>    file.pdf.gz: application/x-gzpdf
> (Of course I checked that before reinstalling shared-mime-info.)

Usually extensions have a slightly higher priority when matching 
freedesktop.org mimetypes, so foo.pdf.gz is usually considered 
application/x-gzpdf (i.e. a PDF document compressed with gzip).
It seems that `mimetype` is part of a perl module for fdo mimetypes 
(never heard of it before); if you want to check closer how kde 
applications see file types, you can use `kmimetypefinder`.

In case of this "compressed XXX document"-like mimetypes, you can easily 
check what is the mimetype (only by checking the file content though, so 
it could not always give a "good" result) of the compresse document 
$ gunzip -c foo.pdf.gz | kmimetypefinder -c -

> Thus, I still consider this to be strange behaviour of okular.

There's nothing strange: basically what okular does when asked to open a 
file is the following:
1. determine its mimetype
2. is it a compressed file (gzip, bzip2, or xz)? if so:
 2a. uncompress it
 2b. consider the uncompresed file the "new" file to open
 2c. determine its mimetype
3. pick the first generator that handles said mimetype
4. does 3. succeed? if so, it ends here
5. determines again the mimetype of the file, but now only considering
   its content (and no more the file name eventually used in previous
6. is this new mimetype different than the previously used one? if so:
 6a. pick the first generator that handles said mimetype
 6b. does 6a. succeed? if so, it ends here
7. all the attempts failed, so complain that the file couldn't be open

As you can see (hope it is clear and I didn't make it confused :) ), 
okular does not mind the "pdf" part in the application/x-gzpdf mimetype 
determined initially for your document; what "gives you the feeling" 
it's a document is, as said above, because the mimetype detection (as 
used normally in dolphin, the file dialog, etc) thinks foo.pdf.gz is a 
gzipped PDF (even if it's, say, a XML file renamed to .pdf and 
compressed with gzip).
Unfortunately there isn't nothing much that can be done in such cases, 
as the only viable alternative should be uncompressing compressed files, 
which you can guess it would be too expensive for what should be a 
"fast" operation.

Pino Toscano

Attachment: signature.asc
Description: This is a digitally signed message part.

Reply to: