Missing file extension when downloading .docx files

[wonx]

Ok, just another one. When I download (via "Open/Download) a .docx document, I get the file without the file extension. The same happens with .doc files.

for example: "Document.docx" -> "Document"

In the case of open document files (.odt), they just don't work. The browser cannot find the file, as if the URL was wrong.

thanks in advance!

[wonx]

PS: for some reason, I experienced the same problem (missing file extension) with some pdf files. Not all of them, just a few, and I don't know why.

[ghost]

Hi again,

I can't reproduce this, which does not mean that the issue is not real, but that we need to qualify it better.

Check that your Recoll version is up to date, kill and restart the webui server. Maybe even run a full recoll indexing pass (recollindex -z) if this does not take too long.

Then I think that the easiest way to study the issue would be for me to look at the recoll log file for the failed queries. If you set it to stderr, it will go to the same place as the webui messages, and if you redirect stderr for the latter, we can get all in one place.

Please run the failing query to a just restarted webui server so that we get the whole log for the query and download and nothing else.

Cheers,

jf

[wonx]

I ran a complete Recoll indexing. The problem seems to have dissapeared now. It's weird.

Thanks anyway :)

[ghost]

Yes, it will sometimes happen that a Recoll index will become weird. There is a bug somewhere, but I can't find it...

Whenever unexplained weirdness occurs (this usually does not happen too often), it's time to reset the index...

I'm not too happy about this, but these bugs which occur infrequently and are detected after the fact are difficult to fix.

Happy that we seem to be making progress anyway :)

[wonx]

I'm sorry, the bug is striking again. And I've been able to isolate it.

In only happens in Firefox (v23.0) when the filename contains spaces in it. When that is the case, the file that gets downloaded only contains the first word of the filename (trucating the rest of the filename and the extension).

ex: "Important document.pdf" would result in a file called "Important".

However, in google chrome (and chromium) (v29 and v28, respectively), it works fine. Even if the filename contains spaces, the file with the whole filename gets downloaded.

As in some browsers it works fine, I am not sure you are responsible for this bug, but I'm reporting it just in case.