File names containing quotes are not uploaded properly

[coderbot16]

I've started to use the consume-dir mode of dsc recently and I dumped a whole load of old documents in. This worked fairly well for the most part (though I had to restart dsc a few times throughout the process because it would time out when uploading a big file) but as I was cataloging the documents today I noticed a few of them were titled integration: No files supplied. When I looked at them, the only attachments were labeled as No name. This seemed very odd, and I had to spend some time going back through the original files I had backed up to figure out what their names used to be.

I started to fix the names and such, by manually reuploading the documents and deleting the broken ones, but a pattern emerged: the files that were broken were the ones that contained a pair of double quotes in their names, such as a file titled Essay on "The Jungle".pdf.

This also seemed to cause the full-text search of Docspell to silently fail with a 422 HTTP code when one of these broken files appears in the results. Since the files are totally missing an attachmentName field, Docspell fails to decode the response:

Caused by: DecodingFailure(Missing required field, List(DownField(attachmentName), DownArray, DownField(docs), DownField(response)))

I only stumbled on this since I tried to search for the content of the documents to see if I could find one with a similar name, but that did not work. Overall this isn't a big issue, but I assume it's a trivial fix as well.

Thanks for your time and making this nice tool!

[eikek]

Thank you very much for reporting! I need to dig into why dsc creates a bad request in these cases.