Keep attached file name intact while downloading

Ok yes this is a bit annoying to me too. I'd like to fix this. The files are hashed, because then 1) if different people upload the same file, no additional disk space will get used. And 2) when downloading, even if files were uploaded using different names, they'll get the same hash and will get cached by the CDN (if a CDN is in use).

There's a new HTML5 download= attribute that works like so: <a href="path-to-file" download="name-to-give-to-saved-file.txt">...<a> (see this StackOverflow answer) that lets one specify the name of the file name after download — but seems it won't work together with a CDN, because requires the same origin. (Hmm maybe there's some way around this)

Anyway, in your case (on-premise), I suppose there is no CDN. Also, maybe most people won't setup a CDN. So maybe downloading files, using their original names, should be the default behavior. And there could be a config option, to enable the hash name optimization.

(On the disk, they could be saved using hashes though, that'd cause no problems.)

Fixing this is a little bit tricky: I'll need a table that maps file names, to file hashes, and maybe review some reference counting code (that tracks how many times each uploaded file, is linked / used — when the ref counts gets 0 they're supposed to get auto deleted). So probably this will have to wait for a while.

Hmm I suppose I can add the HTML5 download=... attribute, always, and then it ought to work, for you since you don't use any CDN. I'll give this a try.