Duplicate NZB detector: Lowercase vs Uppercase.

[Sypher]

This is either a bug or a feature request. I'm using the RSS autodownloader and found one annoying thing:
Yesterday it downloaded the NZB called: "Showname.S03E12.720p.HDTV.x264-RLSGROUP".

Today, it downloaded it again but now in all lower case (showname.s03e12.720p.hdtv.x264-rlsgroup).

This is on Linux where both files can exist right next to each other, so thats why it passed the "duplicate nzb detector".
I however would like to see this tweaked a bit so that it won't happen again, seeing there are lots of possibilities here that can cause duplicate downloads.

[shypike]

So you are asking to the check case-insensitive?
I thought it already did that, but I will check this later.
So indeed it may be a bug or a feature request.
I agree that the check should be case-insensitive.

[Sypher]

So you are asking to the check case-insensitive?

Yeah, if you put it like that... That's what I'm asking
The version used was Beta6 by the way.

[shypike]

There are two checks.
One is the title check in RSS, current case-sensitive, but that's trivial to change.
The other one against the backed-up NZB file is a bit tricky, at least for Linux.
(For Windows and OSX it's a non-issue).
I'll see what I can do.

You would probably have been "saved" by the first check.

[Sypher]

This bug is apparently still alive and kicking in the 0.5.0 stable build.

CHUCK vs Chuck means a new download with the name.

[shypike]

I don't know how you test, but the title check is case-insensitive.
What changes: the case in the title or your filter?

The NZB filename check is case-insensitive on Windows and OSX, not on Linux.
We won't implement a case-insensitive test on Linux.

[Sypher]

We won't implement a case-insensitive test on Linux.

Mind telling me why? I don't see the point on not-implementing it.

[shypike]

Because Linux happens to have a case-sensitive file system.
Adding the code to work around that is rather counter productive.
Besides, the NZB backup system is intended to prevent true duplicates, not similar files.
Having different casing means that the source is different, and therefor not a true duplicate.

I'd see more use in extending the RSS duplicate check to
examine titles from all feeds instead of just the ones from the examined feed (as it does now).

[Sypher]

I know Linux is case-sensitive, but its giving me quite some false-positives.
How about making it an option then? I really do not want the download the same file again if one letter is uppercased and the other one is not.
Or automatically (option?) lowercasing the nzb's.

[shypike]

I'll implement my last suggestion instead.
It's much more useful.
It will solve your issue just as well.

We already have too many crazy options

[Sypher]

I'll implement my last suggestion instead.
It's much more useful.
It will solve your issue just as well.

Ok, cool.

We already have too many crazy options

Gives people a choice eh

[shypike]

Gives people a choice eh

But it does make testing a nightmare.

BTW, I discovered a little error. When you click the download link in the RSS listing
the NZB is downloaded, but the NZB is flagged later as downloaded.
Defeated my duplicate check while testing

BTW: in which situations do you have problems with duplicates?
Within one feed or across feeds?
The first one should not be a problem, since current title comparison is already case-insensitive.

[Sypher]

I'm only using one feed: TVnzb, with lots and lots of filters