New feature : No more Files Duplicate
 

Xolox allow duplicate download if the file name is identical but file size different.

Xolox should be modified to never download a file if an identical file name already exist.


Additionnal sugestion :

For MP3 : Sometime two file name seem different but in fact there are the same music, for example :

morcheeba - fragments of freedom - 04 - let it go.mp3
morcheeba - fragments of freedom - 04 - let it go.mp3
morcheeba - fragments of freedom - let it go.mp3
morcheba - fragments of freedom - let it go.mp3
morcheeba_-_fragments_of_freedom_-_04_-_let_it_go.mp3

The feature we need is to allow a setting in preference to activate a string comparaison more "inteligent"

example of rules :
two space = one space (All files)
"_" = " " (All files)
" - " = "-" (All files)
5 % of additionnal or missing characters allowed (for MP3 only)
"and" = "&" (for MP3 only)

One thing.
 

about the downloads with different names thing...

Sort a search some time by file size: see ANY that are the same size? no? That's because XoloX uses ALL files are returned from a search which have the same size as the same file.

Do a search for a movie, say "matrix" then start the download from the highest score; open your resumeinfo.txt file and look at the remote file names of the cached hosts. see?

As for the first request... I'd have to do some tests... can't say as I've ever tried to download a file I was already sharing...

I have gigabyte of duplicate.

I have make verification, i have same file name and same file size, probably i selected to download a file that i was already have.

but the file may be different, even it have the same name and size.

But most of my duplicate have not exactly the same file name.

It can't just match on size?

Surely it uses some other criterion to decide it's actually the same file, not just the file size?

Xolox doesn't just match on size.

I can do a search and see files listed of the same size. Do a search for Jackass. You'll see many files of the same size under different names listed.

I dissagree
 

I did a search for "Jackass"... I got 450+ results, and no two items on the list have the same size. Then I did a search for "spears" and got 1300+ results, still with no two items on the list with the same size.

If you really are getting the results you say you are, please sort by size and take a screen shot so I can see. The only thing I could think of that might make two files appear with the same file size is is sometimes takes a few seconds for new search results to be sorted into the list.

Trust me, it happens.

I use it to work out which files are in fact the same encoded versions of episodes. People rename them.

It doesn't seem to rationalize the list either. If they're listed seperately, they stay seperate.

Do your Jackass search again and restrict it to video. Then look around the 200MB files. It happens reasonably often. You might have to leave it searching. It just depends who's sharing what but it does happen.

As I said...
 

Show me a screen shot, these forums allow you to upload stuff (attach file) so you don't need a website or anything.

I assume you're running 1.12 ?

checksum matching + user intervention
 

I like the rules based approach mentioned in the first post...

Here's another idea: once checksums are working, show the user a list of files with matching checksums but differing names. Let the user select those that they think are identical.

Re: As I said...
 

Why do *you* need a screen shot? Are *you* a developer?

I'm curious, why don't you believe me?

And yes, I'm running 1.12.

No
 

I'm not a developer, but I am skeptical. I can almost ALWAYS reproduce soemthing that someone says the software does. I've done extensive testing and under no circumstances have I EVER seen two file names with the same size returned from a search... I want to see your proof because I can't duplicate your result no matter what I try.


Here are a few of my observations and ways to observe them yourself:


1) All files that match the search with the same size are counted as the same file.

Duplication procedure:
Do a search for some extremely popular song with as few terms as possible. Pick the highest scoring file and start the download. Before it's done, pause the download, then look at your resumeinfo.txt file and check out the remote file names (in the [HOST] sections)


2) Only the first matched returned file of a particular size from a host is shown by XoloX, the rest are discarded.

Duplication procedure:
Run Bearshare on port 6437, allow one host connection and accept incoming hosts, but don't auto conenct.

Share your windows setup directory, and make sure that the .cab is a shared extension.

Open XoloX, go to your preferences, and connect to localhost on port 6347.

Do a search for the non numeric portion of the cab file names and CAB. In 98 the search would be 'WIN98 CAB" (no quotes) This test shows that XoloX will only show the first file returned of any size from a host.


3) All but the first returned MP3 from the latest versions of BearShare or Gnotella can be downloaded, the rest are corrupted.

Duplication procedure:
Run Bearshare/Gnotella on port 6437, allow one host connection and accept incoming hosts, but don't auto conenct.

Share some MP3s in BearShare/Gnotella

Open XoloX, go to your preferences, and connect to localhost on port 6347.

Search for ".mp3" (no quotes) and look at the first results that come in. (those will be from you)


4) XoloX limits outgoing results to 30, not 100 as was previously suggested to the public (long time ago)

Duplication procedure:
Run XoloX and another client as shown above, and share > 30 of a file type.

Search in your second client for .ext (where ext is the file type you have shared) and count the results that come from your IP (or localhost)

That's an awful lot of typing and I'll admit I'm not going to do all that fannying about just to test it. That's a developer job. I do it at work. I'm not going to test someone elses software for fun.

Bear in mind I don't see any of the problems that lots of people seem to be reporting with Zone Alarm, memory usage, freezing and disk thrashing. PC's are always different even with the same OS and hardware. There are so many variables it's virtually impossible to get a flat comparison. I'm also running a dual cpu machine so it's also entirely possible that you couldn't duplicate my results anyway. You could then merrily accuse me of forging any screen shots! (Which would also be a lot of work that I can't be bothered to do)

While I appreciate you like to see it for yourself in this case you'll just have to believe me. Or not. It makes very little difference either way!

I posted the message because I have duplicate, but you need to understand why :

After the search result (from an author), i select approx 50 files to download (many album if possible).
But i didnt look very very well each title (impossible to spent one hour to select files)
And I select important files I realy want more than once because I cant manage the download process because i am doing something else.
So I selected some files more than once, please look at my example (morcheeba)
Then I go to job, and when i come back (10H later) i have approx 40 on the 50 files and some duplicate.

The post was NOT a "technical" post, NOT telling that there is a bug in Xolox.

It is a new feature request !

Oups I forgot to add something :

Often I make the process (search and download) several times on an author I like very much, (to try to get all the song from the author).

In this case too, i have many duplicate.

And I am sure that most of my duplicate are here, because i selected song that i was already having.

So the feature request is to verify better in /completed about similar song name when downloading a new song.

Thanks to Xolox team

Hmm...
 

ok, so which are you seeing?

A) Two files in the same search with the same size and different names.

B) Two files in seperate searches with the same size but different names.

C) Two files in the same search with the same name and different file sizes.

D) A file with the same size as one you are trying to download but with a different name, recieved from a seperate search.

I've never seen A

I've seen B

I've seen C

I suppose D is possible, but haven't payed much attention... I almost always pick items with scores > 3 so always tend to get what I'm looking for.

well i have see option E and F , G and H :

E) Two files in seperate searches with different size and same names

F) Two files in seperate searches with same size and same names

G) Two files in seperate searches with different size and similar (but little different) names

H) Two files in seperate searches with same size and similar (but little different) names


all 4 option make duplicate song in my computer

I hope this will help you.

I see... I think.
 

There's obviousy been some communication error, as I've been assumeing you meant 'A' all this time. Of course seperate searches will hold the possability of seeing the same files, as they are blind to eachother's results.

I won' t go in to how I think downloads should be handled in the case of multiple overlapping searches, but I do hope the devs have been at least READING the GDF and are busy implementing some of the ideas there.