Hi!

I started using ktorrent two weeks ago and i have to say: I love it!

However there is one thing i miss out:
You can make a cheksum to se if all files are complete however it takes ocer the entire window and you cant make the program do that automaticly.

My opinium is that ktorrent should a least automaticly do a cheksum before it place a finished download in upploads (somtime it isent as complete as it is supposed to).

You like to see the recheck data dialog as a part of the normal gui, plus an auto recheck data on completion. The last one might be better to have optional because it can be a time consuming procedure.

Anyway, it's up to the devs. I know they have a lot on their plate already, so please be nice with them.

avrage_guy wrote:Hi!

I started using ktorrent two weeks ago and i have to say: I love it!

However there is one thing i miss out:
You can make a cheksum to se if all files are complete however it takes ocer the entire window and you cant make the program do that automaticly.

My opinium is that ktorrent should a least automaticly do a cheksum before it place a finished download in upploads (somtime it isent as complete as it is supposed to).

As of 2.0.2, we are not aware of any data corruption bugs, so I don't really see much need for this.

Not to mention it is time consuming and CPU intensive. However when we reload a chunk for uploading, we could maybe check the data of that single chunk.

As an additional safety check. That way we are sure that what is uploaded is OK. Plus it has the advantage that it will detect corruptions continuously. So if a file gets corrupted after the download is completed, we would still be able to detect this.

But I would need to test this, to see if this doesn't consume to much CPU, checking the hash on a big chunk (> 1 MB), might take to much time.

This is sorely needed. I finally had to just give up downloading a torrent of Knoppix 5.0.1 because after about 10 tries at redownloading, the number of corrupted chunks kept increasing. So I downloaded it from an FTP server and it was fine. But I wasted a lot of time manually checking the data integrity and redownloading useless chunks.

It should be simple enough to have a low-priority thread running in the background to check chunks after they are downloaded. If they pass, don't automatically check them again. If they fail, redownload them. By setting it to a low priority, it shouldn't have any noticable effect on the computer's responsiveness, but it should get the job done without the user having to mess with it.

KTorrent already checks chunks on the fly, and from version 2.0.2 corruption of data on big files shouldn't happen anymore either.

Are you using version 2.0.2?

Yes, I'm using 2.0.2. I'm also having to redownload about a hundred megs of an HD video podcast that failed the manual integrity check.

Damn, that's an issue.
Could you please send me this torrent so i can verify it for the devs wheter the problem is still present on the current develop version?

(E-mailed the torrent to stoeptegel.)

Thanks blujay
Nope, no problems here, no failed chunks and no data corruption in svn.

OK, that's good news.

Thanks for checking. I'll let y'all know if I get any more corrupted torrents.