Well, I never parallel download anything to begin with. Anonfiles gives the max speed (for me) when downloading a single file, so I just do one at a time.
Your response didn't really address what I brought up as a possible problem though, the handshakes. I do agree that if there was no https, there probably wouldn't be problems though. No https, no need for the handshakes. (I did not make up this term, to anyone wondering)
Pause and resume usually happened to me when downloading from anonfiles (like I mentioned near the end when downloading), which might have reinitiated handshakes and possibly caused corruption of the archive.
That's just speculation on my part though, and I'd like someone with more expertise in that area to tell me what they think. I don't think it could be fixed on our end.
zipped files are prone to corruption so i usually don't download it. because with my slow dl speed, it would be waste of hours if file downloads corrupted :( try this, when downloading zipped files, don't open parallels. if parallels is cause of corruption, download without acceleration (1 parallel) & try to avoid repeated pause/resume. hope it would minimize corruption issue :) if file still downloads corrupted, either its bad upload or something wrong with anonfiles server.
i think anonfiles should disable https on direct links & upload servers. its the cause of slow speeds for many & corruption too maybe.
I have a theory regarding anonfiles. I'm not well versed in this technical stuff, so bear with me please.
What I've noticed is that the data comes from a secured server, most likely, hence the https. And when initiating the download, there are several handshakes that occur, similar to how rtmp video streaming works, if memory serves correctly. Downloading is fine, usually up until the end, where the downloading (at least on my end) slows significantly. Sometimes it even stops when I'm at that 97-99% completion mark. I believe that anonfiles has this measure in place that causes the slow down when the file download is almost complete. Anyway, in the times that there are several slow downs or stops, I think that the handshakes need to be done again, and at this point, is where I believe that sometimes the data gets corrupted.
Like I said, I'm not all that technical, but this was based on observations I made when downloading. I use orbit downloader and the log shows a bunch of different information while downloading takes place.
In closing, I don't really think that anonfiles is corrupting the files in the way some of us may think they are. Since redownloading usually does the trick, I think that the slow down and handshakes at the end may be the culprit, since they don't always follow the same pattern.
I dropped this after the jokes started becoming stale and repeated themselves constantly .I guess if it really is ending then the reveal will be the predictable option of movie sequels sometime in the future, or a 2000000000000 yen complete box set :P.
The problem of downloading corruption still exists from time to time. Have not seen any uploading corruption so far apart from one that I came across some weeks ago.
Thanks for the suggestion! A concern I have is that many commenters may not bother to set a type (though the link posters most likely all would), so I haven't added many types.
I'm going to keep the homepage list the same as before, and make the comments listing page filterable.
Please could you a filter on the pop-up menu so people can browse the large BD rips of series/movies/etc that have been skipped and have been greyed out .
Cannot answer that as I'm not sitting at your end. However, Anonfiles and Peejeshare do not have restrictions on the number if connections per file you have with them.
Comment in Feedback 20/02/2013 13:50 — Anonymous: "cool_boy_mew"
*how much
So how much you have to redownload to fix it usually?
Comment in Feedback 20/02/2013 12:52 — Anonymous: "cool_boy_mew"
I guess, depending on how you have to redownload from the torrent (hopefully not much), but with Anonfiles in can easily get a 500mb anime in under 5mins, it's that fast
I may reconsider, but I don't really like uploading to a service that's likely to corrupt things :/ Despite the note on the top of the front page, written in nice small font so that it's easily missed, there will be those that try the links only to get a corrupt file, which can be frustrating. There are others that don't seem to mind the corrupt files.
I suppose it depends on whether they improve or not.
Love the ideas! Yeah, it's mostly a fun page to be quite honest. It relies entirely on the script's ability to automatically determine the correct episode of a file, which at times gets things horribly wrong.
1. There can be quite a number of images on the page, so I don't really want to make them too big. Enlarge-able images are possible, I can't think of a way to do it though (any ideas?). 2. I like this idea. A problem would be trying to figure out which file to take the screenshot from... Logically, it should be the first file received (otherwise we don't have an image to display on that page when the entry gets created), and possibly the middle screenshot? 3. This sounds interesting, but I'm a bit lost on how to design this - any design suggestions?
1. Someone suggested a listing of shows (instead of episodes) on the last feedback page - I think this is probably more relevant for sorting purposes. Though sorting by latest files is probably still going to be a major thing. The problem is that links die fast, so having a nice list of dead links may not be that useful... 2/3. Sites like AniDB, ANN, MAL etc do sorting much better than I'd ever do it here, so I'd recommend looking there for recommendations
Each time you download the same file, do you actually get the same file back from AnonFiles? From my quick tests, I actually seem to be getting different files (different hashes).
To check, you can use something like HashTab (on Windows) - check that the hashes of each file to see if they are the same.
Yeah, it could be an upload problem, but that's difficult to determine if downloads are actually being modified unfortunately. Thanks for your input nonetheless.
I'll have to think a bit about whether it can be particularly beneficial. Generating these files would mean that they have to be uploaded. In which case, we may as well upload the whole file. I see your point in being easier to store, although I find the main issue being time-outs on hosts. You may be able to store a few files on an account, but once you start taking up lots of space, the host will probably start looking at culling files. A 5% redundancy, of, say, 50GB/day, is 2.5GB/day, which still builds up over time unfortunately.
The way 7zipping is currently done, each byte is dependent on the previous one, so a bit flip will propagate throughout the rest of the file.
It is due to poor servers on AF side because no-one uses them and they will time-out your downloads due to poor bandwith.Try BU,ZU,SB,180up,etc instead of them.
Comment in Feedback 19/02/2013 16:39 — Anonymous: "cool_boy_mew"
Please don't remove Anonfile. They did give me a bad file once, but the download is so fast it doesn't matter. On JDownloader it can have has many connection as it wants, making the speed as high as 1-2mb/s at school. I download a lot at school because of the terrible bandwidth limit I have at home. There's not a lot of hosts with amazing speeds or no limits
self-selected categories would probably be the best, unless if you want to just separate links from replies then it could be automated. The first thing that came to mind would be to use a tab-like layout for the latest comment section so the homepage doesn't get cluttered.
About benefit, I would say it is not that "much". It's also NOT necessary if corruption doesn't happen. It's just a kind of "smart" backup, and "redundant".
I'm not sure about what you mean about "delta coding" with 7z. Correct me if I'm wrong. In my opinion, a random bit flip is a flip of a cuple of local bit, not some bits disappear or are added, so it can be fixed with a redundancy block.
PAR is not copyrighted, so it's a bit easier to store. I usually use Ge.tt BTW, IIRC TCP in IPv6 is not checksummed.
Comment in Feedback 17/02/2013 11:05 * — A_Registered_Trademark
Latest Comment listings on the homepage: We've probably come to a point where the latest comments should be divided out/sorted a bit. For example, into categories such as:
1. Download links (Thanks Yogicat, Tibbs and everyone, by the way.) 2. Tech help 3. Info on show (Not airing this week, Recap episode, etc.) 4. Opinions 5. Other
Currently the download link comments scroll the other stuff off the homepage screen pretty fast. That may keep a welcome peace as far as opinions go (just another opinion), but tech help needs are probably getting lost.
Self-selected should be adequate.
Don't know how this can be done without using up too much homepage space, offhand.
Comment in Feedback 17/02/2013 10:26 * — A_Registered_Trademark
I forgot to mention --some additional sorting options would be nice. Right now it sorts by time of arrival I guess. 1. Alphabetical sort --to find specific programs of interest 2. Popularity sort -number of downloads for ideas of what to try and see who we are (if 90% of our downloads turn out to be Pokemon...) 3. Genre sort --to find drama, comedy, younger audience, school settings, etc. Maybe via admin's look-up links to AniDb, etc.
admin's "experimental episodes listing" link has scrolled out of sight and I guess disappears with his next system message. Will there be a permanent button or something on the homepage to get there? It's a great way to see what the latest episode is.
Difficult but fun: 1. A picture that is big enough to see and can be enlarged 2. A picture from the actual episode, not just one standard shot 3. How about the whole row of preview pictures that can be clicked on and enlarged for each title, as it is on the file's download page?
Anyways, thank!
Comment in Feedback 17/02/2013 06:55 — A_Registered_Trademark
AnonFiles: My experiences are not in line with other reports in one respect: Whenever I encounter a "broken file", every re-try resulted in another broken file. Based on my experiences, I would assume the problem, whatever it is, is happening at the upload phase, not the download.
I assume others are looking into this problem in more detail, but I thought I would mention this.
I continue to get a strangely consistent 30% failure rate, like today: 14/21 successes.
PAR sounds like it could be useful. My main wonder is, though, how necessary is it really? TCP is checksummed, so it should be able to detect transmission errors - corruption seems somewhat unlikely here. Okay, AnonFiles is problematic, although they're a relatively rare case. Unfortunately as I'm using delta coding, a random bit flip will corrupt the rest of the file, so some redundancy probably won't be enough to fix it. Even with the rare case of corruption, you could fix it via the torrent, as long as it still has peers. Torrents hash pieces, so it will be able to identify which pieces are bad and re-download good copies of these. PAR files, on the other hand, just seem to increase your chances of getting a good file, and may not work depending on how much corruption there is.
Another complexity factor is where to upload the PAR files to. Is there much of a benefit with PAR files?
10/03/2013 22:05 — Anonymous