DC++ NEEDS THIS!!!! (and it's something regular DC has!)

Johner · Post by **Johner** » 2003-03-10 21:41

Hi all. I like DC++ more than direct connect. I am a fan and either way I will stay with DC++ however there is something that would be very easy to change/implement that would make things much easier for DC++ users.

Currently when you download a file it seems to give a crazy rename to it sometimes. (IE: Movieulxaadjf) I wish it would just ask what you wanted to do like in regular direct connect and renames would be simple (1).

MOST IMPORTANTLY, here's the problem...
Often times I will get very far on a download of, say, a movie. I will get disconnected for whatever reason and have a large percentage of the file. Since people use different names for files and DC++ sometimes renames them I can never get the file to resume I have to start all over again or get extremely lucky and find the old use WITH a free slot. What sucks is so many people have the exact same file but with the same name.

So what I did with direct connect was download again from someone really fast, temporarily pause the download delete the file then rename the file I had a large portion of to whatever the name of this download was (because it's the same file but with a different name) then resume the download and WHAA LAA! I would be getting a fast download and shortly complete the file because I just resumed it after almost being done.

YOU CAN'T do this on DC++ 1 because it's hard to "pause downloads" and 2 MOST IMPORTANTLY because whenever I do this it gives me some stupid rollback inconsistency message. DC++ thinks the files are different even though they aren't and I KNOW they are the same. You can try to justify or explain this to death but DC++ would be much better if this were to change. THANKS.

Splicer · Post by **Splicer** » 2003-03-10 23:45

Technically wouldn't these be fixes?

Johner · Post by **Johner** » 2003-03-11 00:30

Splicer wrote:Technically wouldn't these be fixes?

I don't know. But no need to nitpick.

ender · Post by **ender** » 2003-03-11 03:47

You can simply right-click the file you want to downoad and select Download to->Browse. The file will be downloaded to your temp folder with random characters appended to it's name, and moved to the destination folder when it's finished (without the random characters).

HaArD · Post by **HaArD** » 2003-03-12 01:35

Johner,

DC++ can do what you want to resume a partial file and it's a lot easier then the process you described....

First you have a partial download, say 80% of the file:

"The Movie Part 1 of 2.mpg"

and now the user has gone AWOL, so you do a search for the "Movie" and find:

"The Movie pt1.mpg"
"Movie 1of2.mpg"
"The Movie - Part 1.mpg"

all of them appear to be exactly the same same file, based on size of file....

Right Click on one of them and click on "Download To" if DC++ also recognizes the partial file as identical to the one you found then the partial file/path should show up there. Click on it and you've added another source to complete the partial download. The fact that the source, had named the file differently doesn't matter.

HaArD

aDe · Post by **aDe** » 2003-03-12 02:40

..and uninstall Zone Alarm to get rid of the rollback inconsistency message.

Marvin · Post by **Marvin** » 2003-03-12 07:27

Or, if you really are in love with zone alarm and DC++, don't forget to delete your DC++ settings in zone alarm (program Control) when upgrading.

vipah · Post by **vipah** » 2003-03-12 15:09

Why does zone alarm corrupt the downloads? cant it be fixed somehow? Just a while ago i was downloading some files and the next time i checked the file had no users to download from and when i searched it again i got the rollback in-something message.

[KUN.NL]mepmuff · Post by **[KUN.NL]mepmuff** » 2003-03-12 15:39

ZA replaces a part of winsock. for more information, i would recommend the search function on this forum

Johner · Post by **Johner** » 2003-06-11 23:19

HaArD wrote:Johner,

DC++ can do what you want to resume a partial file and it's a lot easier then the process you described....

First you have a partial download, say 80% of the file:

"The Movie Part 1 of 2.mpg"

and now the user has gone AWOL, so you do a search for the "Movie" and find:

"The Movie pt1.mpg"
"Movie 1of2.mpg"
"The Movie - Part 1.mpg"

all of them appear to be exactly the same same file, based on size of file....

Right Click on one of them and click on "Download To" if DC++ also recognizes the partial file as identical to the one you found then the partial file/path should show up there. Click on it and you've added another source to complete the partial download. The fact that the source, had named the file differently doesn't matter.

HaArD

Thanks for the tip but I would still prefer to have the option of not always giving it random alphanmerics so you don't have to worry about it and so sometimes you don't know you have a file you have.

Giving users this option, I feel, is simply a good idea.

Johner · Post by **Johner** » 2003-06-11 23:32

Oh ya and with my way it's all automated so you don't have to manually waste time for just about every download.

Johner · Post by **Johner** » 2003-06-12 00:48

Also when you do "download to..." it often gives many files. I don't know which one to resume because it won't tell me what the sizes are on that menu. So I have to go and look which is a huge (unecesary) pain and often times I can't even find them

TheParanoidOne · Post by **TheParanoidOne** » 2003-06-12 04:23

Johner wrote:Also when you do "download to..." it often gives many files. I don't know which one to resume because it won't tell me what the sizes are on that menu.

The fact that you are seeing them on the menu implies that they are all the same size.

jbyrd · Post by **jbyrd** » 2003-06-12 07:31

Sorry, but your way is not necessary. It is easy enough to go to your download queue and right-click the file you want to resume and click search for alternates. when it finds some, left-click on the first one and hold shift and left-click on the last one to highlight all the files. Then, right-click and select download to...(your file).

Who wants to be prompted for a name when downloading?

Besides the alphanumerics are helpful when searching because it is easy to locate incomplete downloads.

X-level · Post by **X-level** » 2003-06-13 09:08

hmm... shouldn't there be a "incomplete downloads" directory (like in KaZaA

) and not bother anymore with that 8-alphanumerical-stuff at the end of each file ?

Johner has a point though... you should be able to see the size of each file in the "Download to..." submenu... or a tooltip when you hover over a file with your mouse that displays info on that file (see also KaZaA

)...

and yeah, why not "importing" some features from other programs ? (such as KaZaA, Direct Connect - as mentioned by someone -, eDonkey, eMule or even Opera

)) ?

ender · Post by **ender** » 2003-06-13 12:49

X-level wrote:hmm... shouldn't there be a "incomplete downloads" directory (like in KaZaA ) and not bother anymore with that 8-alphanumerical-stuff at the end of each file ?

There is.

Johner has a point though... you should be able to see the size of each file in the "Download to..." submenu...

No point in this, all the displayed files already have the same size

Johner · Post by **Johner** » 2003-06-13 16:28

ender wrote:quote]No point in this, all the displayed files already have the same size

No. You need to read what's going on before you reply. The files are the same size WHEN COMPLETED. I am saying it should say the CURRENT SIZE for each file so you can just pick the largest one, rather than stopping everything then having to go through the pain in the ass of looking in the folder.

Why are some people so objectionable to this?

Post by **GargoyleMT** » 2003-06-13 16:30

Johner wrote:No. You need to read what's going on before you reply. The files are the same size WHEN COMPLETED. I am saying it should say the CURRENT SIZE for each file so you can just pick the largest one, rather than stopping everything then having to go through the pain in the ass of looking in the folder.

Why are some people so objectionable to this?

So you're downloading multiple copies of the same identical file? That's what you're saying, even if you don't realize it.

Johner · Post by **Johner** » 2003-06-13 16:33

GargoyleMT wrote:
Johner wrote:No. You need to read what's going on before you reply. The files are the same size WHEN COMPLETED. I am saying it should say the CURRENT SIZE for each file so you can just pick the largest one, rather than stopping everything then having to go through the pain in the ass of looking in the folder.

Why are some people so objectionable to this?
So you're downloading multiple copies of the same identical file? That's what you're saying, even if you don't realize it.

Here's how it works...let's say I search for "just another movie"

I will get a bunch of sources and when I download them they will all download separately (or at least some of them will). You can't just downlaod the identical files to 1 files (and have every source resume from there) AT THE START.

ender · Post by **ender** » 2003-06-13 17:48

Huh? I don't quite get what you're saying. I often search for anime episodes, and I usually get tens or even hindreds of results for the same episode. I add download of 1 file from the group that has the most files, then select all others and use the Download To -> file I previously selected... If there are two size groups with approx. the same number of results, I do the same for the second group. In both cases I only get 1 file in the Download To box...

Post by **GargoyleMT** » 2003-06-13 18:25

Johner wrote:I will get a bunch of sources and when I download them they will all download separately (or at least some of them will). You can't just downlaod the identical files to 1 files (and have every source resume from there) AT THE START.

No, you're correct, you cannot start one download by selecting multiple sources.

The work-around is to download one file, and add the others as sources.

When DC++ groups search results (which it does not currently), then you can safely bet that selecting the group and using "Download" will download only one file, but with multiple alternate sources. It's not doing that currently, and probably will not do so until there's hash support. (as they might not all be the same file)

Johner · Post by **Johner** » 2003-06-14 00:05

GargoyleMT wrote: When DC++ groups search results (which it does not currently)

That's something that should certainly be added in my opinion, even Kazaa has that! I mean Kazaa! Man!

Anyways I think it would be nice to have all identical files contribute to the same (incomplete) file IF THAT IS POSSIBLE (which I think it is because kazaa does it). That way you don't get a ton of incompletes of the same file, PLUS all your downloads can contribute to the MOST COMPLETE of the incomplete files.

COLYPTiC · Post by **COLYPTiC** » 2003-06-14 03:08

a little late but if you are still having any problems with rollback incosistency change the rollback to 0 in settings, for some reason rollback messes up sometimes in every program I've used to transfer. KaZaa, FTP, WinMX, Grokster, eDonkey, napster, ICQ, AIM, Messenger and a few others. I think its a winsock problem actually.

cologic · Post by **cologic** » 2003-06-14 07:06

a little late but if you are still having any problems with rollback incosistency change the rollback to 0 in settings, for some reason rollback messes up sometimes in every program I've used to transfer. KaZaa, FTP, WinMX, Grokster, eDonkey, napster, ICQ, AIM, Messenger and a few others.

Thanks for spreading corrupted files. Really.

Post by **GargoyleMT** » 2003-06-14 09:02

Johner wrote:That's something that should certainly be added in my opinion, even Kazaa has that! I mean Kazaa! Man!

Uh, yeah, you saw the sign above the door when you entered, right? The one that says "Because Kazaa is the pinacle of P2P programs, we want to be just like it. *cough*bullshit*cough* No, seriously, man."

Johner wrote:Anyways I think it would be nice to have all identical files contribute to the same (incomplete) file IF THAT IS POSSIBLE (which I think it is because kazaa does it). That way you don't get a ton of incompletes of the same file, PLUS all your downloads can contribute to the MOST COMPLETE of the incomplete files.

Well, you cannot tell from filename and size that two files are identical. You need the contents of the file to make sure it truly is identical. Rollback is an approximation of this. Hashes would be even better. Grouping non-identical results (ie. without waiting for hashes) would be a disaster from a support point of view.

COLYPTiC wrote:a little late but if you are still having any problems with rollback incosistency change the rollback to 0 in settings, for some reason rollback messes up sometimes in every program I've used to transfer.

Please don't override the protections that are there to protect the rest of the DC network (and, indeed, yourself) from distributing and downloading corrupted files.

Johner · Post by **Johner** » 2003-06-16 02:48

GargoyleMT wrote: Uh, yeah, you saw the sign above the door when you entered, right? The one that says "Because Kazaa is the pinacle of P2P programs, we want to be just like it. *cough*bullshit*cough* No, seriously, man."

You don't have a clue what I was saying, you also don't seem to have a clue with what you're saying.

Another reason to implement what I suggested is because sometimes files appear to be the exact same size (under the same search of course) but they won't download to the same file. You wouldn't have to worry about this if it was just automated.

sarf · Post by **sarf** » 2003-06-16 04:37

Johner, are you trying to provoke people, or are you some form of stealth troll on a rampage?

You seem to totally ignore what GargoyleMT says (and then turn around and try to blame him for ignoring what you said), even when it actually concurs with some of your points! Now, I'll admit that you might not have gotten that impression from Gargoyles post, but there are a few things that you are right about :

Kazaa does support downloading files from several sources at the same time
Kazaa does support using incomplete files as sources for files
Kazaa does group search results when they are of the same file

Now, how is this possible in Kazaa, and not in DC++ you ask? Well, the answer to that is... HASHING! Kazaa supports this, and DC++ does not, since John Hess did not think it was worth the bother when he created Direct Connect. There are discussions about implementing hashing in DC++, though, and they might even manage to result in some code one day. Until then, however, there is no way of knowing that two files actually contain the same data without downloading them and checking it yourself. In spite of this, DC++ allows you to add files that are of the (EXACT) same size as the original file as alternate sources, since often they actually are the same file. The reason that this works is not in any way supported by any other data than wishful thinking (well... some data does exist), but hoping that people do not do evil things with files and then share them and the usual excuse, "It Works On My Computer" (tm).

I hope you know understand why your position seems a bit ridiculous to me.

So, you're right, it would be good if this was automated. Unfortunately, for it to be automated, hashing would have to be used. For hashing to be used, someone will need to make a few changes to DC++ (hashing the share, making a "hash search", making the client respond to the hash search, using GetZBlock/GetBlock to download files in segments...). This has yet to happen.

Sarf
---
Being good at being stupid doesn't count.

snole · Post by **snole** » 2003-06-22 09:21

Dear Johner......

1. Rollback Inconsistency.
I had Zonealarm FW, and got it all the time. By setting the value to 0, you avoid it, but then you can risk resuming on a file that isn't the same, resulting in bad file. When I removed Zonealarm, and started using Sygate Personal Firewall, all those problems with RI dissapeared. I'd rather have RI activated, so I am sure that when I resume, the file will be checked.

2. Incomplete folder in DC++.
Go to settings, and downloads. There will be a box to specify your incomplete downloads folder. F.example "E:\Download\Incomplete\" Then all your incomplete downloads will be put there.

3. Aplhanumeric things.
The 8 letters put in the end of a incomplete file is GREAT! Then you can see if someone is sharing a incomplete file. If you're looking for files you don't know the filesize of. When file is 100% downloaded, DC++ will put it in your donload DIR, WHITOUT the 8 letters. (They will be magically removed).

4. Resuming.
I resumed files EXACTLY the same way in DC as you described. But I found that DC++'s resuming function is ALOT better. Why you don't realize it, I don't know. When I resume, I right click the file I found in search, and I choose "download to". If the file is a match of the file I have xx% of, it will appear in the "download to" list. It even works when filenames aren't the same. (WOW!)

5. Downloading multiple files.
If you find files with exactly the same filename, and same size, and for some reason want to download them at the same time, you can do so. You doubleclick the first one. It starts to Download. You then rightclick the other(s) and choose "download to". Then "Browse". You can then specify where you want the duplicate, and even what filename you wish to give it. So you can have "Matrix Reloaded.avi" and "Matrix Reloaded (copy).avi" DL'ing at the same time, even if they are the exact same files, and sources have given they the same filename.

In other words, DC++ is ALOT better on file resuming. You haven't realized it for some weird reason.

SNole

cyberal · Post by **cyberal** » 2003-06-25 04:29

I have an idea

today dc++ performs some kind of check on the filesizes when you click on a file you have already partly downloaded, so that the previous file shows in "download to.."

What about if this same check is made when you select multiple files in the search.. and if the files are al the same (based upon size) a new choice in the menu would show... "Download to ONE file" or "Add all sources to ONE file" or something like that..

this would be neat I belive..

jbyrd · Post by **jbyrd** » 2003-06-25 09:49

Yeah, but it would be unnecessary if the existing commands were used correctly. I thought they were foolproof...but I was proved wrong.

TheParanoidOne · Post by **TheParanoidOne** » 2003-06-25 10:12

jbyrd wrote:I thought they were foolproof...but I was proved wrong.

Nothing is foolproof to a talented fool.

It is impossible to make anything foolproof because fools are so ingenious.

X-level · Post by **X-level** » 2003-06-29 12:22

Dear sNole...
as for the "alphanumerical" stuff... there IS a way to share your incomplete downloads without the 8-chars ending part... just rename the file (manually) as you wish, then (while DC++ is closed) modify your "Queue.xml" file so that the TempPath will point to that file and voila...

so this "reason" of yours is not really useful... yes, for "normal" people it should show you the "partially downloaded file" status, but... as i mentioned...
the same thing is with the "Download to..." option... people can share files of the exact size, but not the ones that they supposed to share (changed the filenames and fooled you)...

so... i guess, there should be a distinct topic about getting your files as they are, not dummies...
a solution to this last problem could be a CRC check (& stuff) which should be done when getting a file from someone (and there can be set a lower non-zero limit for the "Rollback" so that to be sure DC++ can check that file)... then the result should be seen at the "Errors" column in the "Download Queue" and the download should continue or not accordingly...
(e.g. a CRC check for every 1024 bytes, and so DC++ should resume a file from a multiple of 1024 bytes size: let's say a file has 5432 bytes and you partially downloaded 3174 bytes... then DC++ should resume it from 3*1024=3072 bytes, so the rollback would be 102 bytes)...

Post by **GargoyleMT** » 2003-07-04 09:27

X-level wrote:a solution to this last problem could be a CRC check (& stuff) which should be done when getting a file from someone (and there can be set a lower non-zero limit for the "Rollback" so that to be sure DC++ can check that file)...

Hashes are better.

SirPlus · Post by **SirPlus** » 2003-07-15 10:03

I love how DC++ works - once I worked it out!

If the file is exactly the same size and the name or path name indicates it is similar then it is "Exactly the same file" to within a infintessimaly small margin of error. If it woofs and smells like a dog then it is a dog.

If you are using it right u can't miss

I bet the zone alarm issue was the real problem here.
so setting the rollback to 0 will have fixed it.

Perhaps adding some clear instructions about "adding sources" and "Rollback" in the Newbie Help will solve these issues.

cologic · Post by **cologic** » 2003-07-15 10:31

setting the rollback to 0 will have fixed it.

Ugh no.

I propose a check in DC++ such that one cannot set the rollback to 0, and to additionally pop up an explanatory message as to why it's fucking stupid.

Post by **GargoyleMT** » 2003-07-19 14:04

cologic wrote:I propose a check in DC++ such that one cannot set the rollback to 0, and to additionally pop up an explanatory message as to why it's fucking stupid.

Amen.

I'd almost go as far as warning on large write buffer values too. It doesn't corrupt files, but it wastes a hell of a lot of resources for the person kind enough to send you the files.

DC++

DC++ NEEDS THIS!!!! (and it's something regular DC has!)

DC++ NEEDS THIS!!!! (and it's something regular DC has!)

Tooltips could be a great feature...

Re: Tooltips could be a great feature...

Re: Tooltips could be a great feature...

Re: Tooltips could be a great feature...

Re: Tooltips could be a great feature...

Re: Tooltips could be a great feature...

Re: Tooltips could be a great feature...

Re: Tooltips could be a great feature...

Re: Tooltips could be a great feature...

Hmmmm