LW 2.3 beta fixes resumes
 

Hi everyone. Many people are complaining about problems with resumed downloads. Indeed, there are a number of cases where 2.0-2.2 code can fail to resume properly. The new LimeWire 2.3.0 beta should fix these problems. Do try it out and let us know if it does the trick for you. Please be aware that this is still a beta. We know of two platform-specific bugs that still need fixing. Here is <a href="http://www.gnutellaforums.com/showthread.php?s=&threadid=9368">more information</a>.

Sincerely,
Christopher Rohrs
LimeWire

YAY!!!!!
THANK YOU, THANK YOU, THANK YOU!!!!!

:cool: Thanks a million...much better

The beta version does seem to work better for actually resuming the downloads but they start over from 0% rather than from where the file download left off when the connection was broken.

Any ideas?

ARE INCOMPLETE WORTH SAVING ?
 

I am using 2.3 - I have tons of 100 - 500 mg files in the incomplete folder with no apparrant way to resume download? They are not auto-resuming as is clearly states in the user guide.

Is something else required of me? I would pay the $8.50 and go PRO if there was some indication that a full featured product could be offered here.

However, it seems that as poor as other gnutella client interfaces are - at least many of them have much better RESUME features.

Somebody is going to figure this ALL out soon, put it all together and CONTROL this market.

LimeWire 2.3 may not be able to resume incomplete files created by older versions. The older versions sometimes failed to write enough information into your downloads.dat file.

BTW, there is an option to have LimeWire automatically delete old incomplete files: Tools->Options->Saving->Incomplete Files

Resumes seem to work better, but LimeWire still doesn't continue to download files with different names (even if that's the same file), and the requery doesn't work well, it can find a file, and download it from 0% (even if I had 99% of the original).

LimeWire can't resume a file with a different name. In other words, if you have 50% of "file one.txt" in your incomplete directory, LimeWire won't be able to tell that "file two.txt" is the same thing. In the future it may using SHA1 hashes, but for now it will start "file two.txt" from scratch.

And what if I'll rename file one.txt to file two.txt? Will it work?
And if LimeWire doensn't use HASH then why do I get different file names mixed in the same group when I search? And why doesn't the requery search for files with the same name?

Unfortunately renaming a file won't work because the state of the file is stored in downloads.dat. That file isn't easily edited.

The result grouping in the GUI is approximate; it tries to groups files that are similar but not necessarily the same, e.g., "File 1.txt" (100 KB) and "FILE 1.txt" (101 KB). LimeWire won't swarm from both files, but it will try them one after the other.

The autorequery feature "canonicalizes" the file name before requerying, i.e., it extracts the most relevant keywords. Any results matching exactly incomplete files will be resumed. Other results may be downloaded anyway.

Then if you won't implement HASH in the near feauture, I stronly suggest that you'll add an options to change the file names in download.dat or select the name e want for the file to be saved under.

Resume Still Doesn't Work
 

I just downloaded 2.3.2 on the promise that it could resume partially downloaded files. Well, it doesn't work for me. I start to download a file, get partially don, quit Limewire and restart. . . does not resume. I thought that maybe if I found the songs again, named exactly the same, and downloaded again, thy ould pick up where they left off. NOPE.

Mac running OS 9.2
705 MB RAM/125 MB allocated to Limewire
Limewire2.3.2

Re: Resume Still Doesn't Work
 

If LimeWire didn't restart your download, it means that the uploaders are no longer availible.
Also, you must find a file with the EXACT same name (Case Sensitive). Also, once the download starts from 0%, go to your "incomplete" directory and make sure that the file isn't added (I noticed that LimeWire sometimes writes file sizes incorrectly (it writes the same size in KB, but it's different in Bytes), so pay attenstion to the filename (T-XXXXXX-FileName.Ext).

Backwards Evolution
 

I also have GB's of incomplete files that wont resume and get started at 0%. What ****** me off most is that I also am running limewire 1.8b (in addition to 2.3) and it resumes files without a hitch. Why Did You Screw This Up in the new versions!!!!??????!! And if you don't fix this, then AT LEAST don't OVERWRITE the partial file - rename it or something. So until you fix 2.3, its back to 1.8b.

Please implement the hashcode feature soon.
This is the thing, that most users I know are angry about.

Is it possible to download the old 1.8b version from anywhere?

Regards
Hugo

Re: Backwards Evolution
 

The partial file is usually not overwritten when resuming a file, however I don't know how LimeWire 2.3 behaves when resuming a file of a version as old as 1.8b...

It used to be that if you restarted limewire while files were being downloaded, the file would immediately try to continue to download the files. I would REALLY like to see limewire go back to that.

should be, but no search results
 

the search tab shows results, but I don't see any results!

What's going on?

how to resume?
 

Hi, I thought that by "resume downloading files" Limewire might start doing this by itself. I have downloaded the 2.3.3 or whatever Beta to resume downloads and to fix some other problems (which are in fact fixed) but I don't know how to resume downloads because it doesn't do it.
Is it just because these people aren't online, or do I have to find the file again? If that is so, how do I (or Limewire) know it is the same one?
My internet connection sucks at the moment and sometimes disconnects me every 5 minutes, let alone staying online long enough to download a whole song, so this would be useful.

M

I paid the $8.50 for 2.3.3 Pro and I still have this issue. It isn't only they don't resume, it actually erases the files from the incomplete folder in my case. I hope someone is looking into this.

Look at the settings, there is an option to erase files after X days.
LimeWire does resume better, it just doesn't resume files with different names (and it's case sensetive).

Missing incomplete files?
 

We just discovered a bug where LimeWire can delete incomplete files immediately, making resumes impossible. The problem happens if you set the "keep incomplete files" option to anywhere between 24 and 49 days. Values between 75 and 99 days will also have the same effect, as will some other values. The problem is caused an overflow error.

I'm committing a fix now that will be released in LW 2.4. Many thanks to William Rucklidge for pointing this out and suggesting a fix. In the meanwhile, I'd recommend keeping the incomplete time at the default: 7 days.

-Christopher Rohrs

Would you please, please, please write something that recreates the downloads.dat file, using the data in the Incomplete folder?

--another one with gigabytes of incomplete files I'd hate to lose

I'll try creating a class, that does so. It won't be working perfect, however.

Aw, bless yer li'l heart.

Perfect, schmerfect--anything at all will be a big help to lots of us out here. Just think of the bandwidth you're saving with all of us not having to re-download files from the beginning!

Re: Missing incomplete files?
 

Could you also put in the request to add the ability to restart downloads automatically when you restart Limewire, like back in version 1.8? Thanks

Don't be too optimistic. When I said "it won't be perfect" I meant "I don't know if it will even work". So far it seems as if you could leave your computer running for a day or two, before it recovered your gigabytes of incompletes. And even if you tried it anyway, LimeWire still mysteriously locks up, when you try the recovered downloads.dat.

It woud be ideal if you could manually tell a file to resume from a new source, or add download servers. I have a 200 meg file here with 3% to go, that can't find the servers it was downloading from, yet there are at least a couple coming up when I search. to be able to tell the file, 'go from here instead ' would be the stuff.

Also, chat on a mac? doesn't seem to work, and there is little documentation. how is this done?

A

Yes, this would be a HUGE step forward for mankind!
It's a sorry fact that the human mind is more capable to make fuzzy decicions like "this file seems to be the same as that other one", so please let us override the software's logic when we want to.

Well, now that the "resuming" problem seems to be fixed, maybe you guys can figure out the next biggest problem, that is, getting the "re-query" to work....especially without the need to constantly re-search

You might as well toss your earler incomplete downloads, as several previous updates have made it impossible to resume them. A crash or freeze can also corrupt the little 4-8 KB download file, and you might as well free the space for stuff you can use. The current version also has a *preference* for new files, it seems to me. Sometimes, however, it will rewrite existing files, though I have no idea why. I have also had files start with a (1) after them, even with identical names, which, I suppose, means that a few bytes or bits are different.

Hashing will also cut down your frustration in another way: it will considerably reduce traffic. I have had up to six different incomplete downloads of the same file, all due to slight variations in name.