Bug 157570

Summary: KGet corrupts downloaded files
Product: [Applications] kget Reporter: Alexey Chernov <4ernov>
Component: MultisegkioAssignee: KGet authors <kget>
Status: RESOLVED FIXED    
Severity: normal    
Priority: NOR    
Version: unspecified   
Target Milestone: ---   
Platform: Compiled Sources   
OS: Linux   
Latest Commit: Version Fixed In:
Sentry Crash Report:

Description Alexey Chernov 2008-02-10 03:15:03 UTC
Version:           2.0.1 (using KDE 4.0.0)
Installed from:    Compiled From Sources
Compiler:          GCC 4.1.2 Target: x86_64-unknown-linux-gnu Configured with: ../gcc-4.1.2/configure --prefix=/usr --libexecdir=/usr/lib --enable-shared --enable-threads=posix --enable-__cxa_atexit --enable-clocale=gnu --enable-languages=c,c++ --disable-multilib Thread model: posix
OS:                Linux

Almost every file downloaded with KGet for the first time is corrupted. Md5sums don't match and tar archives don't unpack. But if I start the same file to download again, sometime it is downloaded correctly (usually after the second time and almost always after the third one). The download task in KGet goes OK to 99%, then it gets "Stalled" status and then it finishes, but the file is bad. Most frequently this problem occurs with sourceforge, but I also noticed it with some ftp's. I also tested the same downloads with Firefox built-in downloader, it gets the files successfully by the first time.
Comment 1 Urs Wolfer 2008-02-10 11:53:29 UTC
Please provide an example link.
Do you something "special" while downloading? Pause the download? What download speed do you have?
Comment 2 Alexey Chernov 2008-02-10 13:35:55 UTC
The problem occurs, for example, with this link: http://ovh.dl.sourceforge.net/sourceforge/gpac/gpac-0.4.4.tar.gz. But note that, as I mentioned before, it's not the persistent problem. For example, while I test this link now it finally was downloaded correctly, after the second try.
I don't actually do anything special, just click the link in browser, select destination in KGet dialog and it starts downloading.
The speed is usually about 150-300 KB/s but at approx. 97-99% it reduces dramatically to 0. I've noticed, if KGet manages to finish the downloading before it reduced, the file is ok, otherwise it gets 'stalled' status and then the file is bad.
Comment 3 Manolo Valdes 2008-02-10 15:23:44 UTC
On Sunday 10 February 2008 12:35:57 pm Alexey Chernov wrote:

Hi Alexey

it works fine here.

can you please check how many threads you have set on the multithreaded plugin 
settings widget?

note that if it is 1 it disable the plugin. and use the regular kio copy-file 
method to fetch the file.

thanks 
manolito

> The problem occurs, for example, with this link:
> http://ovh.dl.sourceforge.net/sourceforge/gpac/gpac-0.4.4.tar.gz. But note
> that, as I mentioned before, it's not the persistent problem. For example,
> while I test this link now it finally was downloaded correctly, after the
> second try. I don't actually do anything special, just click the link in
> browser, select destination in KGet dialog and it starts downloading. The
> speed is usually about 150-300 KB/s but at approx. 97-99% it reduces
> dramatically to 0. I've noticed, if KGet manages to finish the downloading
> before it reduced, the file is ok, otherwise it gets 'stalled' status and
> then the file is bad. _______________________________________________
> Kget mailing list
> Kget@kde.org
> https://mail.kde.org/mailman/listinfo/kget

Comment 4 Alexey Chernov 2008-02-10 15:39:51 UTC
Hello.
It's set 5 in 'number of threads' value, but I don't remember, whether I set it manually or it is so by default. I believe, I didn't change anything in settings. But now I'm in doubt.
Also, maybe the cause of problem is my architecture.. I use x86_64 with SMP, so maybe this is the reason.
Comment 5 Manolo Valdes 2008-02-10 19:28:40 UTC
On Sunday 10 February 2008 2:39:52 pm Alexey Chernov wrote:
> Hello.
> It's set 5 in 'number of threads' value, but I don't remember, whether I
> set it manually or it is so by default. I believe, I didn't change anything
> in settings. But now I'm in doubt. Also, maybe the cause of problem is my
> architecture.. I use x86_64 with SMP, so maybe this is the reason.



5 threads is the default value. so kget is using multithreaded plugin to 
download the file.
can you please check on other architecture?
and let us know

thanks
Manolito
Comment 6 Daniel Nicoletti 2008-02-10 19:35:00 UTC
Manolo I also got this problem too, but these days I'm very busy to post anything.. I still need to do some stuff on kget :)
 btw x86_64 but not sure this is the problem... on doka i experienced this problem too, don't remember exactly what fixed
 

Manolo Valdes <nolis71cu@gmail.com> escreveu: ------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
         
http://bugs.kde.org/show_bug.cgi?id=157570         




------- Additional Comments From nolis71cu gmail com  2008-02-10 19:28 -------
On Sunday 10 February 2008 2:39:52 pm Alexey Chernov wrote:
> Hello.
> It's set 5 in 'number of threads' value, but I don't remember, whether I
> set it manually or it is so by default. I believe, I didn't change anything
> in settings. But now I'm in doubt. Also, maybe the cause of problem is my
> architecture.. I use x86_64 with SMP, so maybe this is the reason.




5 threads is the default value. so kget is using multithreaded plugin to 
download the file.
can you please check on other architecture?
and let us know

thanks
Manolito
_______________________________________________
Kget mailing list
Kget@kde.org
https://mail.kde.org/mailman/listinfo/kget


       
---------------------------------
Abra sua conta no Yahoo! Mail, o 
Comment 7 Alexey Chernov 2008-02-11 00:30:40 UTC
I'm afraid, I won't be able to test it on another architecture... I've compiled only 64 bit libraries, so I don't have multilib binaries and 32 bit support. Really pity..
Comment 8 Alexey Chernov 2008-08-26 17:29:07 UTC
I've thoroughly tested the case for KGet in KDE 4.1.0 (with Qt 4.4.1) and I've found this bug luckily disappeared. I've tested both large and small files and all the md5sums have been similar. It seems to be somewhat fixed.
I think, we can close the bug.