vinux4 on sourceforge.net is corrupted

John G. Heim jheim at math.wisc.edu
Fri Apr 24 23:47:01 UTC 2015


I have an i386 version you can download here:

http://dl.math.wisc.edu/talkies/vinux-4.0-desktop-i386.iso

On 04/24/2015 05:54 PM, Jude DaShiell wrote:
> There is an md5sum file that comes separate from the iso in the same 
> directory and it's what people are urged to download and use to check 
> the iso.  That file fortunately was packaged with gzip so you need to 
> run gunzip on it before you get access to the md5sum inside.  That md5 
> file and its corresponding iso were in the same directory when I ran 
> the check on the iso file itself.  If you do that without that md5sum 
> file you have no assurance you didn't get a corrupt file since md5sum 
> has no independent .md5 file to check against.  The iso does not 
> produce the result for md5 you mentioned either.
>
>
>
> -- Twitter: JudeDaShiell
>
>
> On Fri, 24 Apr 2015, Henry Yen wrote:
>
>> On Fri, Apr 24, 2015 at 12:11:08PM -0400, Jude DaShiell wrote:
>>> Next I ran md5sum --quiet --check vinux-4.0-dvd-amd64.iso and got a 
>>> bunch
>>> of missing files.
>>
>> The "--check" option is supposed to point to a text file containing
>> a list of md5sums and filenames whose md5sums should match those.
>>
>> If you run md5sum without the "--check" option against the .iso file,
>> you should get "2a5089976a0991aee516a14ff080b384".
>>
>> Note that if you loopback-mount the .iso file and changedir to the mount
>> point, the "--check" option will mostly work (although it produces 
>> errors
>> on a handful of files that are simply missing, not corrupted). The files
>> it checks OK are from the "md5sum.txt" file in the root. I've no idea 
>> where
>> it's getting the additional filenames that are missing, but since all of
>> those are just packages/release lists, they shouldn't affect 
>> functionality
>> anyway.
>>
>>> Final conclusion, sourceforge.net has a corrupted master and has
>>> distributed that corruption to all of its mirrors some time ago.
>>
>> I don't think that conclusion necessarily follows, if using the md5sum
>> on the raw .iso directly.
>>
>> _______________________________________________
>> Blinux-list mailing list
>> Blinux-list at redhat.com
>> https://www.redhat.com/mailman/listinfo/blinux-list
>>
>
> _______________________________________________
> Blinux-list mailing list
> Blinux-list at redhat.com
> https://www.redhat.com/mailman/listinfo/blinux-list
>




More information about the Blinux-list mailing list