[R] md5 hash is wrong for 2.6.2 (Windows)

[Duncan Murdoch]

On 09/02/2008 4:41 PM, Paul Lynch wrote:
> Hi,
>    I've downloaded R-2.6.2-win32.exe (base) from two CRAN sites
> (http://mirrors.ibiblio.org/pub/mirrors/CRAN/ and
> http://lib.stat.cmu.edu/R/CRAN/) and for both of those sites the
> expected MD5 hash is reported to be:
>      ed512872818707f27ca05f08be7363c2 *R-2.6.2-win32.exe
> (See, for example, http://lib.stat.cmu.edu/R/CRAN/bin/windows/base/md5sum.txt).
> 
>    However, when I compute the MD5 hash (using two different MD5 hash
> generators) I get
>       4acf8298a829c2787b8fe848f866999a *R-2.6.2-win32.exe
> 
> Could someone check and confirm that the MD5 sums on the CRAN server
> are incorrect?  (And fix it?)  Thanks,
>       --Paul
> 
> p.s.  Given that two CRAN servers have the bad value, I'm guessing
> it's safe to install.  It installed okay is seems to be running fine.

Thanks for noticing this.  I don't know what happened, but the version 
that was uploaded was shorter than the real one, which should have 
30750532 bytes.  The md5sum.txt file is correct for the real one, but 
doesn't match the truncated one.

I've redone the upload, and now the md5sum matches.  It will take some 
time for CRAN to synchronize with the change; it won't be immediately 
visible.  But I also changed the page to say "Last change: 2008-02-09", 
so if you see that date, you should be getting the right download.

I'll have to put some tests into the upload script.

Duncan Murdoch