Re: Broken aquaterm in 5.0.4

[Mojca M. <moj...@gm...>]

On 22 July 2016 at 20:39, Tait wrote:
>
>> > I'm sorry for not having it tested earlier, but my compiler doesn't
>> > seem to be happy.
>>
>> Proving once again that no change is too small to cause problems.
>> I've uploaded an amended tarball that contains the other half of the
>> aquaterm fix.
>
> It's history at this point, but there seems to be confusion and some
> mild alarm about two gnuplot-5.0.4.tar.gz files floating around with
> different SHA checksums. I imagine that for those who aren't aware
> of this thread, the first assumption may be that the distribution
> was compromised. If it happens again, it might be less confusing to
> just say "oops" and bump the version (e.g. to 5.0.5 in this case).

Indeed, this is a bit of a problem and it's generally not advised to
quietly change the tarballs with the same name. I kept postponing an
upgrade in MacPorts because I knew that people would end up with the
wrong file even after it was already fixed (the mirrors still had the
old file) and it would be difficult to get rid of that one.

Then again, we got a bug report today from users who are unable to
install the software. It seems that some mirrors still hold the old
file:

$ wget http://heanet.dl.sourceforge.net/project/gnuplot/gnuplot/5.0.4/gnuplot-5.0.4.tar.gz
$ shasum -a 256 gnuplot-5.0.4.tar.gz
27897103153ec5c8efd517df9bbe690246aaaa9a59a2ce18eb88d08228ac7723
gnuplot-5.0.4.tar.gz

$ wget http://vorboss.dl.sourceforge.net/project/gnuplot/gnuplot/5.0.4/gnuplot-5.0.4.tar.gz
$ shasum -a 256 gnuplot-5.0.4.tar.gz
151cb845728bde75eb9d1561b35140114a05a7c52a52bd35b4b2b3d944e0c31e
gnuplot-5.0.4.tar.gz

Mojca