Re: [Maxima-discuss] bfloat accuracy, %pi

[Raymond T. <toy...@gm...>]

>>>>> "Wolfgang" == Wolfgang Dautermann <wol...@da...> writes:

    Wolfgang> On 2018-06-05 00:17, Raymond Toy wrote:
    >> "dirty" is because you built maxima yourself and you've modified
    >> some (any) file in the tree.  (Or you're using an installer that did
    >> the same before building the result.)

    Wolfgang> Hi,

    Wolfgang> I assume, that was my fault, when I created the installer.

    Wolfgang> Of course did modify the source (update wxMaxima, increase the version
    Wolfgang> number) and created installers (64+32 bit) and tested them. And then I
    Wolfgang> uploaded the installers - and committed my changes since they were
    Wolfgang> tested and they worked.

I think it's dirty only if you have files that aren't checked in.  If
you do, then you'll get something like 5_41_<n>_g<digits> where <n> is
the number of commits after the tag 5.41 and <digits> are the SHA-1
value of the last commit.

    Wolfgang> I should have recompiled the installers after I committed (and pushed?)
    Wolfgang> the changes again for uploading.
    Wolfgang> Crosscompiling the Windows installers takes much time on my Notebook and
    Wolfgang> I did not know, that recompiling is necessary.

    Wolfgang> Sorry. I will consider that, when I crosscompile the next Windows installer.

I don't think it's a disaster that it says "dirty", but it might scare
someone that they didn't get the right version. (I assume if they
cared they could check the md5 hash or something of the installer, but
I know I never do that either.)

--
Ray