It could be that all we need is to build gnubg against python3 for it to
work.
C
On Thu, 24 Feb 2022 at 00.58, Philippe Michel <philippe.mich...@free.fr>
wrote:
> On Sun, Feb 20, 2022 at 10:38:32PM +0100, Jim Segrave wrote:
> > I need to look at this in depth- on my host the default python is 2.75
> > (even thouogh nowadays I almost always use python3) and it does what it
> > used to.
> >
> > I really don't remember how the Python integration works, athough I once
> > used it quite heavily to build databases of matches. I'm sure it's
> solvable
> >
> >
> > On 20/02/2022 16:20, Rose Menninger wrote:
> > > Hi. Thanks for your reply.
> > > Unfortunately I forgot to mention that cube-info is supposed to be a
> dictionary and type(gnubg.cubeinfo()) also returns class ‘dict’.
> > >
> > > So, if I try to convert it using “bytes” function (as you mentioned),
> i will get an error from bytes function like this:
> > > TypeError: encoding without a string argument
>
> I don't think you want to "convert the dictionary to bytes".
>
> cube-info is something like:
>
> {'cube': 1, 'cubeowner': -1, 'move': 1, 'matchto': 0, 'score': (0, 0),
> 'crawford': 0, 'jacoby': 1, 'beavers': 0, 'gammonprice': ((0.0, 0.0),
> (0.0, 0.0)), 'bgv': 0}
>
> In python 2, the identifiers like 'cube' are ascii bytes strings and
> that is what the methods taking such dictionaries as argument expect.
>
> In python 3, the same code that creates a dictionary from a list of C
> variables makes the identifiers utf-8 strings (u'cube', implicitely),
> but the methods still expect ascii strings, causing the error you get:
> "TypeError: expected bytes, str found", str being an utf-8 string
>
> I'm not very familiar with python and what would be considered standard
> in python 3 vs. python 2, but I would guess that what needs to be done
> (when built with python 3) is for the methods to accept utf strings as
> identifiers rather than create the dictionaries using byte strings.
>
>
