Problem while building Python 3.6 from source.

Chris Angelico rosuav at gmail.com
Sun Jan 15 16:09:17 EST 2017


On Mon, Jan 16, 2017 at 7:55 AM, Michael S <michael.s011235 at gmail.com> wrote:
> @ChrisA: I tried your advice and think it worked. So, I just used $
> ./configure and then $ make. Then I was not sure, whether it had
> worked so I used $ make test and got a fairly long result. At the end
> there was a line like "test successful". Nevertheless not all of the
> tests were successful.

Can you run:

$ ./python

That should fire up the newly-built interpreter.

> So, now I got some questions:
>
> 1) Why did not all of the tests in $ make test succeed? But the end
> line was "test successful". That confuses me.

Not sure - I'd have to see what the actual output was. Most likely
there were some tests skipped; not all tests apply on all platforms.
For instance, there are a number of Windows-specific tests, which
won't be running on your system. Also, Python includes a number of
optional modules, and if you don't have their dependencies, you get a
Python that's fully functional with the exception of that/those
module(s).

> 2) This is more general. In order to get the build-dependencies I used
> # apt-get build-dep python3.4. I also googled for the build
> dependencies but did not find anything. So, how could I actually
> figure out the build dependencies for Python3.6?

They don't often change between versions. Certainly between 3.4 and
3.6 you should have no problems. It's only an issue if you're porting
to a brand new platform or something, and then you have to do the
usual dance of "compile, read the error, grab a new library, rinse and
repeat".

ChrisA


More information about the Python-list mailing list