Patch setup

[benjello]

First tentative to start discussion
I am not able to test it on Windows and I am not very happy about the hack
Comments very welcome

[gdementen]

I'm more into the apt-get camp. I never used aptitude, and from what I read about it, it does not seems necessarily worse (nor better) but AFAIK, apt-get is still the more used/known tool of the two, so would you mind changing that?

[gdementen]

Thanks a lot for taking care of this. I have been wanting to improve the Linux experience for ages but never took the time to do it. BTW, if you could also fix the examples to run on Linux out of the box, that would be super great ;-)

So far, the moves look good. The setup.py changes less so. The ideal situation would be if you made cx_freeze entirely optional (whatever the platform), but allow it to work on linux too.

[benjello]

I started tweaking the tests too. More to come. Thank you for fast review.

[gdementen]

btw, this line needs update:
https://github.com/benjello/liam2/blob/patch_setup/INSTALL#L102

[benjello]

I have a proposal for the tests but unfortunately I hacked on the same branch. Lets deal with the setup first then we will move to the tests.

[benjello]

@gdementen are you okay with these modifications ? Should I proceed with the tests on the same PR ?

[gdementen]

eventually does not mean "éventuellement" in French... ;-)

[gdementen]

@benjello sorry for the slow reaction. It looks good to me IF it actually works, which I fear it does not (see inline comment). For the tests, I would prefer another PR, if that's not too much trouble.

[benjello]

@gdementen : tests and new files hierarchy are a bit intricated. Since it is not that urgent, let's go through this part of the PR and then move to the tests

[benjello]

I tested the install this way by reinstalling cx_freeze
aptitude install mercurial
hg clone https://bitbucket.org/anthony_tuininga/cx_freeze
cd cx_freeze
In setup.py remplace
if not vars.get("Py_ENABLE_SHARED", 0):
by
if True:
python setup.py build
python setup.py develop --user

But I didn't tested more extensively since I didn't know what to expect ... and since cx_freeze cannot work out of the box on my debian machine and the install via pip does fit my needs.
But I certainly could write some tests for that if you explain what you are expecting.

[landscape-bot]

Repository health increased by 0.15% when pulling 26b4c82 on benjello:patch_setup into 64c1419 on liam2:master.

• 271 new problems were found (including 15 errors and 173 code smells).
• 268 problems were fixed (including 15 errors and 170 code smells).

[gdementen]

Sorry, I didn't get to this today. I wanted to finish something else. I hope to be able to squeeze it into tomorrow's schedule (just before my holidays).

[benjello]

Thx

Le jeu. 6 août 2015 21:33, Gaëtan de Menten notifications@github.com a
écrit :

Sorry, I didn't get to this today. I wanted to finish something else. I
hope to be able to squeeze it into tomorrow's schedule (just before my
holidays).

—
Reply to this email directly or view it on GitHub
#181 (comment).

[gdementen]

I just tested it. As expected, it did break .exe generation (and the release script). After quite a bit of trial & error I managed to fix the .exe generation (*). I will commit those fixes now, and fix the remaining of the release script later (when I do the next release ;-)).

For some reason, 99% of cxfreeze-driven setup.py examples I could find (and ALL the official ones) have the setup.py in the same directory than the "main" script.

Thanks a lot for this PR and pushing me a bit to do this. I have been postponing this for too long (possibly because of the cx-freeze examples "issue" I just mentioned, I don't remember).

[benjello]

Thanks for going through this.
Should I proceed with a PR about using nose for testing ?