>From my understanding pip is going to install the same package no matter 
what OS is making the request so this is not really an Ubuntu issue, but it 
is very interesting to me because it is cross platform...

I think the best approach would be to create a Python package that is sort 
of a utility for doing a "standard" Web2py installation. The same basic 
functionality is required for "installing" Web2py applications so we might 
want to consider that as a possibility as well. The trick will be insuring 
that the package remains os agnostic as pip is used "universally" across all 
platforms. The "faux installer" approach might work well for other systems 
that like repackaging things so they fit the systems concept of 
installations, so it could work for Debian and so on and so forth. 

So in a nutshell rather than "packaging" Web2py for Debian (which kind of 
destroys the elegance of not having to install Web2py) it we could "package" 
a Web2py "installer' that basically a glorified python based downloader, 
unarchiver and permissions checking utility that unpacks Web2py (and demo 
applicaitons?) to a location that is handy for folks who are just getting 
started with Web2py like $HOME/Web2py

I think if we do it as a Python module (the Web2py "faux installer" for lack 
of a better description, not Web2py itself) then we will have something to 
give to folks who want to create packages for specific OS's and the result 
will be a standard setup across platforms... ie /Users/me/web2py or 
/home/me/web2py and C:/Users?/me/ <- I forget the Windows parlance for home 
directory.

A number of options are available so this will require a bit of reflection. 
I think our primary goal is to give new folks an easy way to get started, 
but it also has some very interesting potential as far as making things 
easier for developers as well (we could have the package install mercurial 
or even fabric if we wanted to for example). 

My preference would be to put out a really basic one A.S.A.P. that does a 
straight up install and works across nix platforms (any Windows users using 
pip?) and over time we can add new features or create a similar tool for 
developers depending on how much it gets used and if we have demand.

What do people think?

Chris

Reply via email to