Ubuntu repositories

[dangra]

Current repositories setup requires maintaining an apt repository per codename (precise, quantal, raring, saucy, trusty,...) and it bugs us on every new ubuntu release.

The true is that we build debian packages on a Precise host and upload the same package to all repositories. There is a legacy reason for using multiples repos per codename, we started publishing and building debian packages in Lucid (Python 2.6), when Precise arrived we had to build for Python 2.7. Lucid packages were published for Karmic, Maverick and Natty, while Precise for the others. There was also the ubuntu switch to Upstart that affected Scrapyd packaging by that time.

There are two ideas flowing around:

1. Unify repositories and install instructions to:

   deb http://archive.scrapy.org/ubuntu scrapy main
   

2. Move repositories to Ubuntu PPA managed by Scrapy team.

option (1) is simple and will work as far as Python2.7 is available in ubuntu.

option (2) has the advantage that a new debian package is built per codename, and we don't rely on ScrapingHub infrastructure to build and distribute debs.

I intentionally left out the discussion about renaming scrapy-VERSION to scrapy, but it may be related if we want to publish oldstable/stable/trunk versions under the same name but in different repository components.

[dangra]

fixed by #549