launchpad-dev team mailing list archive

[Jonathan Lange <jml@xxxxxxxxxxxxx>]

Hello Launchpadders,

Most of us agree that there's a fair bit of polish to do on the
Launchpad web UI. Thing is, whenever we sit down and try to think of
something to do about it, we always get stumped by how big the problem
is. Launchpad feels very big and there's so much of it that's unknown
even to the most advanced users.

So, I've written a program that will list every type of page in
Launchpad. With devel r12574 you can run it like this::

  ./utilities/list-pages

And it will print out something like this:
<http://pastebin.ubuntu.com/578349/>. It's a CSV, and you probably
want to view it in a spreadsheet app. The script's docstring has some
notes on the output format. I've pasted it below the email.

The script is held together with chicken wire and gaffer tape, so I
would appreciate it if you look at the output and check for missing
pages.

For the curious, we have 462 different types of page in Launchpad,
built from 311 templates and 131 different types of content.

jml


"""list-pages -- list the pages that are in Launchpad

Prints out a CSV of all of the pages that exist in Launchpad. A "page" is a
whole, actual, web page. Something you might actually want to see filling a
browser window. Page fragments, portlets etc. are excluded.

The output contains the page ID, content interface, layer, template and
example URL.  Here 'layer' is a clue as to the subdomain.

The example URLs will look like::

  https://launchpad.dev/~<ITeam.name>/+addmember

Which means all segments are present, and you can get a real URL by
substituting data (in this case, a team name),

or::

  https://launchpad.dev/[[ITranslator]]/+admin

Which means we couldn't actually figure out the proper URL. [[ITranslator]]
might actually be made of multiple URL segments,

or::

  <IPOFile>/+export

Which means we have no clue how to figure out the URL.  This is generally
because our cheat objects don't match the app-encoded business logic.
"""