config.json

{
  "vars": {
    "@gray-darker": "lighten(#000, 13.5%)",
    "@gray-dark": "#646464",
    "@gray": "lighten(#000, 33.5%)",
    "@gray-light": "lighten(#000, 46.7%)",
    "@gray-lighter": "lighten(#000, 93.5%)",
    "@brand-primary": "#323232",
    "@brand-success": "#5cb85c",
    "@brand-info": "#5bc0de",

config.json

{
  "vars": {
    "@gray-darker": "lighten(#000, 13.5%)",
    "@gray-dark": "#646464",
    "@gray": "lighten(#000, 33.5%)",
    "@gray-light": "lighten(#000, 46.7%)",
    "@gray-lighter": "lighten(#000, 93.5%)",
    "@brand-primary": "#323232",
    "@brand-success": "#5cb85c",
    "@brand-info": "#5bc0de",

zaius

ctrl-z
bg

touch /tmp/stdout
touch /tmp/stderr
gdb -p $!

# In GDB
p dup2(open("/tmp/stdout", 1), 1)
p dup2(open("/tmp/stderr", 1), 2)

onyxfish

Suggestions from the 11 hackers at the table:

* Use connection pooling (pgpool).
* Don't expect reporters to get excited until you can show them something. (Find a way to appeal to reporters interests.)
* Only update what's changed. (e.g. on election results: show changes, not raw numbers)
* Use the AP's "dbready" format for election results.
* Use CSV for everything.
* Use pdb with runserver for debugging.
* Beware circular imports when using Haystack.
* Make the case for building news apps with government data. (Niran will provide numbers showing that people look at it.)

dcramer

def attach_foreignkey(objects, field, select_related=None):
    """
    Shortcut method which handles a pythonic LEFT OUTER JOIN.
    
    ``attach_foreignkey(posts, Post.thread)``
    """
    field = field.field
    qs = field.rel.to.objects.filter(pk__in=distinct(getattr(o, field.column) for o in objects))
    if select_related:
        qs = qs.select_related(*select_related)