[OSM-dev] No new planet.osm files?

[Immanuel Scholz]

Hi,


> Swap cache: add 776211, delete 776211, find 2065432/2097947, race 0+0
> Out of Memory: Killed process 28432 (planet.rb).
>
> I think planet.rb is loading almost the entire node table in to memory
> before spitting it out.
>
> Nick you hacked on this right? Any way to make it row seek or something
> so it doesn't die?

I coded the reading-from-db part.

The probably offending rows are:

def all_nodes
  $mysql.query "..." do |rows|
    rows.each do |row|
...

If 'rows' is an Array, and this completly reads the data into 'rows'
before iteration starts, this is the place to change. If rows is of some
"cursor-like" type, this should be fine.
Can someone check this ("puts rows.class" before the rows.each...)? I have
no ruby environment here at hand.. :(


However, I don't know how it should look like instead. Does someone know
the MySQL API well enough?


Ciao, Imi

PS: Same for segments and ways.