Hi,
On 26.04.2011 22:29, Frederik Ramm wrote:
You could probably save some time by doing something like this
(shortened by dropping arguments):
curl ... -o - | osmosis --rb /dev/stdin --tee 3 --tf --tf -tf --wx --tf
--tf --un --wx --tf --un --uw --wx
that way you would start processing while the download was still
ongoing, and process all three filters at the same time. Logically it
would even be possible to add the merge step to the back of the chain
but I remember some past discussion on the Osmosis list about potential
deadlocks if you combine what you have split in the same task.
exactly, to do tee+merge in a single process you'll need --buffer/--b
tasks after the tee or you will deadlock for sure. So the pipeline would be:
osmosis --rb /dev/stdin --tee 3 --b --b --b --tf --tf -tf --wx --tf --tf
--un --wx --tf --un --uw --merge -merge --wx
(Though if I'd be doing this for real, I'd spell out the at least the
out/in-pipe names, in my experience the pipe stack is quite easy to
break and quite hard to debug in big pipelines like this operating on
data sets >1G).
Best regards
Igor
_______________________________________________
dev mailing list
[email protected]
http://lists.openstreetmap.org/listinfo/dev
