[OSM-dev] TIGER upload automation

[Dave Hansen]

On Wed, 2007-08-29 at 20:13 +0200, Martijn van Oosterhout wrote:
> On 8/29/07, Dave Hansen <dave at sr71.net> wrote:
> > I added a return value from the OsmChangeReader->process() function so
> > that we can differentiate between uploaded or not objects.
> 
> Ok, that's to do with the timing changes right?

Yep, so it doesn't count the cached objects as having been uploaded.

> > It adds a -t option to accept additional tags to add to the osm file.
> > This is so I can add something like '--tag tiger:county=Sherman, OR' or
> > '--tag tiger:upload_uuid:yyyyyyyyy'.  Just in case an upload goes
> > bonkers, I can go find the objects from just _that_ upload.
> 
> This is a good idea, I like it!
> 
> > I also tried to change some of the time estimates so that they only
> > track the time we spend uploading.  This is still really hackish.
> 
> Ok, The problem you're trying to solve is that sometimes files are
> partially uploaded and when you restart it it skips a lot of the file
> and that screws the timing. Hrm, tricky.

Yeah, I'm not very confident in what I came up with so far.  It's
definitely hackish.  

> Also, I saw you removed the default for the $api, why is that.

I was uploading to a private server for a bit, and I didn't want to
mistakenly upload to the main server.  You should ignore that bit.

-- Dave