[OSM-dev] way 7062297, is this new?

Richard Fairhurst richard at systemeD.net
Thu Oct 9 15:07:47 BST 2008


Stefan de Konink wrote:

> Can I conclude that regression testing is something that should be
> implemented for any editor? Or do you place the blame yourself at the
> API that allows this, while any other editor will drop this?

I love your "should be". You must be new round here.

Yes, maybe regression testing should be implemented. So should a pony.  
And a spaceship.

svn, as they say, is that way ---------------------------------->

But you have also missed the point when Andy asked you to "stop  
reflexively blaming Potlatch". As far as I can tell your  
"investigations" started and ended with the created_by tag, at which  
point your Potlatch gland immediately became inflamed. You didn't even  
look to see whether the version you were complaining about was live  
before demanding it was banned.

Now I'm very very prepared to believe that it could be Potlatch,  
because there was indeed an issue related to tag-handling introduced  
in 0.10c which was fixed in 0.10d (in particular: under some  
circumstances Potlatch might have uploaded empty values in 0.10c).

But it also might not be. Potlatch stores its tags as a hash, it  
transmits them to the server as a hash. Hashes generally don't really  
_do_ duplicate keys.

By way of perspective I will quote from IRC a few minutes ago:

     smsm1: there are 35 nodes and ways that have duplicate tags in the
     current set of data

If it is indeed Potlatch's fault then I will happily agree that what  
it did briefly is bloody stupid. I would even apologise for having  
indirectly contributed to a waste of your time. If, that is, you had  
raised it in a civil fashion.

Richard





More information about the dev mailing list