[OSM-dev] [osmosis-dev] [patch] Merge command an resulting bounding box?

Brett Henderson brett at bretth.com
Mon May 9 11:42:37 BST 2011


Hi Igor,

Thanks for the patch!

I just applied your patch to Osmosis trunk before receiving this latest
email.  I've applied it without making any changes, it all looks good to me.

I'm not too fussed about the default of warning.  As you say, it preserves
backwards compatibility which is generally a good thing.  I'm happy to break
compatibility if there's a good reason (eg. unintentional data corruption),
but I don't think dropping a bound element fits into that category.

Can you please update the development version of the wiki usage page with
your changes?
http://wiki.openstreetmap.org/wiki/Osmosis/Detailed_Usage_0.40

Brett

2011/5/9 Igor Podolskiy <igor.podolskiy at vwi-stuttgart.de>

> Hi Dominik, hi all,
>
>
>  So, basically, I tested your patch sucessfully for the case:
>>
>>  2. If both sources have a bound, then a bound corresponding to the
>>> unions of bounds 0 and 1 is emitted to the merged stream.
>>>
>> cool, thanks for the response, nice to know it worked out for you.
>
>
>  In case 3b - when a bound is "removed" -, there are three
>>>
>> possibilities controlled by the new boundRemovedAction keyword argument:
>
>>
>>> * ignore: don't do anything and continue
>>> * warn: emit a warning to the log but continue (default)
>>> * fail: stop processing altogether
>>>
>>
>> Thinking twice about this, I would perhaps consolidate that to
>> default failure (with detailed explanations in the log how to
>> overcome this problem, i.e. how to specify the following keyword
>> argument) plus  a keyword argument "ignoreBounds" which would allow
>> continuing (equivalent to your "* ignore" case). This would ensure
>> that the user looks at the issue and ensures correct functioning or
>> fixing the input maps. - just my suggestion. Maybe case 1 should be
>> covered by this behavior, too - not sure.
>>
> Well... I'm not so sure about that. The thing is, many people don't really
> care about the <bound> being preserved - well, I didn't care, for example,
> till you reported the bug. And if the patch with the default behavior of
> "fail" makes it into a new version of Osmosis, it will suddenly break a lot
> of pipelines which have been working fine. So I'd let it at a default "warn"
> for backward compatibility reasons if for nothing else.
>
>
>  In any case, thanks a lot for your great work - the whole thing
>> works  nicely for me now.
>>
> You're welcome :)
>
> Best regards
> Igor
>
> _______________________________________________
> dev mailing list
> dev at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20110509/c254f28b/attachment-0001.html>


More information about the dev mailing list