[Potlatch-dev] GPX time elements

[Craig Stanton]

On 12/10/2010, at 9:02 AM, Steve Bennett wrote:

> On Mon, Oct 11, 2010 at 10:57 AM, Craig Stanton <c.stanton at niwa.co.nz> wrote:
>> In my case I know that the single trk is indeed a single track and should remain so. I've written my own pruner to cut out the extra data points and have created entirely valid file. Is there really no way that I can tell Potlatch or OSM that I have a valid file and the 3 minute rule shouldn't apply to long distance hikes?
> 
> IMHO you have a fringe case that you want an unimportant feature of
> Potlatch (and Potlatch 1 at that!) modified to suit. It doesn't make
> sense.
> 
> The GPX import is just a convenience. Fwiw, when I imported an 8 day
> hike, I traced it all rather than importing directly - there were
> plenty of little bits I wanted to chop out. I guess the difference is
> I did the cleanup within Potlatch, whereas you did it before hand.
> 
>> I've yet to hear of anyone that can walk a few hundred miles without a little down time.
> 
> These kind of strawman arguments really piss me off. No one ever said
> such a thing. Just don't do it, ok?
The code says that, by cutting it with every 3 minute break it assumes a new way should be started.

> 
>> I'm probably going to put together a little php script to fake the time stamps on the locations
>> but it seems strange that I'm having to change valid data to fit an assumption that doesn't cover it, what ever that time stamp was going to be used for it's now going to be totally useless.
> 
> Excellent. Generate the fake GPX, import it, convert it to a way, then
> delete it. Then import the real GPX. Best of both worlds.

Yup it worked like a charm. In the end I didn't even need to script it. Just grep out all the lines with a time element, then find and replace to install new time elements with identical values. Does no-one else see this as a bug that I need to modify valid data to get it imported? If it just divided it up into joining ways tI could understand, but actually loosing data is a bug. The points after each cut were missing. 

> 
>> Maybe the import procedure could have a check-box to turn off the automatic splitting feature, or radio buttons to split by time, split by trk or don't split at all.
> 
> There are a million things that could be improved about OSM and its
> software. The developers are working hard on them. In the case of
> Potlatch, they're working on Potlatch 2. The request you're talking
> about would be waaaaaaaay down on the list of priorities. You've
> already identified a suitable workaround, please go for it. OTOH, if
> you want to patch Potlatch to suppport the function you're talking
> about, I'm sure the developers would be happy to look at your patch.
> 
> Steve (probably being un-necessarily pissy)
> <Header>