deleting note while writing comment gives no warning to user writing comment

[mchua]

When a note is deleted as a user is writing a comment on it, attempting to save the comment shows the note as simply having disappeared, with no indication to the commenting user what has happened.

1. load stage.yourstreets.org/maps
2. log in as both user1 and user2 (on two separate browsers or computers)
3. as user1, open a note, click "edit," and edit the text and/or move the note and/or leave a comment. Don't save yet.
4. as user2, open the same note, (note that user1's edits haven't shown up yet), click "edit," and delete the note. Save.
5. as user1, save.
6. force-refresh both browsers (you may have to clear your cache)
7. you should not see the note as either user.

This also happens if you save in the reverse order (save user1's edit first, then user2's deletion).

At this point, user1 might be upset and confused, and wonder if they've done something wrong - maybe they wrote a long and thoughtful review of the timing of traffic lights on this intersection, and now that has inexplicably vanished.

[sbenthall]

debugging depends on #323

[sbenthall]

So, it looks like there's no easy way to handle this just with the back end.

[10:48] <sbenthall> WFST question
[10:49] <sbenthall> I'm sending two requests
[10:49] <sbenthall> the first one Deletes a feature
[10:49] <sbenthall> the second Updates a feature
[10:49] <sbenthall> I'm getting a SUCCESS on the second request
[10:49] <sbenthall> is that expected?
[10:49] * dwinslow really digging the propono code he's poking through... http://rafb.net/p/Nrbbm556.html
[10:50] <aaime> afaik yes
[10:50] <aaime> there is nothing stating that an update
[10:50] <aaime> must update anything
[10:50] <aaime> if there is nothing that matches the filter
[10:50] <aaime> nothing will be updated and it's still a success
[10:50] <aaime> wfs 1.1 changed the transaction response to give you counters of how many features where updated/remove/inserted
[10:50] <sbenthall> aha
[10:50] <dwinslow> good thing I told seb to ask you
[10:51] <aaime> think of doing that against a dbms
[10:51] <aaime> an update that changes no rows is still returning successfully
[10:51] <aaime> the same goes for a delete that does not delete anything

I think this means that to get this to work, we would need to have an additional request, either before or after the update request, that checks against GeoServer? to see if the feature in question still exists. Pain in the butt!