Matt Amos [Sat, 21 May 2011 14:15:49 +0000 (15:15 +0100)]
Updated to libxml-ruby 2.0.5 and fixed code accordingly.
I've tested this through the unit tests and by hitting it with
bulk_upload.py and neither fail or cause the server to crash or
go into an infinite loop. Both of these things happened randomly
with 1.1.3/4 due to an apparent early deregistration of the
expanded nodes.
Tom Hughes [Sat, 21 May 2011 09:44:52 +0000 (10:44 +0100)]
Make sure unused layers are hidden
When parsing an old style layer configuration string, make sure any
layer which isn't active is marked as invisible so that we don't
include it when generating permalinks. Fixes #3773.
Tom Hughes [Tue, 17 May 2011 20:35:59 +0000 (21:35 +0100)]
Use a pessimistic version constraint for libxml
It seems that rails doesn't handle gems wth multiple version
constraints right - it generates invalid install commands. So
we use a pessimistic constraint instead.
Tom Hughes [Fri, 13 May 2011 16:27:12 +0000 (17:27 +0100)]
Handle large bounding box changesets better
If the bounding box of a changeset is much larger than the map then
there seem to be problems with OL not rendering the feature correctly
so reduce the highlight to just be the extent of the map plus a small
margin in that case.
Tom Hughes [Fri, 13 May 2011 09:06:16 +0000 (10:06 +0100)]
Make "all traces" and "your traces" links more consistent
Make sure we always show the "all traces" and "your traces" links
in the same place and that they don't accidentally inherit parameter
values that we don't really want.
Based on work by Andrzej Zaborowski <balrog@zabor.org>.
Allow any valid (per RFC 3986) scheme name in OAuth callback URLs
Changed the callback_url regex so that you can have custom schemes such
as myapp: or my-lovely-app: instead of just http and https. This should
help with mobile development where users can authorize with OSM in their
browser and are automatically redirected to the registered app.
Tom Hughes [Fri, 6 May 2011 23:05:49 +0000 (00:05 +0100)]
Make the edit action default to z17
If an attempt is made to invoke an editor without specifying a zoom
level (as with the GPX edit links) then default to z17 instead of z14
to avoid trying to load too much data. Fixes #3728.
Mikel Maron [Sat, 30 Apr 2011 20:07:47 +0000 (15:07 -0500)]
Add a map to the changeset list page
Add a map to the changeset list page, with a bounding box for each
changeset and highighting when the mouse is placed over either a box
on the map or a row of the table.
Tom Hughes [Wed, 6 Apr 2011 19:21:37 +0000 (20:21 +0100)]
Switch to i18n version 0.5.0 but keep old style interpolation
It's impossible to stop rails 2 loading a newer i18n gem if it is
installed, so go with the flow and use the latest one but override
the interpolation pattern for now.
Tom Hughes [Mon, 4 Apr 2011 21:44:43 +0000 (22:44 +0100)]
Try and patch up poor browser language selections
The HTTP RFC says that a language range specified by a browser only
matches a resource whose language tag is the same, or longer. This
means that a browser language range of de-DE will not match a resource
with a language tag of de.
Because of this browsers should always send the generic range, such
as de, as a secondary choice after a more specific range like de-DE.
Some browers don't this however, so if we don't get a language match
we try and patch up the list of language ranges by inserting the
more generic ranges.
Dan Karran [Sun, 3 Apr 2011 16:50:31 +0000 (17:50 +0100)]
Fixing the user settings page for mobile use
Also add a quick workaround for the occasions when the top right
links overlap the logo and site name. We'll need a better fix for
this in the longer run.
Andy Allan [Mon, 21 Mar 2011 16:29:30 +0000 (16:29 +0000)]
Tweaks to the contributor terms styling
The 60% trick doesn't work now that the terms are what defines the height (rather than the sidebar). Also, add a margin to bring the floating buttons within range of the slim container.
Andy Allan [Mon, 21 Mar 2011 16:21:55 +0000 (16:21 +0000)]
Change around how this works - the container adjusts to the right height, and it's the (fixed-size) header that's absolute positioned instead of the content