* All of these were caused by the presence of multiple body tags on the
page, which happened because networkD3's sankey plot generates SVGs
containing body tags via SVG's foreignObject tag
* In various places, the 'body' jQuery selector string is used under the
assumption there is only one 'body' tag on the page. The presence of
multiple 'body' tags breaks reliant code in strange ways.
* The fix was to use document.body or 'body:first' instead of 'body'.
- Simplified dragHover "plugin" by counting children instead of storing them.
Counting children fixes Firefox 57+ bug (to be found or filed) that causes
text object of input element to produce drag events
- Removed multimethod since it's no longer used anywhere
- Firefox 57+ appears not to trigger a change event when the `files` field is modified,
which prevented uploads from occuring. This commit triggers a change event manually
and doesn't impact the functioning of other browsers.
Actually three separate issues addressed. Fixes#2197.
- brush.importOldBrush() was not being called anymore, due to it being
registered as a load handler after the image was already loaded (this
was a very recent regression, less than 24 hours old).
- Each time the brush changes, the plot is redrawn twice. This was
because importing the old brush introduced floating point errors that
led to a slightly different new brush being created.
- Sometimes the image's load event wasn't firing at all. This is due to
behavior in WebKit where assigning an image's src to its existing
value is a no-op.
* wait for image to be loaded in browser before initializing handlers
reverts similar behavior in 3354a47e8a
* default the height and width to the image clientHeight and clientWidth
* use raw image clientWidth and clientHeight instead of container clientWidth and clientHeight
prevents being able to brush on non image areas
Instead of providing alternate defaults for optgroupField,
optgroupLabelField, and optgroupValueField, respect the
selectize instance's settings for those fields.
Fixes#2192. Two problems here:
1. It's not documented but apparently we supported data frames for
choices in updateSelectInput/updateSelectizeInput (it doesn't
appear to work correctly for selectInput/selectizeInput though).
This was used in 023-optgroup-server as well as by the user who
reported #2172.
2. The example in 023-optgroup-server was also counting on the
default value of optgroupLabelField, which (starting post-Shiny
v1.1) was being set to a new default of "group". That now won't
happen unless optgroupField is also blank. I'm less confident
about the ramifications of this change. The selectize docs with
the relevant bits are here:
https://github.com/selectize/selectize.js/blob/master/docs/usage.md#data_searching
This bug is new since v1.1. When results are returned from selectize's
server-side endpoint, iff no results have been selected before, then
the control should be set to either its specified initial value (the
one specified in selectInput/selectizeInput) or, if none was provided
AND the selectize control is multiple=FALSE, then select the first
entry automatically.
That's the desired behavior; the bug was that last part, "select the
first entry automatically", was happening whether results had already
been selected before or not. This was causing merely typing in the
control to cause the value to be changed.
Fixes#2191
