Resolves#1000. In Flot 0.7 we only calculated tickDecimals and
tickSize once, when creating the tickGenerator for the first time. This
meant that calls to setupGrid failed to recalculate the values, as
reported in #860. #861 fixed the problem by moving calculation into
tickGenerator, but this turned out to cause a new problem, since the
function doesn't run when an explicit ticks array is provided.
This commit solves both by always recalculating the values outside of
the tickGenerator function. As far as I can tell the only reason it
wasn't done this way from the beginning was to avoid unnecessary work in
the case where tickGenerator is already provided in the options. But
the extra work is negligible, and it's actually more consistent for the
properties to always be set.
The core implementation used parseInt, which was originally to catch
text values, but really only stayed due to its side-effect of flooring
the values. The canvas implementation has never rounded coordinates.
This led to various one-pixel rendering glitches between the two
implementations. I've fixed the problem by consistently rounding
coordinates to the nearest whole number.
Line height is a core font property - it can be specified as part of the
unified 'font' definition - and we therefore need to give users access
to it when they're manually defining the font.
The way in which xaxes/yaxes inherit options from xaxis/yaxis resulted
in a minor bug, where tickColor defaulted to the xaxis/yaxis color
instead of the color for its axis. Fixed by applying the default before
extending the per-axis options, resolving #984.
There's still some questionable behavior here; this section should be
revisited for 0.9, especially with an eye towards removing some of the
code that only exists for backwards-compatibility.
The purpose of the stylesheet hack was to provide a default without
having to use inline styles on containers. We can do this much more
neatly by instead just giving the inline styles to a parent container,
leaving users free to customize the children.
Add an extra pixel to ensure that the lines cleanly join both top and
bottom ticks. Also apply the sub-pixel offset only in the necessary
direction, so the lines don't start a pixel off from where they should.
The axis color now controls the color of the axis line, instead of its
ticks and labels, while the tickColor controls the tick color. This
makes a little more sense and provides the minor feature of being able
to change the axis line color separately from that of its ticks. Pull
request #917 ought to be easier to merge now, too.
This allows us to reuse the canvas elements without having to retain the
rest of the Canvas object, which should really be reset when the plot is
reconstructed. It's also a little simpler, and the Canvas constructor
just feels like the right place for this code.
The tickLabel class is deprecated in favor of flot-tick-label, but we'll
continue to use it until the release of version 1.0.0, for
backwards-compatibility.
Previously the cache was divided only by layer, with entries keyed on a
string built from the text and style. Now the style has its own tier in
the cache, i.e. layers > styles > text > info.
This introduces some complexity, since the nested for loops are ugly,
but at the same time we avoid having to create the cache-key strings.
More importantly it solves the problem of uniqueness that exists when we
try to join strings that may contain arbitrary text. It also allows a
further optimization in the canvas plugin, which can now set text style
and color just once per distinct style, instead of with every string.
This lets users 'namespace' text more naturally, i.e. placing x-axis
labels in a different container from y-axis labels, providing more
flexibility when it comes to styling and interactivity.
Internally the text cache now has a second tier: layers > text > info.
The getTextInfo method previously added new text to the top-level
container when measuring it. Now it adds the text to the text layer,
just as it will be when rendered, so that parent-child CSS rules can
resolve correctly.
This also avoids having to safe a reference to the top-level container,
since it wasn't used anywhere else.
Every cache element now contains the actual text element instead of just
its HTML, plus a flag indicating whether it is visible. The addText and
removeText methods control the state of this flag, and the render method
uses it to manage elements within the text container. So where we
previously used drawText to actually render text, now we add each string
once, then let the render method take care of drawing them as necessary.
This dramatically improves performance by eliminating the need to clear
and re-populate HTML text on every drawing cycle. Since the elements
are now static between add/remove calls, this also allows users to add
interactivity, as they could in 0.7. Finally, it eliminates the need
for a separate 'hot' cache.
I also removed the unnecessary 'dimensions' object; it's easier and
faster to store the width and height at the top level of the info
object.
This significantly improves performance, since we already create the
elements when measuring them, and that effort is now no longer wasted.
We must take care to detach, rather than remove, when clearing the text
layer, so we can add the elements back later if necessary.
Since the Canvas .text object is jQuery-wrapped, it was not preserved as
expected when clearing the canvas of junk. I've replaced the selection
with one based on element classes.
Instead of giving the axis its own font property, we simply look at its
options, where the font comes from in the first place. A separate
property is unnecessary and inconsistent with the way other axis options
are handled.
The base implementation uses the new drawText and getTextInfo methods to
draw text in HTML. Canvas rendering has been moved to overrides of
these methods within the canvas-render plugin.
These methods provide a common way to draw HTML text above a canvas.
The getTextInfo method generates div HTML for text with a given font
style/class and angle, measures the element's dimensions, and saves
everything in a cache. The drawText method takes the resulting entry,
finishes generating the inline styles necessary to position the div, and
adds the result to a buffer. The render method dumps the buffer into an
overlay and expires unused cache entries.
Moved canvas creation and size management into a new Canvas class.
This is the first step towards a more object-oriented architecture.
Since we create multiple canvases, and have to maintain several
module-global variables to track their properties, they are the ideal
place to start.
This commit also removes sizing code that was duplicated between
makeCanvas and resizeCanvas.
This also includes a tweak to zero's default. Previously zero only
received a value if lines were visible; now it always receives a value,
matching the behavior of other contextual options.
Use bind, rather than .mouseleave, because we officially still support
jQuery 1.2.6, which doesn't define a shortcut for mouseenter or
mouseleave. This was a bug/oversight that was fixed somewhere around
1.3.x. We can return to using .mouseleave when we drop support for
1.2.6.
Added a format option 'autoscale' that controls whether the given point
is considered when determining an automatic scale.
The lines & bars 'zero' option controls whether autoscale is set on the
dummy point that is inserted to create the series lower-bound.
Area and bar plots normally use a minimum of zero, since their purpose
is to show size, and using an auto-scaled minimum distorts the plot's
meaning. But this behavior is undesirable in cases where the plot type
is used in more of a decorative sense.
The zero option provides a way to control this behavior. It defauls to
true for bars and filled lines.
David Schnur
martinqt
Sean Jordan
Jamie Hamel-Smith
hizhengfu