Bugs: axes

The following does not work:

add_axis (Y_AXIS, 0,0,10,"y.label=FOO label.angle=45")

The axis label is not rotated 45degrees. It must be rotated separately with a call to >set_axis("ay1", "label.angle=45").

The bug depends on what attributes are changed.

chips-1> x = 10**np.linspace(-1,1,20)
chips-2> y = x**2
chips-3> add_curve(x, y)
chips-4> get_plot_range()
         [-0.395, 10.495, -4.9895000000000005, 104.9995]
chips-5> log_scale(X_AXIS)
chips-6> get_plot_range()
         [0.07943282347242815, 12.589254117941673, -4.9895000000000005, 104.9995]
chips-7> set_axis(["pad", 0])
chips-8> get_plot_range()
         [0.1, 10.0, 0.010000000000000002, 100.0]
chips-9> undo()
chips-10> get_plot_range()
          [0.1, 10.0, 0.010000000000000002, 100.0]

The plot and ranges have not been reverted to the state at line chips-6. Calling log_scale on the axis again fixes up the plot:

chips-11> log_scale(X_AXIS)
chips-12> get_plot_range()
          [0.07943282347242815, 12.589254117941673, -4.9895000000000005, 104.9995]

As a result, plotted contours or image may appear incorrect, and adjusting limits may not work correctly.

If the an axis label is set using the axis.x.label or axis.y.label attributes of an add_<object> call then the string must be surrounded by double quotes if it contains spaces; for example:

chips> add_curve(time, zoff, ['axis.y.label', '"Z offset (arcsec)"'])

Axis labels displayed in sexagesimal format - e.g. tickformat=%ra - may be invalid when the axis range is small (several arcseconds) and a WCS transform is not attached to the axes.

This only affects plots with linear axes. It does not occur if the plot was created by either add_image or add_contour, or a WCS transform object was used when calling add_axis.

If you display an image that contains an unsupported axis transformation (e.g. RA--SIN,DEC--SIN), it looks like the projection is supported - i.e. the axes are drawn - but they are not correct.

In the following example, after line 6, the plot axes are not correctly displayed. Changing the ticks to nice and then back to arbitrary causes the plot to correctly refresh.

chips-1> split(2)
chips-2> add_curve([1,2,3], [1,2,3])
chips-3> set_arbitrary_tick_positions('ax1', [1,2,3], ["foo", "bar", "baz"])
chips-4> current_plot('plot2')
chips-5> add_curve([1,2,3], [1,2,3])
chips-6> bind_axes('plot1', 'ax1', 'plot2', 'ax1')
chips-7> set_axis("ax1","majortick.mode=nice")
chips-8> set_axis("ax1","majortick.mode=arbitrary") 

In CIAO 4.4, set_arbitrary_tick_positions could not be used to set the first major tick mark to be empty. This is fixed in CIAO 4.5, as shown in the Adjusting axis labels gallery example.