Bad axis labels when using both plot and layer data

[ASmirnov-HORIS]

Consider the following plots:

from lets_plot import *
LetsPlot.setup_html()

data1 = {
    'x': ['a', 'b', 'c'],
    'ymin': [1, 1, 2],
    'ymax': [3, 2, 3],
}
data2 = {
    'x': ['b', 'a', 'c', 'a', 'b', 'c'],
    'y': [2, 1, 3, 3, 1, 2],
}

ggplot(data1) + \
    geom_point(aes(x='x', y='y'), data=data2) + \
    geom_linerange(aes(x='x', ymin='ymin', ymax='ymax'))

ggplot(data1) + \
    geom_point(aes(x='x', y='y'), data=data2) + \
    geom_linerange(aes(x='x', ymin='ymin', ymax='ymax')) + \
    scale_x_discrete(reverse=True)

Here we see:

• Inconsistent order of axis labels for reverse=True when using both plot and layer data.
• Invalid axis labels when using both plot and layer data.
• Unexpected rotation of axis labels for reverse=True when using both plot and layer data.

[alshan]

Was fixed by recent refactorings.