re-ordered the list of artists returned by legend_handler.HandlerErrorbar
#2532
Conversation
|
Is there a way to write an easy test for this? |
|
It would be easy to test if the returned artist is a I suspect that this band-aid which hides an underlying picking issue when dealing with over-lapping artists (which may be un-fixable). I also suspect that is a down-payment on a more substantial re-work of |
| artists.extend(handle_barlinecols) | ||
| artists.extend(handle_caplines) | ||
| artists.append(legline) | ||
| artists.append(legline_marker) | ||
|
|
There was a problem hiding this comment.
Does this result in 2 newlines? (its hard to tell with github). If so, would you mind removing one of them?
There was a problem hiding this comment.
good catch, I also pep8'd the entire file while I was at it.
|
rebased onto current master. |
|
re-based again, my local master wasn't really pointing at upstream master, sorry about that. |
6caa4ac to
a256a2c
Compare
a256a2c to
6882780
Compare
|
I am going to abandon this PR. |
This is to address an issue raised on SO http://stackoverflow.com/questions/19470104/python-matplotlib-errobar-legend-picking.
As near as I can tell,
create_artistsis only ever called byHandlerBase.__call__()which then drops all but the first artist, which is returned by__call__and added tolegend.ledgendHandles.Re-ordering the artist makes picking on them behave better (see the SO question at the top).
This technically breaks the API, but I would be surprised if anyone is actually using this (yes, http://xkcd.com/1172/).