Mateusz Paprocki

> It looks like the DataRange1d evaluates all the given data if not given explicit renderers (using renderers keyword). It is not specific to the data per y-range. That's the...

> (...) unless a parallel codepath for canvas-rendered legends is retained alongside the new DOM legend rendering, which would be an extremely heavy maintenance burden. This is exactly how current...

I think that I have an unfinished branch laying around for this, though I need to check this. The main reason I didn't expose it in the Python API thus...

``` I think that we should simply use isympy -c qtconsole. **Labels:** CodeInCategory-Code CodeInDifficulty-Medium ``` Original comment: http://code.google.com/p/sympy/issues/detail?id=2632#c2 Original author: https://code.google.com/u/101069955704897915480/

``` **Owner:** --- ``` Original comment: http://code.google.com/p/sympy/issues/detail?id=2632#c3 Original author: https://code.google.com/u/101069955704897915480/

This is probably duplicating earlier reports of collapsing layouts (marked as `tag: layout`), but I need to verify this.

I've been stewing over this for days and I don't have a good solution for now, at least not one I could commit to. I think the best approach would...

I think that at this point this shouldn't be rushed and designed properly in 3.0.x or 3.1. As shown in PR #12494, extensions can be upgraded easily without this, because...

No, not yet. I intent to add stylesheet related models as part of fixing this issue.

> Is there anything captured here that is not already part of your plans? If not, we can close. Sorry, missed this. Yes it's all covered by the existing issue...