| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
Previously, collections would return a "cycled" version of linewidths which is internally used to scale dashes (this scaling needs to be done at property setting time, because of the interaction with styles). While it makes sense to store the cycled scaled dashes, there's no need to also store the cycled linewidths or pass that to the backend code -- directly using the uncycled linewidths seems simpler. This also makes collection.get_linewidths() return something much closer to whatever was passed in -- see the test change in test_lslw_bcast.
|
If we are to change the return values of these things, I'd like to chip in #23056 I foresee a major rebase effort if this goes in before (although I may be wrong). |
Sorry, something went wrong.
|
I don't really mind either way, though I'd say the change here (making get_linewidths() return something closer to what was passed in) is more useful than the change in #23056 (which makes get_linestyle() return something different from what was passed in, if I understand it correctly). (Notwithstanding the other benefits of #23056.) |
Sorry, something went wrong.
| Back | FazBrowse Home | New Git URL |
Previously, collections would return a "cycled" version of linewidths which is internally used to scale dashes (this scaling needs to be done at property setting time, because of the interaction with styles).
While it makes sense to store the cycled scaled dashes, there's no need to also store the cycled linewidths or pass that to the backend code -- directly using the uncycled linewidths seems simpler. This also makes collection.get_linewidths() return something much closer to whatever was passed in -- see the test change in test_lslw_bcast.
Also, "broadcasting" has a technical meaning in numpy; it's not the same thing as cycling.
PR summary
PR checklist