Repository navigation
Backport light deflection - #10885
Conversation
…e_ldsun" This reverts commit 30d8abb.
|
Thanks @mhvk! Lets see how the tests do before I look more closely. But re:
from #10666, I think it's best to leave this as-is with the revert as that way we don't have to re-write public history. (And it reflects the reality of how this developed anyway!) |
|
@eteq - the failures seem unrelated. |
|
And, yes, good to keep this as is, including the history. |
|
Also: I just checked: the commit backports cleanly to 4.0.x (except the changelog entry, which goes in 4.0.3 for some reason). Shall I push that to 4.0.x directly? |
|
Thanks for the fixed backport, @mhvk! I see now how intertwined these are but this seems to do it. Re: v4.0.x, Not yet @mhvk - we should leave that to the next 4.0.x releaser (which I think will be @astrofrog). Also, sorry, but this raises one more question I want to bring back to #10666 (just trying to put the discussion there instead of here so it's all in one place). To be clear, though, if we do agree to do the 4.0.x backport there, whoever does the backport should backport #10666 and then this, and that should do the trick. |
This reverts the earlier attempt to backport #10666, and then adds a single commit that includes corrected versions of #10666 and #10881. This single commit is likely easy to backport to 4.0.