Thomas Duvernay
Thanks, @zachjweiner, it is true that the discussion you linked is the one that made me think of this `alt_segment` argument. It took me forever to finally get to write...
Hi @nmayorov, Thank you for looking into this PR. As for providing an example of the problem when `order` decreases, the new test implemented in this PR does exactly that....
`t_span` is given as an argument to `map`: [`t0, tf = map(float, t_span)`](https://github.com/scipy/scipy/blob/2e5883ef7af4f5ed4a5b80a1759a45e43163bf3f/scipy/integrate/_ivp/ivp.py#L512). Therefore, it just needs to be a 2-member iterable, which means that any sequence object would in...
@laurent90git Thanks for linking to the bug I had reported, I appreciate it. @Sbozzolo I am really happy to see that someone else has encountered the same issue as me....
@Sbozzolo Did you also add the `if` condition as shown in the screenshot above?
I can still see one mistake; you should have: `order = iwork[13]`
There could be more to it than what you described. `brentq` includes the bracketing values in the root-finding process. So, say the initial-value problem integration goes from `t1` to `t2`,...
But `t_old` has to be `t` of the previous step. So why did not `brentq` find `t` of the previous step as a root?
I believe this issue should have been closed by gh-14552.
This issue should have been closed by gh-14552.