Tony Hirst

Re: the multiple canvases, this could also be complicated by how to arrange the canvases; eg size by side, all on top of each other if they are the same...

Ah, thanks... I started exploring in another context and think I was getting a race condition too? On Wed, 26 Feb 2020 at 14:36, Aaron Watters wrote: > I got...

Thanks... one thing I notice if I run this is that I get the `Uninitialized Proxy Widget` warning as if things were loaded twice?

Looking at discussion in https://github.com/mwouts/jupytext/issues/691 re: the use of timestamps as the basis for syncing, I think that some git branch changing and checkout operations can lead to situations where...

Re: the `pre-commit` hook, I have been experimenting with that and it's a bit of a faff in the Mac Github application. With the simple approach (eg add `jupytext --sync...

Hi Re: pre-commit - I'm pretty sure I was using the pre-commit framework recipe in the docs which was breaking for me, compared to the original "naive precommit" which seemed...

@JohnPaton Ah, okay, thanks. Makes is less than useful for me in this use case, which is trusting that Jupytext works correctly to create pair and add documents in an...

@johnPaton Adding more plumbing could lead to more confusion though. I think the stumbling block in current workflwo for me is that the Github Desktop client at least doesn't register...

I've just tried the `pre-commit` action again in a Github Desktop App environment, and it's completely unusable: various errors are raised when trying to commit with suggestions of various `git`...

@FeryET If you use the original (deprecated) approach (`jupytext --sync --pre-commit` line in a "traditional" `.git/hooks/pre-commit` file), things seem to work fine.