Michael Kraus

Not surprising, the same kind of problem occurs if there is a closing square bracket in a code snippet within some class, e.g. ```` .someClass[ ``` ] ``` ] ````

Splendid! Thanks a lot!

Adding an empty julia block ```` ```julia ``` ```` to the beginning of the first slide gets rid of the error, but produces an output file, where - the second...

> I’m actually surprised OrthogonalPolynomialsQuasi.jl works at all: I don’t remember implementing kronecker products with it. Yes, in fact I patched this together in a very hacky way myself... >...

Sounds good. Is this available somewhere or can you send it to me (together with the aforementioned thesis)? Thanks!

I spent considerable time trying to understand this behaviour, but it is utmost puzzling. I can identify the code where the `__pyx_t_npy_float64_complex`type is used, but if I isolate it, I...

Update: It seems I could identify the reason: generated functions! The [aforementioned file](https://github.com/DDMGNI/GeometricIntegrators.jl/blob/master/src/integrators/stochastic/integrators_wirk.jl) has two of them. A makedocs run for just this file takes 88s. If I comment those...

I refactored my code and got rid of all the `@generated` functions (something I had anyhow planned). Now the whole makedocs takes about 700s. Which is much faster than before...

In my cases, Documenter always finishes, but a run that takes one or two minutes with Julia v1.1 can take up to one or two hours on Julia v1.2 -...

The latter, unfortunately. I wasn't able to pin this down further. But indeed, refactoring our code, removing many of the generated functions, sped up the doc build quite dramatically. Not...