Julian Hall
Can you generate an example that is (say) ten times smaller? This is likely to make it much easier to study. Thanks
Thanks, I can see that the overhead of probing is exhibited by the smallest of these. However, I'm currently looking at the LP iteration count "explosion", and it only happens...
Note that `lp_iterations` is a pseudo-measure of work, rather than actual number of simplex iterations performed, so the value quoted here isn't "real"  I guess that this measure of...
We may be OK now, but the "8" problem might still be useful if it's easy to generate, thanks
> The documentation fix is awaiting review in [Documenter.jl](https://github.com/JuliaDocs/Documenter.jl/pulls) Meanwhile, in #2360 I've switched off link checking
> It's not much effort. (Note that it is not the "8" problem, I could also generate one with more/less columns.) Thanks, let's see how @Opt-Mucca gets on
> @jajhall I would be in favour of removing the adjustment factor. > > I guess the intended purpose is to inflate the numbers so that they better translate to...
OK thanks, we'll see what the regression tests make of these changes
> @jajhall I've added a fix for the adjustment factor. The error was that its query to the submip for number of columns seemed to include the columns that were...
> * `to_array` accessor of `readonly_ptr_wrapper` returns wrong values after the first one. > > > ``` > and solution[0] = 14.0 > and solution[1] = 118.0 > to_array(2) =...