cjchristopher

@taroxd 's fix works for the version of GS in texlive 2017 for Windows. The install just seems to not set the relevant paths / registry entries for GS (that...

No dice with either/both. Fractional scaling still has the same problems (primary looks fine, secondary just upper right corner), and --pan fails outright with: ``` BadMatch (invalid parameter attributes) Major...

I get different behaviour depending on the state before (resolutions and scaling), but in any case, no matter what state I have things in before, the after state is still...

Right. Seems related to https://github.com/eclipse-theia/theia/issues/6844 as well. I'm digging around Theia and can't see an obvious place to set this for Express - could any contributor point me in the...

I did find some evidence that you can set the base path in Express, I just couldn't locate where in the Theia codebase I would add that call. Although potentially...

Just checking in on whether bounded optimisation is still a goal? I notice a few branches/PRs got closed with the 0.11 release (https://github.com/patrick-kidger/optimistix/pull/143 and https://github.com/patrick-kidger/optimistix/pull/150 - apologies for not getting...

> @cjchristopher thanks for the warning! Can you please point to the bugs? I started with a naive implementation, but over time the zoom linesearch ended up being quite similar...

> Yikes, that's a bit scary. That probably means using papers and other implementations as references here. I don't intend to cause a panic haha. But this is one such...

> Tagging @cjchristopher - this is the routine to identify the generalised Cauchy point, required for `BFGS-B` and `L-BFGS-B`. I'm pretty confident that it does the right thing, but another...

Ah to clarify, that missing clip is in the l-bfgs-b implementation, after zoom is called and returned. I assume an eventual l-bfgs-b implementing for optimistix will correctly enforce bounds both...