Fix potential dip in profile plot when using Gaussian blur + minor display issues

[atsju]

fix #333

Problem was due to Gaussian blur not using same inclusive/exclusive boundaries as cicle mask code (second commit)
I also changed some float to int truncature into rounding as small epsilon could cause issues with sin/cos (first commit).

Fix #319
Fix #317

[github-actions]

🚀 New build available for commit 99c48c9
Download installer here

[github-actions]

🚀 New build available for commit 4e983c2
Download installer here

[github-actions]

🚀 New build available for commit a7ebba5
Download installer here

[gr5]

Something wrong with github - these "download installer here" links don't work and if I go to actions to get them from there - also not working. Instead I have to build these myself. No big deal but I've heard that github has lots of new bugs related to AI somehow?

[atsju]

Very strange. I can download the installers (I did not even build once on my own computer this week).
I didn't hear of new bugs but sure GitHub business model is changing for enterprise users.

[atsju]

It's probably easier you just wait for it to work again.
Eventually you can check if you can download artifacts form a pull request openned by yourself. And I will try your artifact. This would cross check if it's an issue with rights/privilege management.

[atsju]

DFTFringe/surfacemanager.cpp

Line 638 in 7b78459

double radm =wf->m_outside.m_radius + outsideOffset-2;

DFTFringe/surfacemanager.cpp

Line 641 in 7b78459

rado += (insideOffset + 1);

From my understanding the arbitrary shrink comes from there and this is the "other place" to mess with.

[gr5]

Yes. Probably. It's tough to know if one caught all the spots.

[gr5]

Please don't merge if we are doing a release where I don't have a chance to test every feature in DFTF.

I tested the hell out of this feature (odd and even diameter wavefronts, averaging wavefronts, subtracting, ronchi, foucault, saving and loading wavefronts) and it looks fantastic.

But it's a little crazy how changing one feature can affect a feature that is seemingly unrelated so I'd like to test everything else e.g. "test stand astig removal" on the next release with this code (again - it shouldn't be affected but I've gotten burned before).

[gr5]

To be clear, I recommend we merge unless we are doing a release in the next 8 days.