kicad-developers team mailing list archive

[jp charras <jp.charras@xxxxxxxxxx>]

Le 05/06/2019 à 19:00, Wayne Stambaugh a écrit :
> 
> I wouldn't mind doing some more testing as well given the complexity and
> potential issues with this change.  Having a public repo to clone from
> would make life easier than patch sets given the amount of code changes.
> 
> Wayne
> 

Hi Wayne,

The amount of code *actually* involved in the zone filling algo change
is very small: perhaps less than 30 lines of code ( 60 to 80 lines
including comments).

Of course my commits are bigger, but "large" changes are mainly in .fbp
files and are cosmetic changes.
The actual change in panel_setup_feature_change is only adding two
wxCheckBox to enable/disable the new fill algo.
All others are cosmetic changes (I hope: enhancements).

The main risk is not in our code (30 lines of code directly related in
zone_filler.cpp) : it is in Clipper Inflate transform code:
We are using inflate since a long time in our code with no issue, but
mainly to inflate convex polygons.
The risk is for polygons having acute angles between 2 edge segments, in
some corner cases we don't yet know. (well: "corner case" is the right
word here...).
I am confident with Clipper library. As I said previously, it is the
best library we used for polygon handling.
However we cannot exclude some annoying bug for us.
Only many tests made by many users on many boards will give us the answer.

I just committed my last changes (the 30 risky lines of code).

Most of my tests were related to "how to chose the number of segments in
arc to polygons transform", and this is the reason I needed the Seth's
work on arc to polygon changes.

In fact, removing the thick outlines in solid areas was a basic and
minor work.


-- 
Jean-Pierre CHARRAS