Clipper1 compatability .

[J-Dunn]

Hi , thanks for the continued updates. It's good to see so many improvements it merits a new version.

There is no comment in the Readme so is Clipper2 call-compatible with Clipper1 ?

If not, will it require heavy rewriting of code which was based on Clipper1 or just some tweaks (assuming C++ 17).

Some clarification in Readme maybe useful.

In particular , I'm interested in resolving some bugs in FreeCAD which seems to be based on 6.2.0 code. What is the most recent Clipper that I can expect to use to profit from the many bug fixes done here, without making more work that it's worth?

I expect most of the issues are in FreeCAD itself but it seems sensible to eliminate known and fixed bugs in Clipper before going higher up the food chain.

Thanks.

[yoavmil]

#209
I was testing once migration from clipper 1 to clipper 2.
I wrote in in the link above notes as I was going by.
After making it compile, I found 2 issues. one was the inflating lines was differenting from clipper1. the other is that in my usecase clipper2 was slower.
one last thing, I feel still that clipper2 isnt mature enough yet.

[J-Dunn]

Thanks for making those notes available. That at least tells it's doable.

Some of the notes are not too clear what you substitute with what but I guess I'll work it out once I did into the code.

I feel still that clipper2 isnt mature enough yet.

do you mean it is less mature than clipper1 ?

[yoavmil]

Indeed. Clipper2 isn't 100% production ready AFAIK. I follow the issues posted here, and sometimes I see poeple complainig about critical bugs.
Clipper1 has a been then for years, used by many.

[sethhillbrand]

FWIW, I disagree with the assessment of Clipper 2 as not production ready. We (KiCad) had a number of long standing issues with Clipper 1 that could not be resolved due to the architecture. They have been resolved by Clipper2. We have reported two bugs in Clipper2 and both have been fixed quickly.

Clipper 2 is not 100% call compatible. You need to read the documentation where all differences are clearly stated and compared with equivalent Clipper1 ideas.

Lastly, I'd recommend moving this to a discussion rather than an issue.

[philstopford]

I can only echo @sethhillbrand's comments. I have several production tools that were ported to Clipper2 and tracked its progress through to the current day. Angus fixed issues reported and I've not had an issue to report for a long while now. The performance boost has been phenomenal and in many cases, the current system requires fewer workarounds (e.g. for open path clipping in terms of raycasting) than I needed for Clipper 1.

Porting should actually be fairly painless in most cases. Switch the library over and see what the IDE throws up. Fixing it from there is largely just adapting to new syntax. The one minor thing to note is that open path offsetting is per-side, not width, in Clipper 2.

[yoavmil]

OK then, @sethhillbrand and @philstopford , you got me curious. I tested Clipper2 back in September, so alot time passed sinse. I retest Clipper2 soon.

[philstopford]

Good - it would certainly be better to provide an opinion based on direct experience of the current state of the project rather than some 9 months ago. To do otherwise seems unfair, especially when you use strong/provocative words to assert those opinions ('not production ready'). I'm aware that English is not everyone's first language, so this is just friendly feedback - tone matters.

[J-Dunn]

I don't think 'not production ready' is bad tone if it is justified. 9mo is not a long time but maybe few issues have been fixed.

I appreciate all perspectives, even it they are contradictory. I don't think anyone should be getting annoyed. It will be interesting to see what yovamil finds with current version.

I am certainly appreciative of the effort Angus has put into maintaining and improving this over the years. It's one thing to make something and release it as FOSS, it is another to put consistent effort into maintaining a project. It seems sometimes there is a expectation of continued free support once you release code under an open licence, especially a very permissive MIT licence which allows commercial exploitation of free code.