Hello it's me again.
I come to say few things about my project, I come to create animations having two characters on screen (and you know one is inside otherbut truthfully here my skills and design showed flaws. It just either i couldn't make a good template for mouth both open and close and/or making animation in unity sucks especially with not enough morphs.
So i thought hard how to make animation easy process and i come to one thing, just change my approach, instead of using runtime created meshes from svg I just use runtime created sprites from svg, so basically i change asset i use for it, the test shows that created an single image for character with many clothes takes about 0.1 sec, so if we for each animation create 10 images it will take 1 sec to create spritesheet, of course we need to add time that will take to manipulate xml file(or svg) that we make during runtime changes to base characters (like breast size morphs, elf ears morphs etc.), but it will allow us to create a different shape of mouth if we need for animation, also making easier to create good quality animation. Also this approach allow us to move creating animations/poses from unity to Inkscape. It's has one problem Inkscape doesn't have any animation tools so we need to create it handmade (or use adobe flash but i do not believe it could export animation to svg file but might be wrong).
The other minor problem is quality, as you know it would be difference between vector and raster art, so you might think about it, but for test characters with size 256x512px it looks good enough that i do not have any problems with that.
Of course current algorithms i create for svg should be written from zero since i change the object of manipulation (from mesh to svg data).
But on other hand with current way there are also exists problems one most important, that i could not display characters on UI layer since sorting layer of objects becomes mess. the other is slow loading, so if i use new approach it would only create svg file from scratch and render single file (which i said takes 0.1sec) so i wouldn't need to load it before needing it.
So what do you think i should do go with bigger modifiability which cost rewriting current svg algorithms or just use old approach and have a faith that someone in future will help me with creating animations (As for current i do not believe that i have enough skill to do that in unity, in inkscape i believe i could make a single pose files which should be doable to create spritesheet based animations from them).
ps. One thing i believe similar library to rendering also exist for c++ so if someone wants to try and modify my current algorithm to use by whoremaster it's quite possible.