Initialization between `part()`s and `Morph`s are different

[linusha]

See #673 for an example.

[linusha]

See #760 as well.

[merryman]

I believe this is mainly due to the style policies applying themselves immediately after the morph is initialized. This may (1) cause some weird behavior in the morphs when created and (2) is also quite performance intensive. I believe we can adress this issue, once we move to an implementation of part() which does not need to apply the policies right after the morphs are created.

[linusha]

I believe this is mainly due to the style policies applying themselves immediately after the morph is initialized.

Not sure if you refer to the same thing and/or if you saw this, but when originally discovered, we put the following description of this problem:

// This is horrible, but a functioning workaround for the current state of the way that masters are applied.
// The problem arises from the fact, that the initial constructor call does not come with all specified properties, but some are applied later.
// This messes up our logic wrt how textmorph styling in `backdocument()` is working.

[merryman]

Not sure if you refer to the same thing and/or if you saw this, but when originally discovered, we put the following description of this problem:

Yes this is exactly the same thing.

[merryman]

This is now fixed due to #1696.