I've also made an editor/IDE from scratch and it is painful, for example implementing text rendering and Unicode support (emojis) is a PITA and takes a lot of optimization if you want below 1ms screen of text rendering. I however took the approach "every feature is a plugin", so when adding features I first made the core API able to handle the feature then implemented the feature as a plugin. So a few years into development I rarely need to touch the core API and can implement new features with a high level easy to use API in just days.

About adaption, it's very hard to make people switch over to a new editor, so by forking an popular editor you shaved off years of development, and it will be easier for users to switch over. But probably at the cost of implementing new features takes longer and large refactors more difficult.