Page History

...

Looking at the page now, some of the things listed above are not part of nb-javac anymore (e.g., cancelling and some part of the error recovery paragraph) but it still gives a reasonable overview and is good to get an idea what kinds of things nb-javac does. 

What is the reason that nb-javac is a fork and not part of the javac main repository?

1. The pace at which javac/JDK moves and at which NetBeans moves is simply different. E.g., imagine a refactoring breaking a user's code because of error recovery misbehaving. Once the problem is found it may be fixed in nb-javac in hours or days. Fixing it in javac can easily take weeks or even months (and much more effort in testing and review). 
    
2. Even some changes that from the NetBeans point may look innocent may be hard to get into vanilla javac. For example, when javac detects a syntax error, it normally ends without getting to a semantic analysis. This does not work for an IDE (noone probably wants to lose almost all features just because a semicolon is missing). But it may be difficult to push through a fix to semantic analysis for a syntactically broken code (sometimes this passed before, other times, there was a pushback on such changes, depending on the nature of the change needed). 
    
3. Some of the changes in nb-javac may be very hard to get in unless there is another user than just NetBeans.