Re: [jdt-dev] Null pointers in the java compiler

[Walter Harley <eclipse@xxxxxxxxxxxxxx>]

On Sun, Oct 18, 2015 at 8:32 AM, Stephan Herrmann <stephan.herrmann@xxxxxxxxx> wrote:

JDT core has different null policies in different parts, even within the compiler.

Indeed...  this was one of the challenges even back when we wrote the APT plugins.  What was perhaps obvious to the JDT developers at the time was not obvious to newcomers.

 

Unfortunately, annotation processing may interact with the compiler in ways
that influence the processing order. At least that's the typical APT-related
NPE from my experience.

Very much so.  The JDT compiler was not written to support the back-and-forth between plugins that APT implies; it was like taking a house and trying to turn it into a boat.  You discover a lot of leaks that didn't matter when it was just a house :-)

 

To sum up: for this class of bugs, all it needs is:
- a reproducing example
- Jay and (Sasi or me) sitting down together to figure out the interaction between
  annotation processing and compilation proper

I was one of the devs who contributed a lot of the original implementation, back when it was being driven by BEA.  Although it's been many years since I was actively involved and the details are blurry now, I'm willing to be a historical resource to the extent that I'm able.  I might in some cases be able to answer questions along the lines of "why the heck did those morons do it that way" :-)

Cheers,

  -walter