Manu Sridharan

> 1. We definitely need to keep the ECJ-based Java source front end working with the latest WALA, as that is being actively used in current projects. I think we...

I am almost done with the work to complete this task locally. However, having done the work, I now wonder whether a different solution might be better. Rather than moving...

My WIP is here: https://github.com/msridhar/WALA/tree/shrink-maven Getting things to work in-repo is basically a pre-requisite for moving the projects out of the repo, so I'm going to continue working on that...

Compiling these projects already works with Gradle. The key problem is that the tests in `ide.jdt.test` and `ide.jsdt.test` are Eclipse plug-in tests, which need to run inside a running Eclipse...

Glad we documented all this stuff before! When I get time I will look into creating the separate Git repo. Dealing with tests will be a bit tricky due to...

So here is what makes things tricky. First, the tests in both `com.ibm.wala.cast.java.ecj` and `com.ibm.wala.ide.jdt.test` depend on the `testFixtures` configuration of `com.ibm.wala.cast.java`. I _think_ we could turn this into a...

A note: there is _some_ code in `com.ibm.wala.ide.tests` that is runnable under Gradle and that I would like to keep working with the latest WALA master; see #965. We may...

From your expected result, it seems you want to do a _forward_ slice from your seed statement, to see what else depends on it. Did you try that?

@161250029 my guess is that your pointer analysis is very incomplete. In particular, if the points-to set for `connection` is empty, you might see issues like this. This will depend...

I would first check the call graph and confirm it’s an issue. Did you do that? Here is some documentation on entry points: https://github.com/wala/WALA/wiki/Pointer-Analysis#entry-points