How to cancel copy propagation during the process of converting Java bytecode into sdg file?

[Andres981231]

I want to perform a taint analysis using Joana, and the sinks I annotate are the SDGNodes with the kind of expression. However, I found that Phi Nodes can not be annotated as source or sink. As the figure showed below.

Moreover, I have also found that the Phi nodes in sdg file seems do not correctly record its position in source code (red block) and bytecode (green block). I wonder that whether this is a special implementation or just a bug in Joana.

Then I studied a specific case, the source code and the sdg file are showed below.


I have noticed that the assigment instructions of variables like max and richer are taken as PHI Nodes (blue block) and they are seperated from the Instructions (red block). However, the original assignment instructions after two if statement instructions are missed (green block). I think that these instructions are likely to be eliminated by copy propagation optimization during the process of converting Java bytecode into sdg file. So I want to cancel this in order to get assigment instructions and annotate them as sink instead of annotating PHI nodes. I wonder whether the joana team is aware of where do they perform the copy propagation and can give me some hints. Thanks a lot!