Spinfer is a tool for transformation rule inference being developed as part of the ANR-NRF funded project "Automatic inference of software transformation rules for automatically back and forward porting legacy infrastructure software". This is a joint project between Sorbonne University, in the Whisper team at LIP6, and Singapore Management University.

Spinfer targets inferring transformation rules for use with the Coccinelle program matching and transformation tool for C code. The goal of Spinfer is to be able to infer transformation rules that rely on the interaction between separated fragments of context and transformed code that are connected by control-flow and data-flow relationships. The project mainly focus on porting of Linux kernel device drivers, but will also consider the applicability of the developed techniques to Java code.

Spinfer is currently work in progress. We hope to have an initial release in 2019.


Fast and Precise Retrieval of Forward and Back Porting Information for Linux Device Drivers
Julia Lawall, Derek Palinski, Lukas Gnirke, and Gilles Muller.
USENIX ATC, July 2017, Santa Clara, California.