|
| 1 | ++++ |
| 2 | +title = "Place Capability Systems" |
| 3 | +date = 2025-11-24T19:00:00+01:00 # (Paris/Zurich) |
| 4 | ++++ |
| 5 | + |
| 6 | +Rust's novel type system has proved an attractive target for verification and |
| 7 | +program analysis tools, due to the rich guarantees it provides for controlling |
| 8 | +aliasing and mutability. However, fully understanding, extracting, and |
| 9 | +exploiting these guarantees is subtle and challenging: existing models for |
| 10 | +Rust's type-checking either support a smaller idealised language disconnected |
| 11 | +from real-world Rust code, or come with severe limitations in terms of precise |
| 12 | +modelling of Rust borrows, composite types storing them, function signatures and |
| 13 | +loops. |
| 14 | + |
| 15 | +In this talk, Zack will present *Place Capability Graphs*: a novel model of Rust's |
| 16 | +type-checking results, which lifts these limitations, and is directly calculated |
| 17 | +from the Rust compiler's own programmatic representations and analyses. The PCG |
| 18 | +model supports almost all real-world Rust code (97% of Rust functions in the |
| 19 | +most popular public crates), and is suitable as a general-purpose basis for |
| 20 | +verification and program analysis tools. |
| 21 | + |
| 22 | +**Presenter**: [Zack Grannan](https://zackg.me) is a PhD student in the [Software |
| 23 | +Practices Lab](https://spl.cs.ubc.ca) at the University of British Columbia, |
| 24 | +working under the supervision of Alex Summers. His research focuses on Rust |
| 25 | +program verification, and he is also involved in the development of the |
| 26 | +[Prusti](http://prusti.org/) program verifier. |
| 27 | + |
| 28 | +**Meeting Link**: TBA |
| 29 | + |
| 30 | +**Recording Link**: TBD |
0 commit comments