Algorithmic debugging of real-world haskell programs: deriving dependencies from the cost centre stack

Abstract

Existing algorithmic debuggers for Haskell require a transformation of all modules in a program, even libraries that the user does not want to debug and which may use language features not supported by the debugger. This is a pity, because a promising approach to debugging is therefore not applicable to many real-world programs. We use the cost centre stack from the Glasgow Haskell Compiler profiling environment together with runtime value observations as provided by the Haskell Object Observation Debugger (HOOD) to collect enough information for algorithmic debugging. Program annotations are in suspected modules only. With this technique algorithmic debugging is applicable to a much larger set of Haskell programs. This demonstrates that for functional languages in general a simple stack trace extension is useful to support tasks such as profiling and debugging.

Lightweight computation tree tracing for lazy functional languages

“…Faddegon and Chitil [12] gave a first formal definition of Gill's technique by extending Launchbury's lazy semantics [17] with observation tracing. That definition is insufficient here, because it omits the request events that Hood provides.…”

“…Hoed-cc [12] is the first algorithmic debugger that works for real-world Haskell programs. Continuous evolution of the Haskell language and the complex implementation of Freja, Hat and Buddha means that these debuggers only support subsets of the language features used in real-world Haskell programs.…”

“…Hoed-cc was the first algorithmic debugger that could handle Raincat [12]. Hoed-pure can also be used to debug Raincat.…”

A Lightweight Hat

“…Hood does not record such chains; it does not need them for observing only values. Faddegon and Chitil (2015) combined Hood's instrumentation with the cost centre stacks of the pro ling system of the Glasgow Haskell compiler to obtain a computation tree for algorithmic debugging of Haskell programs. The implementation hoed-stack uses Hood's event sequence to obtain computation statements as nodes of the computation tree and the cost centre stacks are used to connect these nodes to a tree.…”

Optimization Techniques for Algorithmic Debugging

“…Some advances have been done to improve the scalability of the technique. For instance, in [28] the authors improved the technique by using HOOD to obtain the ET, which avoids them from instrumenting the whole source code. But it does not matter how much the creation of the ET is improved, it still has to be stored somewhere.…”