This paper explores the generation of conformance test cases for recursive tile systems (RTSs) in the framework of the classical ioco testing theory. The RTS model allows the description of reactive systems with recursion and is very similar to other models like pushdown automata, hyperedge replacement grammars or recursive state machines. Test generation for this kind of infinite state labelled transition systems is seldom explored in the literature. The first part presents an off-line test generation algorithm for weighted RTSs, a determinizable sub-class of RTSs, and the second one an on-line test generation algorithm for the full RTS model. Both algorithms use test purposes to guide test selection through targeted behaviours. Additionally, essential properties relating verdicts produced by generated test cases with both the soundness, with respect to the specification, and the precision, with respect to a test purpose, are proved.TEST GENERATION FROM RTS 533 account its reactions to stimuli of the test cases. In both cases, formal properties of test suites need be considered, for example, soundness reflects that no conformant implementation may be rejected, while exhaustiveness expresses that every non-conformant implementation is detected by at least one test in the suite.When considering infinite state systems, undecidability is often an issue. Very simple models like two counters machines lead to the undecidability of the most basic properties (e.g. reachability of a given configuration and occurrence of a given output). Furthermore, provided the description of a reactive system in a given model, the observable behaviour of such a system may not be expressible in this model. In order to establish properties like soundness and exhaustiveness of a generated test suite, it is convenient to have both a formal description of the system and to be able to prove properties relative to the generated tests. There are several models between finite state and Turing powerful systems; this paper considers a variant of pushdown automata (PDAs), which provide a nice middle-ground between expressivity and decidability. They form a model for reactive recursive programmes, like the running example that represents an abstraction of the one in Figure 1. This example is presented in some Java-like syntax and involves exceptions (a shortcut that is used whenever the keyword throw is used). More precisely, the programme asks for some integer, then calls the recursive function comp, which asks for a boolean, and, depending on its value, proceeds by making a recursive call or stopping. Whenever exceptions are raised, the programme branches directly to the catch block of the main function. Along the paper, we will only focus on the control flow of this programme and abstract data values away.There exist several ways to define recursive behaviours: PDAs, recursive state machines by Alur et al. [9] or regulars graphs, defined by functional (or deterministic) hyperedge replacement grammars (HR-grammars) [10,11]. Each of these models ha...