Generate-and-validate (G&V) automated program repair (APR) techniques have been extensively studied during the past decade. Meanwhile, such techniques can be extremely time-consuming due to the manipulation of program code to fabricate a large number of patches and also the repeated executions of tests on patches to identify potential fixes. PraPR, a recent G&V APR technique, reduces such costs by modifying program code directly at the level of compiled JVM bytecode with on-the-fly patch validation, which directly allows multiple bytecode patches to be tested within the same JVM process. However, PraPR is limited due to its bytecode repair design and it is basically unsound/imprecise as it assumes that patch executions do not change global JVM state and affect later patch executions on the same JVM process. In this paper, we propose a unified patch validation framework, named UniAPR, to perform the first empirical study of on-the-fly patch validation for state-of-the-art source-code-level APR techniques (in fact, almost all G&V APR techniques except PraPR fix at the source-code level); furthermore, UniAPR addresses the imprecise patch validation issue by resetting the JVM global state via runtime bytecode transformation. We have implemented UniAPR as a publicly available fully automated Maven Plugin. Our study demonstrates for the first time that on-the-fly patch validation can often speed up state-of-the-art source-code-level APR by over an order of magnitude, enabling all existing APR techniques to explore a larger search space to fix more bugs in the near future. Furthermore, our study shows the first empirical evidence that vanilla on-the-fly patch validation can be imprecise/unsound, while UniAPR with JVM reset is able to mitigate such issues with negligible overhead.