Safe Commits for Transactional Featherweight Java

Abstract

Concurrency is a ubiquitous phenomenon in modern software ranging from distributed systems communicating over the Internet to communicating processes running on multi-core processors and multi-processors. Therefore modern programming languages offer ways to program concurrency effectively. Still, writing correct concurrent programs is notoriously difficult because of the complexity of possible interactions between concurrent processes and because concurrency-related errors are often subtle and hard to reproduce, especially for safety-critical applications. This thesis develops and formally investigates different static analysis methods for various concurrency-related problems in concurrent object-oriented programming languages to guarantee the absence of common concurrency-related errors, hence contribute to the quality of concurrent programs. My PhD period has been so far the most exciting, enjoyable, and memorable part of my life. I have learnt and grown so much both in knowledge and life. I was given a lot of opportunities to study and do research. I have met many interesting people and learnt so much from them. Now, at the end of my thesis time, first of all I would like to thank my main supervisor, Martin Steffen. I am greatly thankful for being his student. His broad and deep knowledge has opened a new world for me in the field of formal methods. His love and passion for research have influenced me a lot and kept me always motivated during 4 years of my PhD. It is known that during a PhD, anyone will at some point reach a low-point of motivation and enthusiasm. And I was not an exception. Sometimes I really wanted to give up, but when I looked at him, I said to myself "shame on me", because I have not seen any moment where he is bored with research, even when he is exhausted and could not move his body anymore but if one asks him a research question, he will suddenly become energetic again. From him, I have learnt things from basic to deep levels, how to appreciate one's own effort, how to be open-minded to new ideas, how to be more professional in working, how to think and do research independently. From him, I have learnt to be precise, concrete, and very honest when doing research. I could never expect anything better than that. As a leader of the group, Olaf Owe is a great supporter both in research and personal life. Olaf created all possibilities for me to access research resources like supporting me to go to conferences and schools where I met so many interesting people and my eyes were opened for new things. I always got back with so much passion for research. He always tried to share and find solutions for problems and difficulties I had to face from time to time. He is also the person who taught me how to keep a good balance between life and work. Also Einar Broch Johnsen has been very helpful and supportive for me. I benefited very much from his clear vision and direction in research, from his deep knowledge and his strict attitude in doing research and in writing papers (and in not tolerating ba...

Safe Locking for Multi-threaded Java

“…As mentioned, the work presented here extends our previous work [12], dealing with transactions as a concurrency control mechanism instead of locks. The extension is non-trivial, mainly because locks have user-level identities.…”

“…Furthermore we plan to implement the system for empirical results. The combination of our two type and effect systems, one for TFJ [12] and one for the calculus in this paper, could be a step in setting up an integrated system for the applications where locks and transactions are reconciled.…”

“…A static type and effect system for TFJ to prevent unsafe usage of transactions was introduced in [12]. This paper applies that approach to a calculus which supports lock handling as in Java 5.…”

“…Furthermore, the analysis needs to handle dynamic lock creation, aliasing, and passing of locks between threads. As transactions have no identity at program level and are not re-entrant, these problems are absent in [12]. Fortunately, they can be solved under reasonable assumptions on lock usage.…”

Software language engineers’ worst nightmare

“…Featherweight Java [33] is a minimal core calculus for Java, proposed to facilitate understanding the consequences of extensions and variations, proposed almost two decades ago. Featherweight Java has received substantial attention since then, with a definition of its denotational semantics [65], small-step operation semantics [23], type-preserving compilation [42], received proposed extensions like Generic FJ [79], Feature FJ [4], Corecursive FJ [2], Transactional FJ [70], Feath-erTrait [44], Welterweight Java [52], Middleweight Java [9], Featherweight Wrap Java [8], and perhaps others.…”