Review queue
Why a queue and how to use it?
- The queue is intended as a TODO list for Malte. We hope that it can be a useful tool for Malte to keep track of who wants want and to keep the waiting times reasonably fair.
- Use the queue for code reviews, discussions and other Fast Downward things that Malte should work on.
- As of now, the queue is intended to be used by the Basel AI group only. If you think this should be changed, please get in touch.
- Add new items to the end of the queue. Usually, Malte will work from the front of the queue, but there may be reasons to promote/demote things.
There is no need to put everything into the queue. It is intended for things that you would ideally like to get done right now, but that are not terribly urgent.
- The queue should be easy to use and useful for everyone involved. If it turns out that this is not the case, we should get rid of it again.
The queue
http://issues.fast-downward.org/issue26: integrate preprocessor in search code (Gabi)
http://issues.fast-downward.org/issue629: task interface for pruning techniques (code review, Florian)
http://issues.fast-downward.org/issue509: task interface next steps (discuss long-term plan, Florian)
http://issues.fast-downward.org/issue667: Integrate SCCs based merge strategy (code review, Silvan)
http://issues.fast-downward.org/issue668: Integrate dynamic MIASM merge strategy (code review, Silvan)
http://issues.fast-downward.org/issue688: replace GlobalOperator* by operator ids in successor generator and for preferred operators (code review, Florian, Jendrik)
http://issues.fast-downward.org/issue693: new hash function (use mini-benchmark to reduce the overhead of new hash function, Jendrik)
http://issues.fast-downward.org/issue719: clean up option parser (code review, Jendrik)
http://issues.fast-downward.org/issue638: bugfix in systematic pattern generation (code review, Florian)
http://issues.fast-downward.org/issue707: M&S: make pruning of unreachable and irrelevant states configurable (code review, Silvan)
http://issues.fast-downward.org/issue693: new hash function (come up with implementation strategy, Jendrik)
http://issues.fast-downward.org/issue705: Reduce memory usage of successor generator (code review, Florian)
http://issues.fast-downward.org/issue725: Use OperatorID instead of GlobalOperator* more often (code review, Florian)
http://issues.fast-downward.org/issue726: Change StateRegistry::get_successor_state from GlobalOperator to OperatorProxy (code review, Florian)
http://issues.fast-downward.org/issue198: issues when using multiple heuristics (added by Malte, to be discussed with Salomé)
http://issues.fast-downward.org/issue718: merge heuristic into evaluator (added by Malte, code review for Manuel)
http://issues.fast-downward.org/issue717: Lama-Synergy: get rid of the synergy object (added by Malte, code review for Cedric)
http://issues.fast-downward.org/issue694: new hash table implementation (code review, Jendrik)
http://issues.fast-downward.org/issue742: Fix a bug in M&S (code review, Silvan)
http://issues.fast-downward.org/issue700: RootTask class (code review, Florian)
http://issues.fast-downward.org/issue668: Integrate dynamic MIASM merge strategy (code review, Silvan)
http://issues.fast-downward.org/issue739: Add meaningful exit codes to translator, change search exit codes (code review, Silvan)
http://issues.fast-downward.org/issue731: remove legacy hash functions (analyze experiment results, Jendrik)
https://bitbucket.org/malte/autonice/pull-requests/1: Python autonice (code review, Jendrik)
http://issues.fast-downward.org/issue749: allow putting plugins into sections in the documentation (code review, Jendrik)
http://issues.fast-downward.org/issue740: Make translator output file configurable (code review, Guillem)
http://issues.fast-downward.org/issue766: convert Heuristic options to Evaluator (code review, Manuel)
http://issues.fast-downward.org/issue774: move axiom evaluator and state packer into task interface (code review, Florian)
http://issues.fast-downward.org/issue198: issues when using multiple heuristics (code review, Salomé)
http://issues.fast-downward.org/issue786: prepare code base for code from saturated cost partitioning papers (code review, Jendrik)
http://issues.fast-downward.org/issue790: fix various minor problems in headers (code review, Jendrik)
http://issues.fast-downward.org/issue784: use clearer and/or more consistent parameter names in a few places (code review, Jendrik)
http://issues.fast-downward.org/issue791: Use task interface for StateRegistry (code review, Florian)
http://issues.fast-downward.org/issue792: Get rid of g_successor_generator (code review, Florian)
http://issues.fast-downward.org/issue793: Introduce/collect utility methods for printing proxy objects (code review, Florian)
http://issues.fast-downward.org/issue794: Reduce code duplication in axiom evaluator (code review, Florian)
http://issues.fast-downward.org/issue802: add support for time limit and computing partial abstractions in general (code review, Silvan)
http://issues.fast-downward.org/issue849: separate merge-and-shrink algorithm from heuristic (code review, Silvan)
http://issues.fast-downward.org/issue802: add time limit for merge-and-shrink (review by Florian done, proper printing of time stamps done (cf. discussion @bitbucket), waiting for confirmation to merge/further reviewing if deemed necessary, Silvan)
http://issues.fast-downward.org/issue780: add code from saturated cost partitioning papers (think about how to change cost functions in projections, Jendrik)
Discuss changing the interface of PatternInformation and PatternCollectionInformation (my current idea is to let these classes own the result of pattern (collection) generators which is then passed onto the user of such generators) http://issues.fast-downward.org/issue911 and http://issues.fast-downward.org/issue906 (Silvan)