Commits

Author Commit Message Labels Comments Date
Nzen
Project submitted It turned out that Timothy's Dhcp implementation won out, so I am glad he made it. This morning I changed Client to permit the user to enter an address or let one be automatically generated. I also refactored much of the code in all three so they are easier to understand. (Just remembered that I should have put Dhcp's try in run() and had everything else throw the IOException, oh well.) I endange…
Nzen
Timothy's server Couldn't connect again, so he sent his changes via email. He also sent 'mature' drivers and mocks, but other team members are responsible for those aspects. So they've been left out.
Nzen
Representative counter action WithSyncOnly displays intended behavior: threads find an untimed counter in a shared array, decrement from it, and search for another one. The run pasted in DataCorruptor shows that the threads can take different orders if looking before the releaser starts looking. As noted, fulfilling our DHCP spec utilizes a similar sequence. An analagous class would notify client when it choseCounter() and in…
Nzen
Sync toy looks better, moved to package Moved them to a package since they are for our reference rather than relevant to the project. I may try out a concurrentSkipListMap later. Committing at this point because this algorithm is too simple to simulate the dhcp operations, but is still clear enough to identify what to replace. In the next phase, WithSyncOnly threads will search a shared array of counters for those at zero and then slee…
Nzen
Added multithread/static_DS toy, unsatisfied To test my hope that a static data structure, wrapped in a syncronization layer, could accommodate various mutations, I made a toy & driver: ThreadwithArrList. I will check how referenced DS handles multiple thread mutations as this one didn't seem to change it as hoped. ThrWarrList makes a bool list and then should set two of the values to true (list initialized with false, & true at head). It t…
Nzen
Code adapted, untested Untested because I've nothing is implemented. I guess I'll test the template by itself against the client and have it echo later. Otherwise this is ready for hacking. However, one coordination remains: the data structure's location. Will the threads initialize with a structure housed in the server or use a static data structure? I don't recall the consequences of each at the moment.
Nzen
Pre-skeleton, template commented Initializing project with the unchanged 4B5B classes with the multithreaded template included within block comments. I'll clean this up before I sleep tonight but the repository shouldn't have to wait. It seems as though the ds for the table will need to be concurrent as the template had each reading and writing in seclusion. Oh well, at least the collections can easily adopt a concurrency wrappe…