Conversation
…delays to keep the IO from hanging
…ts since the daily count may not be what we think it is when the clear data command is sent, it looks like the hourly data pointer is correctly reset but the daily data is not. In addition, when iterating through to the end of the daily data count after a delete seems to provide bad data at some point in the past. I will continue more testing to see if i can figure out the behavior.
…for more than 9 days ago
My branch has fixes for numerous pedometer bugs with the format, includes all the data read by the windows client and includes a delete flag. I've now been using it exclusively for months and the data appears to be 100% correct. I still occasionally get a hang in some of the lower level IO code on Mac OS. Killing the csv read and trying again almost always works the second time. I tried adding a sleep which seems to help, but not completely eliminate this issue. I think when reads happen too fast there are data errors and the internal state machine gets out of sync with the code's expectations. Sample Process says things like: Call graph: Call graph: Call graph: |
All merged. May have to go in and fix the unistd thing in the CSV util, since that'll kill windows compatibility, but for now, eh. Thanks for much for the additions, I'll add you to the credits and all. :) |
Thanks! I had fun hacking on it and I'm very appreciative of your reverse engineering effort. I can get rid of the unistd stuff, i suppose i can parse a single "-d" without using getopt ;) i didn't even think about the windows compatibility. |
Hi,
I managed to add the code for the daily data. I changed the struct a bit because the time is actually just for aerobic steps, not daily total. I figure this shouldn't break anyone since the data wasn't filled in anyway (and you're pre 1.0 anyway ;)
I have the device and I tested in in Mac OS 10.6 and it seems to be working. I will test Linux as well later tonight. I will probably also push some test code later as well. This is my first pull request so I hope this is right. If you need any other help with the pedometer code, I'm happy to help.
Thanks,
Scott White