Added shutoff when current is sensed when enabling tool
Also made a number of fixes to current sensing, which was quite broken.
Shorted modbus timeout
Fixed a few issues setting up tooltron on Marvin
Adding init script for mainbox installation
Fixed problems with previous fix
Fixed exception in event add page
Fixed card removal detection
Also switched from using seek command to select command with the reader.Select returns a result immediately, rather than seek which firstreturns an "in progress" error then later returns the card (as far as Ican tell).
Removed pid file necessity
Still uses it if running as root, but otherwise doesn't bother. Shouldbe a bit easier for local testing.
Added clarifying comment for code that confused me
TODO document all this lol
Fixed event queue bug
Implemented code for new RFID readers
It's not fully working yet. When the checksum is enabled, it ignores allresponses. Also, it no longer recognizes a missing card.
Added LRU eviction to cache. Not tested yet.
Fixed build warnings from last commit
Simplified read_file
Added automatic subscription to roboclub-gb on user creation
Minor typo
Added stop command and fixed pid file deletion
Added new view roboauthall
It returns bitmask of all tool permissions for a given user
Major changes to mainbox code.
-Added local cache for RFID -> tool permissions-Queries new page /crm/roboauth/%08x, which gives bitmask of all toolpermissions instead of just one-Added pid file at /var/run/tooltron.pid-Now must be run as "tooltron run" and also has "tooltron refresh" and...
Added new logging functions which inject timestamps
Changed URLs to match actual server
Adding rfid-web
The javascript will query the rfid reader (using main.c, set up ontooltron.roboclub.org with CGI) and insert it into a form.
Removed some of the fields on the add user admin page
Removed rfid_restart flag
This way the AVR will always repoll the rfid reader after the timeout,so hopefully it will never hang. Who knows?
query.c has optional debug output for /add_card_event
Fixed missing import in views.py
Adding program_tool script so we remember how to set tool id
Added environment setup to wsgi.py for usage with apache
Removed migrations from repo
Fixed Add User page error
Added add_card_event view and made some minor changes to models
Event submissions from ./tooltron tested and working
Added username/password to add_event query
Fixed curl SSL queries failing and queue bug
Added query_add_event to mainbox code
Removed a bunch of .pyc files
Added some scripts for working with virtualenv and running the server
Removed virtualenv and added requirements.txt in its place
Now each developer can have their own virtual environment built fromrequirements.txt, without the annoying path problems that come from theenvironment being under source control
Fixed another overflow bug and removed isqrt
Fixed timing in current sensing
The ADC interrupt didn't clear the TIFR timer interrupt flag, whichmeans the ADC wasn't triggered again on the next compare match.
Fixed overflow in current calculation
Switch ADC to timer 1 because timer 0 was taken. Oops.
Untested current RMS calculation
It's not actually RMS, it's variance. Same general idea.
Fixed annoying warnings about avr/signal.h
mainbox now prints error on 404 or 500 from server
Moved tool address to Makefile
Made errors less noisy
If it can't connect to a tool, it will print an error once, and thenprint another message when it is able to reconnect.
Merge branch 'master' of ssh://roboclub.org/home/svn/tooltron
jeff made changes to model and added migrations
updated boards to include holes.
Added getopt for inputting serial device and CRM server
updated urls page.
added tool authentication system. It works!
working on urls. views not working right
changed model to allow for hex rfid cards
added first attempt to do machine authentication
Added query.c for queries from database web interface
customizing user admin
preformed database migrations to fix null issues
added admin interface.
still trying to fix null
fixed not allowing null for db for dates
fixed issue with models for form validation. Allowed fields to be blankthat should be able to be blank.
added unicode functions to models
fixed a couple errors in the models.py. Then preformed a migration.
Finished defining models. Also added function to create RoboUserinstances. Also modified settings to look for RoboUser as the defaultuser class.
Updated user model
created new app roboclub crm. I will be working on user authentication andthe user model
added south for database migration
initial commit of virtual python enviornment and django project. this shouldcreate a fully reproduceable devolopment enviornment. for info seewww.jeffknupp.com/blog/2012/02/09/starting-a-django-project-the-right-way/
Added flushing before write to rfid reader
And other various fixes and changes, which I don't remember. Changedrfid reading to read every tick, but only try writing if enough tickshave passed.
Changed read util to flush input and wait for reads to complete
Moved timing to time.c, and increased period of rfid polling
initial commit of tooltron boards
Added current sensing to toolbox
Tool stops waiting for access response if id is removed
Added c program for reading from rfid reader on computer
Added tool state TS_DENY, with red LED until card is removed
Fixed timer error correction
Made rfid.c more robust against false negatives
Also changed LED code to be more generally usable, and changed mainboxto grant access to a particular user id instead of randomly choosing
Fixes. It tooltrons at a basic level now.
Added LED blinking, and other fixes
Fixed compile errors; finished converting char -> uint8_t
Revert "Removed superfluous printf in rfid read script"
This reverts commit e66d0e5c23eaaf3db53409d05beb0e1fcdbe1610.
Eradicated use of signed chars in serial d'oh
Changed F_CPU to 8MHz instead of 1MHz
Added LED header, and some rfid debugging
Added serial_flush
Added printing the response timeout, and other fixes
- Fixed using SLAVE_ADDR instead of tool->address- Replaced read_registers with read_input_registers
Various fixes to tool main.c
- Input register callback no longer copies current_user to reg_buf prematurely, and then write the uninitialized serno into reg_buf (?!)- tool_disable was missing a ~- Updated coil read/write to use new 1 indexed addresses
Added RTS_ENABLE to Makefile
Also added UL to end of F_CPU
Fixed U2X1 being set in the wrong control register
Changed register/coil addresses to be 1 indexed
Removed superfluous printf in rfid read script
Deleted accidently commited vim swap file
Fixed problem with high error in serial clock
Also available in: Atom