//
// Picp revision history
//
//-----------------------------------------------------------------------------
//
// Special thanks to Todd Squires, Brian Chamberlain, Scott Johnston, Mike
//  Crowe, Jim Gasbarro, Tim Voght, Ahmi Wolf, and Antonio Augusto Todo Bom Neto
//  for various suggestions and contributions
//
// Many thanks to Jim Robertson (Newfound Electronics) and Wayne Patterson for
// their excellent support and advice in developing version 0.6.0 and higher.
// Thanks to Alexandre Nunes for his modifications for 18xxx devices.
//
// Thanks to Joost Versteegh of Amsterdam, and Rodney Brooks of MIT for their
// help in debugging problems with 18F chips. Special thanks to Dr. Brooks for
// providing hardware with an 18f458 with which to test ISP programming.
//
// Thanks to Eric Lofstad for providing patches for the Windows port.
//
// Thanks to Adam Jurkiewicz for additional information about the processor
// definition data in picdev.c, and for donating a JuPic programmer for testing.
//
// Thanks to Mark Stanley for providing an Olimex programmer for testing.
//
//-----------------------------------------------------------------------------
//
// 0.6.8 (19 December 2005)
//	Read PIC_DEFINITION data from picdevrc file (picdev.c no longer used).
//	Added convert and convertshort programs to read picdev.c file and
//	generate verbose and short forms of picdevrc.
//	Fixed program size for 12F509.
//	Fixed SetRange to set address correctly for older PicStart Plus firmware.
//
// 0.6.7 (08 July 2005)
//	Changed name of picstart.c to main.c since picp now works with three
//	different programmers (PicStart Plus, Warp-13, and JuPic).
//	Modified to handle new protocol for PS+ firmware version 4.30.04.
//	Removed automatic blank check after erase flash function (do blank
//	check as a separate operation).
//	Fixed 12F683 definition data (thanks to Charles Rennolet).
//	Added definition for 18F4550 (thanks to Christian Klippel).
//
// 0.6.6 (20 April 2005)
//	Added support for JuPic programmer. Fixed timing problem with flash erase
//	followed by blank check command. Added new devices to picdev.c and modified
//	pic definition structure to include programmer type support. Picsnoop updated
//	to include new structure definition. Miscellaneous code cleanup and bug fixes.
//
// 0.6.5 (11 January 2005)
//	Fixed long-standing problem with 18Fxxx devices. Should now work with either
//		PicStart Plus or Warp-13.
//	Corrected data in pic defs for 10F200 and 10F202. 10F204 and 10F206 are still
//	probably not correct, but I don't have those chips to test with.
//	Added new command line options.
//	Added support for 10F chips (very unlikely to be working yet!).
//	Save and then restore oscillator calibration data on erase flash and erase
//		program functions.
//	Fixed bug in DoWriteOscCalBits.
//	Added warning message on erase functions other than erase flash. These other
//		functions MAY work on SOME devices.
//	Added command line option to suppress writes for picp debugging.
//	Removed WriteBytes conditional on WIN32.
//
// 0.6.4 (3 October 2004)
//	Fixed DoWriteIDLocs to use byte programming only for 18xxx devices when the ISP
//		flag (-i) is used.
//	DoWriteEepromData now accepts a 'start' parameter to deal with partial
//		eeprom data in hex file. DoWriteData modified to deal with partial data.
//	Makefile can now build a Windows executable (picp.exe) in Linux.
//	Show command line options in log file.
//	Added data eeprom address entries for 12F629/75 and 16F630/76.
//	Changed word alignment for writes for all 18F devices to 8 words (16 bytes).
//	There is a report that USB serial ports don't work in RedHat9. Using the 2.4.26
//		kernel solves the problem.
//	Added -wpx option to suppress actually writing to program space, for debugging
//		picp. Picpcomm.log file will show data that would have been sent.
//	Fixed problem in DoGetProgrammerType and DoGetVersion that caused failure on
//		Mac OS X with slow USB-serial ports (thanks to Jose Commins for the
//		patches).
//
// 0.6.3 (1 August 2004)
//	Calls to WriteHexRecord now pass a blankData value so program reads won't dump
//		out lines to the Intel hex file containing nothing but unprogrammed data.
//	Checksum verification added to GetIntelRecord (in parse.c)
//	Added code to save factory set bits in configuration registers before erasing
//		and to restore them after the erase. This applies only to erasing the
//		configuration registers; the user can still shoot himself in the foot
//		by writing the config regs without setting the appropriate bits.
//	Implemented algorithm for writing 18xxx config bits one word at a time.
//	Added code to write ID locations, configuration bits, and eeprom data that is
//		embedded in a hex file written with -wp.
//	DoWriteConfigBits - apply picDevice->defx mask to remove invalid configuration bits.
//		Pre-read factory set bits before writing.
//	Fixed bug in DoBlankCheck that would report non-blank config bits for some chips
//		when they really are blank.
//	Eliminated picp-isp by making ISP programming a runtime option (-i). This option
//		seems to be needed only when using a Warp-13 with BluePole version 1.5 firmware
//		and 18fxxx devices.
//
// 0.6.2 (15 June 2004) - Jeff Post
//	Added routine to check if a Warp-13 programmer is connected. NOTE: the Warp-13
//		firmware may not support this feature in future versions.
//	Added serial comm line debug option (-c). NOTE: the -c option MUST be the first
//		option on the command line (eg: picp -c /dev/ttyS1 16f84 -ri). If used, the -c
//		option will record all bytes sent/received to/from the PS+ or Warp-13 programmer
//		in a file named 'picpcomm.log'. This file is cumulative--each time it is used
//		it will append data to the log file. The log file can grow quite large unless
//		you delete it occasionally. To report problems with picp, use the -c option to
//		record the comm line data in picpcomm.log, put it in a zip file, and email it
//		to me at j_post <AT> pacbell <DOT> net. Please do not send humungously large
//		log files--try to limit the log file to the minimum that shows the problem.
//	Added code to change character timeout for 18xxx PICs from 1 sec to 5 sec.
//	Fixed dumb bug that forced even number of words when writing program space.
//
//	Thanks to Alexandre Nunes for the following:
//	Add working support for PIC18F252, PIC18F452:
//		- Support for configuration bits with size > 1 word;
//		- Added some data to the device structure to supply
//		   some information not detected from the headers;
//		- Support for word/multiword address alignment (at least 18Fxx2 requires 8 bytes);
//		- Picstart addressing seems to be in byte fashion, not word (size is in words);
//
//		- Modified atoi_base to support octal (JLP).
//
//	Fixed endian issue with configuration data in hex file.
//	Handle comment lines in Intel hex file.
//	Fixed cfgsize error in DoReadCfg.
//	Fixed extended info in PIC_DEFINITION of 16f818/9.
//	Multiple "failed to verify" messages suppressed.
//
// 0.6.1 (18 May 2004) - Jeff Post
//	Added read/write/erase data (eeprom) space.
//	Added erase ID locations and erase configuration bits.
//	Modified DoEraseFlash to explicitly program blank any area that did not
//		get blanked by the CMD_ERASE_FLASH command.
//
// 0.6.0 (14 April 2004) - Jeff Post
//	Added ID location programming.
//	Added compatibility with Warp-13 programmer.
//
// 0.5d (12 June 2003)
//	added erase flash command
//	fixed segfault in WritePgmRange()
//
// 0.5c (12 November 2002)
//	added signal handing; if any signal is received reset the PICSTART and exit
//	added libc++ to Makefile to eliminate link errors in OS X and Red Hat 8.0
//
// 0.5b (3 November 2002)
//	fixed bug which crept back into DoWritePgm() which messed up config word
//	increased BUFFERSIZE to 128k (was 8k)
//
// 0.5a (2 November 2002)
//	updated to match MPLAB 5.70.40 (PICSTART Plus firmware 3.00.40)
//	added all parts from MPLAB 5.70.40
//
// 0.5 (31 October, 2002)
//	added support for more parts (current as of MPLAB 5.40.00 -- thanks Mike!)
//	increased character timeout to 1 second from 1/2 second (helps to eliminate
//		false "no response from PICSTART" errors)
//	added support for extended linear address records in Intel hex files.  this
//		used to fail silently and prevented programming if any byte address in
//		the file exceeded 0x8000 (like certain configuration words)
//	corrected size of 'startAddr' and 'curAddr' (were 16 bit, now need to be 32 bit)
//	added warning if segment address record is encountered in Intel hex files.
//		such records are now ignored; previously they caused the parser to fail
//		silently.
//	added test for buffer overrun when reading hex files.  this was absent before,
//		and writing files larger than 8K bytes would fail.  now hex files may be
//		arbitrarily large.
//	added extended linear address record generation to output when reading device
//	replaced some ioctl calls in serial.c with wrapper functions which makes it
//		more portable (now builds under OS X without modification) -- thanks to
//		Ahmi Wolf.
//	fixed error reporting in main() when you specify an illegal part number or
//		serial port
//	reorganized some of the code.
//	added support for V3 of the PS+.
//	added support for the 16F627.
//
// 0.4d (20 February, 2001)
//	changed CMD_REQUEST_ACK to CMD_REQUEST_MODEL (actually asking for programmer
//		model, not just an acknowledge)
//	dropped baud rates over 115200 from serial.c, they aren't useful here and
//		were limiting cross compatibility
//	added bool as a typedef so picp can be compiled under straight c instead of c++
//	changed Makefile so picp compiles using gcc instead of c++
//	modified OpenDevice() and CloseDevice() to save the previous serial port
//		settings and restore them on exit
//	made some improvements to ConfigureDevice() courtesy Brian Chamberlain
//
// 0.4c (14 July, 2000)
//	added support for PIC18C442 and PIC18C452
//
// 0.4b (14 July 2000)
//	added support for several additional parts (culled from MPLAB 5.00.00)
//	added extended device profile (or something) to initialization -- newer
//		versions of MPLAB send more stuff at device init; these are now sent
//		by picp as well
//
// 0.4a (28 Oct. 1999)
//	improved hex record parsing
//	added timeout to InitDevice() to allow the PICSTART to respond to DTR
//
// 0.4  (1 Oct. 1999)
//	cleaned up some argument handling (better, more generalized)
//	broke picstart.c up into smaller modules
//	added support for -wo flag (only works on single-word calibration spaces)
//	added oscillator calibration space size checking
//	added dump of device info if invoked without any flags
//
// 0.3  (12 Sept. 1999)
//	cleaned up a couple of mistakes in the usage text
//	added support for many more parts (everything supported by MPLAB)
//
// 0.2a (11 Sept. 1999)
//	removed timebits.h from includes.h (wasn't actually used anywhere)
//
// 0.2  (10 Sept. 1999) -- first real release
//