Professional VisSim 6.0 Fix List
-----------------------------------------
This document is organized into three sections.
Section I lists problems fixed in Professional VisSim. 
Section II lists problems fixed in VisSim Real-TimePRO.
Section III lists problems fixed in VisSim C-Code.

Section I: PROFESSIONAL VISSIM
-------------------------------

VisSim Bugs Fixed in 6.0a (Released 10/14/05)
--------------------------------------------
o If plot/bezel scrolled above top window in display mode, then display mode was exited, blocks shifed down a few pixels
o printing failed to print wires
o rotated buttons didn't paint right
o integrator gave wrong results if inside conditional compound and static solver activated but no constraint or unknown blocks present
- fixed in A2
o Integrators inside conditional compound with static solver could run at multiple of proper rate
o bezel blocks grew in size if in display mode and user entered and exited compound block
- fixed in A3
o restored interface to ISA bus RT to allow proper DIO operation
o copy of FIR filter lost band info and could cause GPF
- fixed in A4
o keep bezel size the same across display mode settings
o Mathcad/OLE object 1 dim matrix output outputs caused GPF
- fixed in A5
o ctrl+right click on transcendental caused GPF
o over 1024 user menu items gave GPF, now checks and warns of limit and bumped limit to 4096
- fixed in A6
o var with no input connected to screen edge parent compound pin caused GPF on level entry
o set type of hex const to int, had been double
- fixed in A7
o Bezel with bitmap grew bitmap slightly on entry to display mode. Now no change.
o Button text was not centered properly on entry to display mode.
- fixed in A8
o set type of enums defined in startup .c file to int (went to unknown in A6)
o added EXPORT32 CMATRIX *PASCAL EXPORT cmatReDecl(CMATRIX *m, unsigned dim1, unsigned dim2);
o updated vissim32.lib and vsuser.h for above
o added WM_DESTROY message send to user DLL vsmEvent()'s upon VisSim exit
- fixed in A9
o codegen of compound containing variable with removed input connector caused bad code
- fixed in A10
o Print preview in tiled mode for diagram that spanned > 1 page was garbled
o Left mouse click now activates buttons in display mode
- fixed in A13
o smoothed analyze freq response
o tighter 1/z & SH fixed point codegen for const clock
o negative step if bounds/step gives over 4e9 steps
- fixed in A14
o plot traces made thicker when printing
o analyze gave empty plots after Bode or rootlocus was deleted and then recreated if reuse plots enabled
o codegen of pulseTrain for DSP target ignored delay field
o Monitor Buffer Read sometimes gave "Can't find buffer"

VisSim Bugs Fixed in 6.0B (Released 8/28/06)
--------------------------------------------
o expression and state transition blocks failed to give correct value for scale_int variables
- fixed in B2
o Maximize of child window caused bleed of other window contents
o Open of nonexistant file caused flaky behavior after "Can't open" message
o Help/Contents gave "Help file not found" on first try
o Autorestart failed to propagate reference to $runCount
- fixed in B3
o doubly nested conditional compounds caused bogus "algebraic loop" error
o open of empty file then close caused strangeness
o screen repaint event sometimes caused hidden digrams to "bleed" through to top diagram
- fixed in B4
o $firstPass codegenned to '1'
- fixed in B5
o added VR_GET_IMPLICIT_SOLVER_RUNNING vissimRequest() message.
- fixed in B6
o codegen for abs of 32 bit fixed point gave wrong value.
o 1/Z outside of enabled compound sometimes failed to clock if 1/Z inside enabled compound did clock
- fixed in B7
o unmaximize of plot sometimes failed to paint connector wires
o Find tag after plot maximize caused all wires to disappear
o only relevent subplot labels appear in subplot (previously all trace labels would appear in all subplots)
o resize of plot using top edge caused wire schmutz
- fixed in B8
o Radix point propagation during codegen sometimes failed when 1/Z in feedback had old radix point from prior use.
- fixed in B9
o enabled subsystems got wrong value of input arg when targeted at fixed point CPU
o enabled subsystem with 1/Z inside would sometimes give wrong results if a dependency on an external 1/z existed
- fixed in B10
o import block counted the 'e' in 1e10 as a delimiter and gave too many columns in matrix
o delayedSwitch block ignored dialog settings
- fixed in B11
o codegen for conditional compound that contained user blocks could fail to properly declare data structures
- fixed in B12
o reload of .m file that initialized variables could cause GPF
- fixed in B13
o leading zero property of display block was not saved to .vsm file
o Bezel block property dialog file field would not scroll for long names
- fixed in B16
o codegen of scaled int xfer func could have wrong scaling
- fixed in B17
o codegen of reset integrals could give wrong results for integrators after 1st (ballxxx.vsm)
- fixed in B18
o codegen of conditional compound with "num=den" xfer func input could give "use before def" warning on compile

VisSim Bugs Fixed in 6.0C (Released 5/13/07)
--------------------------------------------
o rt mouse while moving selected blocks, backed up a level and confused display
o del of connected var input, failed to remove logical connection
- fixed in C2
o codegen of local vars in stateTransition block gave errors
- fixed in C3
o leading '!' character in data file was not treated as comment char. It was in v5
o pow of complex base gave wrong answers
- fixed in C4
o Optimize with DSPinterface block gave error on 1st restart
- fixed in C5
o pow with complex base scalar exp was bad
- fixed in C8
o pow with one input gave GPF
o optimize with DSP crapped out after st iteration
- fixed in C9
o optimize fix in C8 failed to read unknown initial val on opt start
- fixed in C10
o codegen of fixed point pulse train with external pin for time set to $timeStep was off by 1/2 
- fixed in C11
o codegen of fixed pt filter in feedback with discrete integrator could give wrong answer
- fixed in C12
- Allows timed licensing
o got bogus "pow overflow" error when using min/maxElement
o Codegen of scaled-int transfer function, could give floating point results
o codegen of float gain fed by fixed point block gave wrong conversion value.
- fixed in C13
o sim of relational that had been connected to 2 scaled ints, could give quiet overflow error and highlight in red when reconnected to float types.
- fixed in C14
o sim or codegen caused diagram to think it needed to be saved
- fixed in C15
o scaled fixed point tranfer function could give wrong simulation and codegen results if input pin data type mismatched
- fixed in C16
o lower case DLL spelling in registry could cause failure to properly setup addon DLL resulting in failure to save/restore addon state changes
- fixed in C17
o codegen of scaled int xfer function with scaling mismatch on input gave wrong scaling of output
o DSP failed to wait for PC communciation before starting
o "sign" block would not convert to scaled fixed point
- fixed in C18
o interrupts were started before hotlink connection, now waits for connection
o codegen of "return 0;" in "make callable from user app" caused compiler error for some compilers
o fixed point scaling was off when merge block had operands with different fixed point scaling (introduced in C15)
- fixed in C19
o reference to undefined local var in state transition block caused "OOPS", now gives diagnostic
- fixed in C20
o convert of int to scaled int use a float multiply
o Ref of local var in state transition could give erroneous "undefined" message
o Codgen marked .vsm file as modified causing "Save changes" dialog on exit

- fixed in 6.0D
o Use of undefined symbol in const block failed to give warning
- fixed in 6.0D2
o histogram autoscale sometimes gave glitching
o histogram title was ignored
- fixed in 6.0D3
o only gave 6 digits of precision for integrator IC, now gives 15
o menu directories were case dependent alphabetic order, now case independent
- fixed in 6.0D4
o updated const blocks with expressions after 1st local opt iteration
- fixed in 6.0D5
o Added "noAudioAutorepeat" .ini file option to suppress .wav file looping mode for Light blocks
- fixed in 6.0D6
o after 6.0D4, sliders sometimes failed to update output values after range was altered  
o added vissim.ini setting "ignoreNonNumericData=1" to ignore import data lines starting with nonnumeric data
- fixed in 6.0D7
o codegen of unit delay clocked by expression that required autoconvert could cause ref to undefined temp var "t0"
- fixed in 6.0D8
o codegen of diagram with "make callable from user app" selected would prefix function with INTERRUPT. Now removed.
- fixed in 6.0D9
o time delay sometimes caused sim abort when buffers grew/ then shrank/ then grew again over 2x
o abs of -1 for for 1.n fixed point resulted in -1, which is correct, but failed to flag red as overflow

=================================
Section III: VISSIM ECD
=================================

VisSim Bugs Fixed in 6.0 (Released 9/9/05)
--------------------------------------------
o support for SB MSP430 LCD 

- fixed in A2 10/17/05
o f2812 eZdsp occasional "did not start on DSP" msg. Increased pause from 20ms to 50ms before "Awake response" check.
- fixed in A4 10/17/05
o use of 32 bit fxDiv gave wrong answers due to failure to include DMC32.h
- Build 16
o eventCapture was missing from 243, 240x and 281x targets
- Build 162
o added support for long, unsigned long, char and short typed pins on dspInterface block
o added GPIO MUX selection for F280x periph blocks
- Build 361	10/13/06
o Shrank RAM used by flash_pie_init
o allocated .switch segment to same as .text (could cause case block to hang on DSP)
o MSP430 had FLL faults for large crystal multipliers over 0x40, caused serial baud rate skew
o Added string Tx to serial write
o disabled serial interrupts when adding/removing serial queue items to avoid race condition
o only bring in uart1_isr's if used to use less RAM
o allowed Radix point of 0 for fixed pt const
o limited upper bound of ePWM period to 65535
o MSP430 sometimes had wrong CPU speed for high crystal multiples
o MSP430 serial TX required presence of SCI status block to work.
o read/reset interrupt block genned wrong code for F280x
o added i2c support for F280x and MSP430
- Build 602	2/19/07
o added CMPCTL setting for ePWM CMPA/B load event for F280x
o TBCTRn had wrong spelling in regs24x.h file
o Interactive ePWM mode F280x had limited duty cycle from 0-50%

=================================
Section II: VISSIM REAL-TIME PRO
=================================

Real-TimePRO Bugs Fixed in 6.0 (Released 10/01/2005)
-----------------------------------------------------
o Support added for analog out modes of PMD-1208xx
o USB PMD devices speed up 

=================================
Section III: VISSIM C-CODE
=================================
Bugs Fixed in 6.0a (Released 10/14/05)
--------------------------------------------

o Added MAP_TABLE_MEM_SECTION decl macro for mem placement
o Added feature to generate code in top down order of screen placement