Jump to content


  • Content count

  • Joined

  • Last visited

Community Reputation

0 Neutral

About Pavel

  • Rank
    Super Maniac

Contact Methods

  • Website URL
  • ICQ

Profile Information

  • Gender
  • Location
    Melbourne, Australia
  1. Looks like a bug. Will take a look. Pavel
  2. Plugins have not changed and we don't plan to release 7.41 plugins. You can use 7.40 plugins with this release.
  3. Limited support means that only core information that is necessary to compile and debug for this target is included into system headers and TDF files:- only core registers are defined in the system header files, if you need other registers you need to add your own defines to either your code or system header - full config data is added to the system headers (PIC16) or TDF(PIC18) files - target architecture is fully described in the TDF files but non-core registers and register groups are not. You are welcome to add missing information. To compile it's only necessary to add it to system header files. Missing information in the TDF files is used in debugging under SourceBoost IDE and if you use Mplab or Mplab X you don't need it. For example look at the Port B support that is not defined in the limited support targets but is fully supported in PIC18F8722. This target has the following information in its system header file PIC18F8722.h (used in compilation): ... #define PORTB 0x00000F81 ... volatile char portb @PORTB; .. and in PIC18F8722.tdf file (used for debugging): Configure PORTB { // create PinNames = "RB0|INT0","RB1|INT1","RB2|INT2","RB3|INT3|ECCP2|P2A","RB4|KBI0","RB5|KBI1|PGM","RB6|KBI2|PGC","RB7|KBI3|PGD"; } RegisterSF PORTB { Description = "PORTB",""; Address = F81h; BitNames = "RB7","RB6","RB5","RB4","RB3","RB2","RB1","RB0"; }
  4. The reason we use xlaunch.exe instead of the actual compiler in the MplabX plugin is to allow different programming languages to be used in the same project. Xlaunch analyses the command line and based on the extension of the input source file launched a relevant compiler For example a project can consist of .c, .c++ and .bas files an when compiling xlaunch.exe will use C, C++ and Basic compilers to compile these files.
  5. When you assign an integer to a float like in float fnum = 0x41bba752; you expect compiler to do copy number into the float directly byte by byte. This is not what's happening. Behind the scene compiler generates code like fnum = float32_from_int32(1102817106); so you get the same value but presented as a float.
  6. Use either this: asm { movlw 0x52 movwf _fnum+0 movlw 0xa7 movwf _fnum+1 movlw 0xbb movwf _fnum+2 movlw 0x41 movwf _fnum+3 } or this: union { unsigned long l; float f; } f; f.l = 0x41bba752; fnum = f.f;
  7. Check http://forum.sourceboost.com/index.php?showtopic=1456&do=findComment&comment=20995
  8. Make sure you start preg.exe as administrator and make sure you select the correct product inside the preg.exe dialog (for example a BoostC++ key won't work with BoostC). If this fails try to register from command line. This can be done for any product that is an executable (so this method won't work for plugins and those are DLLs).Do do this: - open command prompt as administrator - navigate into SourceBoost installation directory - start the product executable followed with the "register" command line argument. For example for the BoostC compiler issue the command line like : C:\Program Files (x86)\SourceBoost>boostc_pic18.exe register This will bring up a dialog where you enter your registration name and license key and press OK. If registration was successful you will get a relevant prompt - Again it's important to use these steps as administrator, if not you still may bet an successful prompt but the registration won't be applied properly If still unsuccessful send a mail to support@sourceboost.com and quote your registration name and license key and briefly describe the problem and what you have done. UPDATE: We migrated on a new faster web server and seems that our key generation system got broken If you experience similar problem please email to support@sourceboost.com and we will sort it out. Sorry about the confusion. Regards, Pavel
  9. That's sounds like a compiler bug. We'll look into this. Chameleon does not support C++ classes but has a very good support of data and function templates up to the point when template based meta programming can be done. In fact its code generation logic is based on this. For an example check very bottom of include\sys\sys_assign.h file where meta programming is used to generate code for assigning structs. It also supports auto data type and explicit enum types (like in the latest C++ standard). From the top of my head: - rom type specifier is ignored - dynamic memory allocation is not supported (new/delete) - complex array and struct initialisation works a bit differently - system header system.h does not have to be included (it's included behind the scene) - Chameleon does support bit fields - Chameleon does support floating point data types natively Pavel
  10. You can download support files for PIC18F6xK40 from: http://www.sourceboo...PIC18F6xK40.zip 7.40 also includes instructions how to add support for new targets. Look into the readme.txt files inside include and config directories and how_to_add_new_target.txt file in the root of SourceBoost installation. Pavel
  11. Currently the Chameleon compiler is in beta stage and does not have a corresponding MplabX plugin yet. The only way I can think of how to work around this is to save the BoostC compiler executable in a safe place, rename the Chameleon executable into BoostC and use MplabX with BoostC plugin. The command lines between these 2 is mostly compatible and this makes it possible for this approach to work. Pavel
  12. No difference. GOTO is ok as well. My point was to use built-in assembler rather than C but whatever you use CALL or GOTO doesn't matter. Regards, Pavel
  13. I's tricky to call a function that is not linked to an application. The way to proceed is to use built in assembler like 'call 0x70'. The only issue with this approach is that it's impossible to use call arguments so the solution is to declare a fixed address array to pass data to your bootloader function. Regards, Pavel
  14. Please explain why your method is more preferable? Regards, Pavel
  15. There is no known way to do this under MPLAB And MPLABX should work just fine