SX/B includes / arrays
threepointone
Posts: 10
I've been working on a relatively big SX/B program the past week, and I encountered a couple of problems:
a) Is there any way to include code snippets into sx/b? I wrote a sort of library for an LCD and a couple of other components, and the file is a bit unmanageable at the moment. "INCLUDE" will only include stuff directly in assembly, without compiling it.
b) Of course, with a program this big, I need arrays. I've started using them, but I've noticed that there's a couple of places where they just don't work. I know that you can't use them as loop variables in for loops, but they seem to also not work if you're making a "custom" for loop (i.e. do, if var = 0 then exit, inc var, loop) without any compiler errors. The behavior becomes really, really strange (my program works, but it does some really strange things with the LCD--this was in a loop for writing characters). I assume that for some reason banking isn't done properly in loops with arrays without returning any compiler errors? Does this also happen with "goto" loops, and does it only apply for counter variables?
Thanks for the help, and sorry if I'm asking anything stupid--I just started working with the SX this week.
a) Is there any way to include code snippets into sx/b? I wrote a sort of library for an LCD and a couple of other components, and the file is a bit unmanageable at the moment. "INCLUDE" will only include stuff directly in assembly, without compiling it.
b) Of course, with a program this big, I need arrays. I've started using them, but I've noticed that there's a couple of places where they just don't work. I know that you can't use them as loop variables in for loops, but they seem to also not work if you're making a "custom" for loop (i.e. do, if var = 0 then exit, inc var, loop) without any compiler errors. The behavior becomes really, really strange (my program works, but it does some really strange things with the LCD--this was in a loop for writing characters). I assume that for some reason banking isn't done properly in loops with arrays without returning any compiler errors? Does this also happen with "goto" loops, and does it only apply for counter variables?
Thanks for the help, and sorry if I'm asking anything stupid--I just started working with the SX this week.
Comments
a) Use LOAD for SX/B files, INCLUDE for ASM files.
b) Array elements cannot be used for certain things. For example as a FOR...NEXT control variable, or as an index to another array. If you post your code we could probably figure out what you need to do. Banking is all handled by the compiler and you should be able to make a loop using an array element as you describe. Again please post your code and we can help you out.
Bean.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
www.iElectronicDesigns.com
·
However, I did notice something interesting going on in the assembly listing.
[noparse][[/noparse]code]
· 1378· =00000019······ LCD_TempStr·· EQU· 0x19········· ;LCD_TempStr·· Var···· Word
· 1379· =00000019······ LCD_TempStr_LSB· EQU· LCD_TempStr
· 1380· =0000001A······ LCD_TempStr_MSB· EQU· LCD_TempStr+1
· 1381·················
· 1382· =0000002C······ LCDVars······ EQU· 0x2C········· ;LCDVars·············· Var···· Byte(5)
· 1383·················
· 1384· =0000002C······ LCD_Temp1···· EQU· LCDVars+0···· ;LCD_Temp1···· Var···· LCDVars(0)
· 1385·················
· 1386· =0000002D······ LCD_TempHex·· EQU· LCDVars+1···· ;LCD_TempHex·· Var···· LCDVars(1)
· 1387·················
· 1388· =0000002E······ LCD_TempHexChar· EQU· LCDVars+2· ;LCD_TempHexChar······ Var···· LCDVars(2)
· 1389·················
· 1390· =00000030······ LCD_TempCharCnt· EQU· LCDVars+4· ;LCD_TempCharCnt······ Var···· LCDVars(4)
· 1391·················
· 1392· =0000002F······ LCD_TempChar· EQU· LCDVars+3···· ;LCD_TempChar· Var···· LCDVars(3)
· 1393·················
[noparse][[/noparse]/code]
The variable that was causing problems, LCD_TempCharCnt, happens to overflow to the next bank in the array. I'm not really sure how that'd mess up with indirect addressing, but it's something interesting to note.
By the way, I'm running on an SX48.
here's the actual code that appears to be causing the problem. The problem that occurs is too consistent to describe without all of the code. It might have to do with the I2C bus commands somewhere, as putting pauses in certain places in the main code magically made everything work again (a whole lot slower, of course). Again, changing LCD_TempCharCnt to a global variable fixed everything in the code. Unfortunately, my SX-Key is a bit defective and won't debug anything (I'm doing an RMA after this project is done), so that makes things a bit more difficult to figure out. A lot of the functions require an actual I2C device to be connected, so I can't really do this in SXSim
[noparse][[/noparse]code]
' parameters: character (Byte), # of characters (Byte)
' LCD_I2Cxxxx commands are elsewhere in the code
LCD_DispChar:
·LCD_TempChar = __Param1
·If __ParamCnt = 1 then
··LCD_TempCharCnt = 1
·else
··LCD_TempCharCnt = __Param2
·endif
·LCD_I2CStart LCD_DataMode
·Do
··LCD_I2CSend LCD_TempChar
··LCD_TempCharCnt = LCD_TempCharCnt - 1
··if LCD_TempCharCnt = 0 then exit
·Loop
·LCD_I2CStop
·Return
[noparse][[/noparse]/code]
(/edit)
Alright, I've trimmed down on everything extraneous (there is a lot of code in here that's specific to how the sensor and LCD I use respond using the I2C protocol, just ignore that). I tested it again, and it's completely noticable--when the character count is a byte in an array, the program starts doing strange things (specifically, the measured capacitance will never update even though it's clearly not stuck in some outside loop and will still respond to button presses)
I've attached a heavily condensed version of the code, and I'm pasting the specific code regions which seem to be the culprits.
Sorry if the code's messy/unoptimized; I actually discovered this problem while I was trying to tidy it up and stuff everything I could into the array variable space, as I was running low on RAM before.
A condensed form of the code is attached--basically, this dispChar subroutine is called by two other high level routines, RetHome (which basically writes spaces to the LCD and returns the cursor) and DispHex, which converts a variable into hex ascii.
I'm completely stumped at the moment. Although I've been able to fix it, I really want to know why array variables won't work here--I don't see any obvious reasons. I've messed around with the order of the variables (since the variable causing problems just happens to be at the edge of a bank) and it didn't do anything. For some odd reason, adding pauses in the main code that calls all of the LCD writes makes it work (in a really strange manner)--this leads me to think that I've messed something up with my LCD_I2CSend routine and the acknowledgement bit, but I've been looking through the assembly and I can't seem to find anything wrong.
Thanks for your assistance!
Post Edited (threepointone) : 1/31/2008 10:32:09 PM GMT
I've been trying to convert my sensor code to use arrays, and I'm having similar problems. Is there some sort of conflict with using the I2C commands and array variables?
Post Edited (threepointone) : 2/1/2008 5:56:58 AM GMT