ICD
Moderators: David Barker, Jerry Messina
ICD
This Is a great Forum. I am just getting started with the compiler and must say I like it a lot.Any chance there is or going to be a Software ICD incorperated into the IDE?I only have the SE version for now.
Regards CharlieM.
Dave has not replied as yet so since this is a subject I have bent his ear over a few times. I will give you my version of his answer.
Yes an ICD will be implemented, not straight away but there will be more than one systems available.
The basic problem is the use of local and global variables. With a language such as Proton you have the ability to map the pic ram to one variable, with variables being reusable you have an issue. If you listed some variables used in one subroutine and your stepping through another you would not expect the first set to start altering with no connection to the code your looking at.
To get round this you have to build a debug file, listing what variables are associated with which sub and module. Chuck in some more info like which source line is associated with which sub/function/module etc and you can build in your ICD an association to the variable. Like if your stepping through a sub you only update the variables used in association with the sub.
Once you have this debug file you can then use that to make an software ICD. A bit more work and you can build a COFF file. This then opens up the use of SF in ISIS and Mplabs.
Currently you can work with ISIS at ASM level in isis and in the pretty near future when you will have a command line interface you will be able to ICD at asm level in Mplabs.
So the answer is yes, you will be able to do it and in many ways. The work just has to be finished first.
Yes an ICD will be implemented, not straight away but there will be more than one systems available.
The basic problem is the use of local and global variables. With a language such as Proton you have the ability to map the pic ram to one variable, with variables being reusable you have an issue. If you listed some variables used in one subroutine and your stepping through another you would not expect the first set to start altering with no connection to the code your looking at.
To get round this you have to build a debug file, listing what variables are associated with which sub and module. Chuck in some more info like which source line is associated with which sub/function/module etc and you can build in your ICD an association to the variable. Like if your stepping through a sub you only update the variables used in association with the sub.
Once you have this debug file you can then use that to make an software ICD. A bit more work and you can build a COFF file. This then opens up the use of SF in ISIS and Mplabs.
Currently you can work with ISIS at ASM level in isis and in the pretty near future when you will have a command line interface you will be able to ICD at asm level in Mplabs.
So the answer is yes, you will be able to do it and in many ways. The work just has to be finished first.