This commit is contained in:
Irmen de Jong 2023-08-01 22:49:55 +02:00
parent 548721e306
commit f2c62bee7e
4 changed files with 24 additions and 24 deletions

View File

@ -106,7 +106,7 @@ class TestCodegen: FunSpec({
Files.deleteIfExists(Path("${result.name}.asm")) Files.deleteIfExists(Path("${result.name}.asm"))
} }
test("64tass assembler available? - if this fails you need to install 64tass in the path") { test("64tass assembler available? - if this fails you need to install 64tass version 1.58 or newer in the path") {
val command = mutableListOf("64tass", "--version") val command = mutableListOf("64tass", "--version")
shouldNotThrowAny { shouldNotThrowAny {
val proc = ProcessBuilder(command).start() val proc = ProcessBuilder(command).start()

View File

@ -8,6 +8,12 @@ Porting Guide
Here is a guide for porting Prog8 to other compilation targets. Here is a guide for porting Prog8 to other compilation targets.
Answers to the questions below are used to configure the new target and supporting libraries. Answers to the questions below are used to configure the new target and supporting libraries.
.. note::
The assembly code that prog8 generates is not suitable to be put into ROM. (It contains
embedded variables, and self-modifying code).
If the target system is designed to run programs from ROM, and has just a little bit of RAM
intended for variables, prog8 is likely not a feasible language for such a system right now.
CPU CPU
--- ---
@ -17,8 +23,8 @@ CPU
Memory Map Memory Map
---------- ----------
zeropage Zeropage
========= ========
#. *Absolute requirement:* Provide three times 2 consecutive bytes (i.e. three 16-bit pointers) in the zeropage that are free to use at all times. #. *Absolute requirement:* Provide three times 2 consecutive bytes (i.e. three 16-bit pointers) in the zeropage that are free to use at all times.
#. Provide list of any additional free zeropage locations for a normal running system (BASIC + Kernal enabled) #. Provide list of any additional free zeropage locations for a normal running system (BASIC + Kernal enabled)
#. Provide list of any additional free zeropage locations when BASIC is off, but floating point routines should still work #. Provide list of any additional free zeropage locations when BASIC is off, but floating point routines should still work
@ -33,11 +39,11 @@ RAM, ROM, I/O
#. what part(s) of the address space is RAM? What parts of the RAM can be used by user programs? #. what part(s) of the address space is RAM? What parts of the RAM can be used by user programs?
#. what is the usual starting memory address of programs? #. what is the usual starting memory address of programs?
#. what is the best place to put a page (256 bytes total) of scratch area data in RAM?
#. what part(s) of the address space is ROM? #. what part(s) of the address space is ROM?
#. what part(s) of the address space is memory mapped I/O registers? #. what part(s) of the address space is memory mapped I/O registers?
#. is there a block of "high ram" available (ram that is not the main ram used to load programs in) that could be used for variables? #. is there a block of "high ram" available (ram that is not the main ram used to load programs in) that could be used for variables?
#. is there a banking system? How does it work (how do you select Ram/Rom banks)? How is the default bank configuration set? #. is there a banking system? How does it work (how do you select Ram/Rom banks)? How is the default bank configuration set?
Note that prog8 itself has no notion of banking, but this knowledge may be required for proper system initialization.
Character encodings Character encodings
------------------- -------------------
@ -57,8 +63,7 @@ The more the merrier.
Floating point Floating point
============== ==============
Prog8 supports floating point math *if* the target system has floating point math routines in ROM. Prog8 can support floating point math *if* the target system has floating point math routines in ROM. If that is the case:
If the machine has this:
#. what is the binary representation format of the floating point numbers? (how many bytes, how the bits are set up) #. what is the binary representation format of the floating point numbers? (how many bytes, how the bits are set up)
#. what are the valid minimum negative and maximum positive floating point values? #. what are the valid minimum negative and maximum positive floating point values?
@ -71,11 +76,10 @@ The most important libraries are ``syslib`` and ``textio``.
``syslib`` *has* to provide several system level functions such as how to initialize the machine to a sane state, ``syslib`` *has* to provide several system level functions such as how to initialize the machine to a sane state,
and how to warm reset it, etc. and how to warm reset it, etc.
``textio`` contains the text output and input routines, it's very welcome if they are implemented also for ``textio`` contains the text output and input routines, it's very welcome if they are implemented also for
the new target system. the new target system. But not required.
There are several other support libraries that you may want to port (``diskio``, ``graphics`` to name a few). There are several other support libraries that you may want to port (``diskio``, ``graphics`` to name a few).
Also of course if there are unique things available on the new target system, don't hesitate to provide Also of course if there are unique things available on the new target system, don't hesitate to provide
extensions to the ``syslib`` or perhaps a new special custom library altogether. extensions to the ``syslib`` or perhaps a new special custom library altogether.

View File

@ -1,6 +1,9 @@
TODO TODO
==== ====
- add a mechanism to put the original p8 source lines into the generated assembly as comments (not only the line numbers).
- add a mechanism to pass the original p8 source lines into the p8ir file as comments. Remove the position xml tags.
- IR: reduce the number of branch instructions such as BEQ, BEQR, etc (gradually), replace with CMP(I) + status branch instruction - IR: reduce the number of branch instructions such as BEQ, BEQR, etc (gradually), replace with CMP(I) + status branch instruction
- IR: reduce amount of CMP/CMPI after instructions that set the status bits correctly (LOADs? INC? etc), but only after setting the status bits is verified! - IR: reduce amount of CMP/CMPI after instructions that set the status bits correctly (LOADs? INC? etc), but only after setting the status bits is verified!

View File

@ -1,23 +1,16 @@
%import textio %import textio
main { main {
ubyte[2] data = [100, 100] const ubyte ATTR_FALLING=$04
uword dRef = &data const ubyte ATTRF_EATABLE=$80
sub start() { sub start() {
dRef[0]-- void test(10,20)
dRef[1]++
cx16.r0L = 0
cx16.r1L = 1
dRef[cx16.r0L]--
dRef[cx16.r1L]++
txt.print_ub(data[0])
txt.spc()
txt.print_ub(data[1])
repeat {
} }
ubyte[10] attributes
sub test(ubyte tattr, ubyte tobject) -> bool {
return tattr&ATTR_FALLING==0 and attributes[tobject]&ATTRF_EATABLE
} }
} }