ruby-x/rubyx
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
2,578 Commits 2 Branches 0 Tags 5.6 MiB
f264aec94a
Commit Graph

206 Commits

Author SHA1 Message Date
Torsten Ruger
25d7b8bd83 chaning calling convention not to add frame/arg types 
these types are only needed to debug and can be gotten from the method (also in the mesage)
just saving the 6 instructions for every call
This was made possible through previous commits on fake_memory access
 2018-08-12 14:47:05 +03:00
Torsten Ruger
3d113b4d83 fix more of the changed names 
add a few self.
 2018-08-12 13:10:44 +03:00
Torsten Ruger
393ac873c9 change return sequence to return messages to space 
as it was before blocks
(thought blocks would make reuse of messages impossible, but was wrong, this only appilies to lambdas)
(too) many tests affected
 2018-08-06 14:07:17 +03:00
Torsten Ruger
fd25f997ce reqrote return sequence with builder 
added test _before_ , same code, just much easier to read
in preparation for returning messages
 2018-08-06 13:04:47 +03:00
Torsten Ruger
595e032edf add function_return helper to risc_value 
create a function_return from within builder
 2018-08-06 13:03:33 +03:00
Torsten Ruger
4d4b691a4b adds a return jump 
that will be used to jump to the return sequence
 2018-08-02 17:34:44 +03:00
Torsten Ruger
9c6a099cde block test 
working on return semanitcs
(still avoiding the implicit return)
 2018-07-30 14:10:24 +03:00
Torsten Ruger
a3059108eb compile the blocks too 
method_compilers have block_compilers just like
methods have blocks
But assemblers are a flat list
 2018-07-30 10:27:32 +03:00
Torsten Ruger
4839e0d245 minor naming and spaces 2018-07-27 10:46:22 +03:00
Torsten Ruger
f5c284b3a0 bring the blocks down to mom level 
reusing message_setup, but adding yield specific instructions
 2018-07-24 11:35:49 +03:00
Torsten Ruger
d80ef4bf4e store method, not name, in message 
changes the calling and thus too many test
 2018-07-22 22:29:32 +03:00
Torsten Ruger
4cc04787e9 remove Risc.resolve_to_index 
mostly by using better typed registers,
which cleans up the code where it was used
 2018-07-16 19:00:04 +03:00
Torsten Ruger
3bc85805a4 must pass registers to slot_to_reg and reg_to_slot 
as they are typed, those functions don't resolve on Risc, but the register type
miscother changes from previous commits
 2018-07-15 16:30:50 +03:00
Torsten Ruger
a71a6d34fb use message reg not message 
for upcoming change
 2018-07-15 15:14:38 +03:00
Torsten Ruger
ab8b574e9c also recode the slot_definition resolution 
much like in the slot_load. In fact suspiciously so.
wip, as also changing to storing type (not sym) in reg
 2018-07-14 22:39:00 +03:00
Torsten Ruger
58c7294abd finish the loop in slot_load 
thus rewrite of that old nested if thing is done
 2018-07-14 11:04:21 +03:00
Torsten Ruger
61c840c023 start to move slot_load code to register_value 
the iea is to iterate through register_values while reducing the slot_load to a number of Slot_to_regs
wip
 2018-07-13 21:56:55 +03:00
Torsten Ruger
a095515b0e new get_type_by_class_name helper for space 
sorely needed, with 36 occurences replaced
 2018-07-13 21:50:40 +03:00
Torsten Ruger
edea9ac080 makeing the method of the compiler more or less private 
in preparation for using the same code for bocks
 2018-07-09 19:32:17 +03:00
Torsten Ruger
dd544214b3 start with block_compiler 
as a copy of method_compiler
re-merge later, when we know what's needed
 2018-07-09 16:48:23 +03:00
Torsten Ruger
9005513368 rename typed_method to callable_method 
seems to make the essence clearer
also extracted base class
 2018-07-07 09:11:09 +03:00
Torsten Ruger
d8b58d8da6 fix inital main (again) 
stop even compiling a fake main if a real exists (in builtin)
previous version was still buggy: using builting methods types (especially locals) even comiled version had different
 2018-07-04 09:17:30 +03:00
Torsten Ruger
6f936f190d misc to_s and small fixes 2018-07-04 08:28:29 +03:00
Torsten Ruger
8b9fd29ce9 Only add Builtin methods that are not in the current compilation 
Current bug looks like the original main is used in parts
(not the compiled one, but the builtin on)
 2018-07-04 08:28:05 +03:00
Torsten Ruger
e099014d63 fix dunamic jump in interpreter and misc 2018-07-03 19:15:36 +03:00
Torsten Ruger
bb1d1495db fix constant propagation through the layers 
so they can end up in the binary
 2018-07-03 10:12:22 +03:00
Torsten Ruger
63dfee0978 move some tests around (and fix them) 2018-07-02 23:20:54 +03:00
Torsten Ruger
9bd8eab999 remove class attribute that was left dangling 
not used, which prompted rename. just forgot to delete it
 2018-07-02 15:50:54 +03:00
Torsten Ruger
1132309f6a unify space collection attribute naming 
currently space is still acting as a sort of memory manager.
For proper linking, all objects must be reachable from space, hence the plural versions like messages and addresses (even they are instances, it is the list that is important)
To dish out instance to use, the head must be kept, ie next_XXX for intergers, return addresses and messages
 2018-07-02 15:49:51 +03:00
Torsten Ruger
22408b86c6 return the linker from mom_compiler 
linker holds assemblers
assemblers come from method compilation
and so the layers come into focus
 2018-07-01 21:51:48 +03:00
Torsten Ruger
bb49f1be78 rename ClassCompiler to MomComplier 
ended up not even using the class, it just came from there
It actually compiles methods, and it turns out is the point where builtin comes into the picture as it's boot process also returns method compilers
 2018-07-01 21:26:45 +03:00
Torsten Ruger
e75aef933d have to pass constants around now 
as they are not global anymore
 2018-07-01 14:11:29 +03:00
Torsten Ruger
f7dfa1c45e return assemblers from translation 
result of translate is cpu instructions, our equivalent of assembly.
So return Assemblers for next stage
 2018-07-01 11:56:09 +03:00
Torsten Ruger
05669065ca back to method_compiler 
it is what it is
 2018-06-30 23:26:28 +03:00
Torsten Ruger
208b98d709 start to move translate code 2018-06-30 22:53:32 +03:00
Torsten Ruger
daf1b56062 start on class compiler 
idea is to get cleaner layer seperation
reduce machine and rework builtin boot
 2018-06-30 19:20:17 +03:00
Torsten Ruger
6bd01fd55f rename method_compiler 
in line with other compiler XX_Compiler being the compiler for that layer
remove type from compiler as it is in method available
 2018-06-29 14:48:52 +03:00
Torsten Ruger
86b1edb40c add seperate builders 
remove if with polymorphism for different builders
(easier to understand by naming)
 2018-06-29 13:27:57 +03:00
Torsten Ruger
53107d3ef8 reducing the label's int for return 
placing the raw address in the register to jump to
puts the extra instruction at the risc level (not arm), thus changing a
lot of (brittle?) tests
 2018-05-30 12:54:40 +03:00
Torsten Ruger
0dc89c772a get the label int to work consistently 
still need to use it in the return
 2018-05-30 10:54:18 +03:00
Torsten Ruger
074ec34659 wip, fixed some label, need more fixing 2018-05-30 10:29:38 +03:00
Torsten Ruger
8322fca7b3 give labels an integer that will end up being the position at runtime 
Since integers are first class objects, we need to use an integer object
as the return address. The actual address can not be stored in an
instance variable since it is not an object.
The address is unique to the label and never changes after positioning
(using the int is next up)
 2018-05-29 20:26:00 +03:00
Torsten Ruger
7847420d49 make ints compile time mutable 
to change the label value and use the allocated stubs
 2018-05-29 17:03:55 +03:00
Torsten Ruger
a350325b6b fix function call and simple call logic 
Before creating DynamicJump, the FunctionCall got a register for a
possible jump address. Now that is handled by DynamicJump and
FunctionCall just needs the method, from which it determines the
binaryCode address
 2018-05-19 12:21:20 +03:00
Torsten Ruger
dc8afce444 use existing constant 2018-05-16 12:55:51 +03:00
Torsten Ruger
000bf0a9b3 separate SlotDefinition into own file 
what started as three lines has grown up to demand it’s own file and
tests
 2018-05-15 19:29:06 +03:00
Torsten Ruger
020bce740a more complex while test 
that works, but does leave one wondering what is tested apart from the
result (1)
 2018-04-20 19:38:33 +03:00
Torsten Ruger
1849522a54 fix TruthCheck 
mixing up the false and true, such a basic human mistake
 2018-04-19 10:34:15 +03:00
Torsten Ruger
3a50b7dd0e fix mod4 name 
really did div4
 2018-04-19 10:00:55 +03:00
Torsten Ruger
059ff4a868 pass a source into the builder 
for debugging
 2018-04-18 19:12:30 +03:00
Torsten Ruger
df08cb78e2 adds a lot of to_ssss 2018-04-17 20:26:15 +03:00
Torsten Ruger
d19bd5f98d fix dynamic jump 
which gets dynamic resolve and call to work.
FIRST DYNAMIC CALL working
 2018-04-09 15:06:46 +03:00
Torsten Ruger
59e6298879 first resolved call running though 
and returning an int, as it should
just the value .  .  .
 2018-04-08 23:45:23 +03:00
Torsten Ruger
fabe4db4f6 fix logic error in vool dynamic send 
using receiver of current method
instead of receiver of next message
 2018-04-08 22:59:42 +03:00
Torsten Ruger
580c53cdae fix the resolve 
without return and not being a function
 2018-04-08 22:30:18 +03:00
Torsten Ruger
33ffcf1d88 move resolve_method code from word to mom 
rather make resolve an instruction.
Since it was coded in risc anyway, we not only save the setup and call
But also makes the mom instruction flow clearer
The method really came from not wanting to code it in risc, but with
the Builder, that is now surprisingly painless
 2018-04-08 18:55:17 +03:00
Torsten Ruger
9867234c38 move most code from method_compiler to builder 
rather use builder in two ways, than sometimes compiler and sometimes
builder
Also makes it possible to reuse builtin code in mom’s to_risc, as both
use builder. The builtin code by directly adding to compiler, the mom
code not.
 2018-04-08 18:51:20 +03:00
Torsten Ruger
6958fc31ab rework resolve_method, using builder 2018-04-07 23:07:44 +03:00
Torsten Ruger
26cf911a5c change build names a little and document 2018-04-07 22:35:40 +03:00
Torsten Ruger
db55ec3fd6 slightly changing built return logic 
very tempted to add immediately, but bigger change
 2018-04-07 00:35:54 +03:00
Torsten Ruger
dd0d162ebf fix the init also, was using first message twice 
not advancing after first load
 2018-04-07 00:14:02 +03:00
Torsten Ruger
eea92399ee concludes static message setup 2018-04-06 23:11:08 +03:00
Torsten Ruger
1ddbde1191 return single message for "message" ie r0 
other names, like next_message allocate  a new register
 2018-04-06 22:54:54 +03:00
Torsten Ruger
22409c93ee remove >>, consistent use of << 
makes code easier to read, like assignments
does remind of Passengers
 2018-04-06 22:40:58 +03:00
Torsten Ruger
c042dd9faa allow reverse syntax for >> (ie <<) 
slot >> reg makes sense, being a slot_to_reg
but . . .
consistently use of << (as meaning assignment , =) also makes sense
allow both and let time tell which makes more sense
 2018-04-06 21:05:26 +03:00
Torsten Ruger
41d573d571 wrought one more instruction for message setup 
now with dsl ready, should be faster
 2018-04-06 20:58:58 +03:00
Torsten Ruger
cae5e323ec add reg to slot 2018-04-06 20:21:14 +03:00
Torsten Ruger
c233bd82d6 implement [] for RiscValue for the dsl 2018-04-06 16:08:35 +03:00
Torsten Ruger
e396807ee5 start work on dsl 
so i can read my own code
 2018-04-06 14:21:38 +03:00
Torsten Ruger
d52e14d201 continue to rewrite message_setup 
message come from space already.
next types
 2018-04-05 20:37:03 +03:00
Torsten Ruger
f09086e524 unite the two resolve_to_index functions 2018-04-05 20:10:00 +03:00
Torsten Ruger
f4ce6d6253 starting to redo message_setup 
apart from previous commits bug, it also needs to deal with
dynamic/static correctly
and while at it will do the getting from space
 2018-04-05 12:24:49 +03:00
Torsten Ruger
ee0a1ca823 renaming methods args and frame 
to arguments_type and frame_type, because that is what they are
In honour of setup bug, where the types of those types were loaded,
instead of just them types
 2018-04-05 12:22:14 +03:00
Torsten Ruger
c51fc67ba5 make interpreter tests less brittle 
by factoring __init code away
just counting main
 2018-04-04 20:05:09 +03:00
Torsten Ruger
65d57c8c7c removing unconditional 
just Branch is fine
 2018-04-02 19:30:34 +03:00
Torsten Ruger
a2173645b3 remove the :int shorthand 2018-03-31 19:17:55 +03:00
Torsten Ruger
cb9f6973d9 fix true false and nil constant 
going from mom to parfait
basics in place, more ripples to fix
 2018-03-31 13:47:02 +03:00
Torsten Ruger
9e9b5c7f37 move to parfait integers in risc layer 
loading constants means loading parfait objects
objects have to me collected in machine
integer ok, string/true/false/nil next
 2018-03-31 13:25:59 +03:00
Torsten Ruger
4cc1d8455e fix util namespace 
and instruction move ripples
 2018-03-26 20:05:30 +03:00
Torsten Ruger
bc4d4b428a change boot to new hash syntax 2018-03-25 13:27:15 +03:00
Torsten Ruger
de7e02b0b8 remove IsSame branch from risc 
mixing up levels, IsSame is Mom
at risc this is a minus and Zero check
fix all tests
 2018-03-24 18:54:36 +02:00
Torsten Ruger
ad3e73d931 start on dynamic call test 
fix cache entry being not loaded
test incomplete because of missing resolve_method
 2018-03-24 17:55:01 +02:00
Torsten Ruger
6a538624c5 remove NotSame from risc 
instead use a - b and then  isZero
 2018-03-24 17:54:15 +02:00
Torsten Ruger
0f183b3a74 fix value return and test 
slot load was wrong way around
 2018-03-23 20:02:17 +02:00
Torsten Ruger
6721153456 fix return sequence 
logic error of swapping messages too soon
simplify by folding message unto itself
thus only need one extra register for the address
 2018-03-23 18:58:42 +02:00
Torsten Ruger
a306c464b7 start using tmp registers at 1 
which used to be reserved for the next message
 2018-03-23 18:57:16 +02:00
Torsten Ruger
55832315eb more fix for multilevel constant load 
was ignoring first level which is already the second for a constant
as the constant is the first load.
first interpreter test working but looking dodgy
 2018-03-22 19:14:22 +02:00
Torsten Ruger
e505856af7 fix multi level right slot load 
was done for left, but forgotten for right
 2018-03-22 18:54:07 +02:00
Torsten Ruger
6e901e1718 allow setting the source for slot loads 
so we can track more exactly which instruction created the risc
 2018-03-22 18:45:03 +02:00
Torsten Ruger
ca3bf6acfa fix constants being passed down 2018-03-22 02:38:06 +05:30
Torsten Ruger
01151b4ba7 make continue labels unique 2018-03-21 22:05:51 +05:30
Torsten Ruger
f424e58715 finish the simple call 
moving jump address
 2018-03-21 19:29:00 +05:30
Torsten Ruger
49880267bb start to test call 
as per tdd noticing logic errors, have to swap message out
 2018-03-21 19:20:51 +05:30
Torsten Ruger
b5ef929c9c add method to risc function call 
just so we still know at compile time
 2018-03-21 19:05:53 +05:30
Torsten Ruger
a9196e9cd6 implement simple_calls to_risc 2018-03-21 18:54:42 +05:30
Torsten Ruger
61a801b00c Return to_risc 
remove the index from FunctionReturn, just jump to the register address
 2018-03-21 15:48:50 +05:30
Torsten Ruger
fa797f722d to_risc for NotSameCheck 
which is only used in call cache checking
some fixing, needed to add a abel for the cache check jump
 2018-03-21 12:38:28 +05:30
Torsten Ruger
12c71fa394 first go at message setups translation to risc 
simplest possible implementation, ie the method and next_messages are
loaded many times.

But the layer design shines, it’s easy to understand
 2018-03-21 12:20:55 +05:30
Torsten Ruger
b99fdc3425 rename jump label 2018-03-21 11:52:53 +05:30
Torsten Ruger
d98e55907e first go at translating DynamicCall to risc 2018-03-21 11:51:10 +05:30