Commit Graph

408 Commits

Author SHA1 Message Date
e61c5d4a55 Simplify Parfait booting
Since some weeks, Parfait uses instance variables instead of generated attribute getters (that needed type)
This makes it possible to simplify the boot process, getting rid of separate boot Space and class versions.
It is still quite order dependent, but all "normal" ruby code, (less magic) so easier to understand.

Also moved all code that can never run at runtime into the adapter. This included Space and Object new, space as the space will only ever be created at compile time and object, since that is quite different at run-time (which is where i am working towards)
2019-09-22 19:10:47 +03:00
a496ea7e4b Derive Class and MetaClass from Behaviour
Finally! If i remember the module was from before i had any non Object superclasses
unified code, unified bugs :-) just makes sense
2019-09-21 18:50:33 +03:00
fd46826b9c Implement class instance variables
as they are just the type of the meta_class, that was relatively simple.
I feel this is what oo is meant to be, instance variables and methods for the objects, and since classes are objects, for them too.
Class variables seem like a design mistake, weird scoping rules and no data hiding (left as an exercise to the reader)
2019-09-19 15:48:27 +03:00
b0d1948800 Moving space to object class
away from Parfait module, as that gets collapsed
Leaving shortcut outside parfait for now
2019-09-18 22:36:56 +03:00
38491d120b removed unused NamedList
args and locals got inlined into message, forgot to delete then
ripples out due to type creation
small type class api change, more ripples, but also more consistent
2019-09-18 22:07:05 +03:00
e56db0a3ac get method name out from method_missing 2019-09-17 20:18:00 +03:00
b46512a1b8 tests for mom check instructions 2019-09-15 19:57:15 +03:00
7ee57f2b08 generalize get_main and get_init to get_method
to get at those know methods that really
__must__ exists, hence the bang, raise if don't
about to add method missing and raise to the list
2019-09-15 12:58:43 +03:00
b36ba42990 Test complied parfait tests
this makes it obvious that we need a working raise
and a correct method_missing, so we can diagnose the 
resulting errors
2019-09-15 12:18:31 +03:00
4bf23defc8 fix many tests with preloading
preloading, something akin to builtin, loads some very small predefined (macro) methods for the tests to work (ie call)
2019-09-12 22:27:26 +03:00
5ea91df4c1 Integer macros tests and defs 2019-09-11 19:23:56 +03:00
72643ebb08 Fix last parfait bug
missing return statement (duh)
2019-09-10 14:49:02 +03:00
63323376e4 use more instances in parfait
and misc
2019-09-10 12:33:57 +03:00
81e3c0c270 moving to instance variables in parfait 2019-09-09 20:26:54 +03:00
0ae7c5d8aa hacking method not found
was using exit, since raise is not implemented. This was ambiguous as all programs exit.
Using :died as special kernel code and bending it, and reporting it in interpreter.
2019-09-09 11:47:37 +03:00
6811fc4174 fix interpreter to output symbols
which amazingly lets us get at classnames etc
2019-09-08 15:31:03 +03:00
b13c19def3 cache booted functions
remove more redundant parfait boots
2019-09-07 17:56:06 +03:00
7d92ee9e6a add a statistics command to compiler
just to see how many objects make up a binary
2019-09-05 13:25:40 +03:00
02261ad79d changing factory size per factory
Before it was one class variable, but ints and messages are not created in equal amounts.
2019-08-24 09:46:33 +03:00
1eb6430880 last fixes for cc
seems like setup had just worked by chance and change broke that.
2019-08-23 19:22:27 +03:00
89f32ed74a Changing the call setup and return to be more efficient
Now we are using a statically linked list of messages. This will not work with procs, but that can be solved then.

Previous (wrong) thinking was that because of procs messages have to be allocated for every call. This was too slow, and not neccessary
2019-08-23 15:30:27 +03:00
ec1e8c8f3a Fix risc layer from cc changes 2019-08-23 10:20:39 +03:00
7ca3599c5a Fixed all after changing argument handling
arguments are now fully inlined into the message
locals next
2019-08-22 23:10:29 +03:00
0a1b05b2ee Inline arguments into message
thus removing indirection for access
does affect rather much, several commits
2019-08-22 17:55:19 +03:00
5dc8c046e7 increase binary_code size to 32
save a few jump, adds some size to binary
16 just seemed kind of small
2019-08-22 12:26:40 +03:00
15d1c07a1c Remove dead code
some copy paste in creating full mom layer left dome unused (untested) 
code
thanks to code climate stats
2019-08-18 10:19:52 +03:00
4c76ff3388 Remove Padded module
Parfait was depending on it, ie it created a dependency out of Parfait. But Parfiat needs to be self contained.
Moved 2 methods into parfait object, and resolved single call from text_writer to third.
2019-08-17 21:07:07 +03:00
155c042009 Fix forgotten block compiler
Especially on the way down to risc
2019-08-13 19:32:17 +03:00
aaf169ad8d Fixed almost all but Interpreter
150 only failing, seems only 1 bug though
and one in linker
2019-08-13 00:13:29 +03:00
9474932320 fixing risc collection tests 2019-08-12 16:12:17 +03:00
fa0aa30386 Move builtin wholesale to Mom
Since Builtin generates risc, just like mom instructions, it was a design mistake to put builtin into risc in the first place. Now that borders are coming more into focus, it make much more sense to have the builtin in mom.
In fact the instructions should be moved out and a seperate invocation mechanism used , so functions can be parsed, not generated (wip)
2019-08-12 12:38:29 +03:00
2326081161 Now booting all original methods
Just need to refactor now
2019-08-12 11:31:47 +03:00
91ddae2251 Start to convert integer operations
All apart from operators, which are its own thing
2019-08-12 11:08:09 +03:00
97488c4e5b Transform builtin word functions
according to same schema
2019-08-12 10:45:07 +03:00
3282e0ae06 transformed object builtins 2019-08-11 20:36:10 +03:00
0725f02e9a starting to fix builtin
start at Object get_interna_word
using the pattern to replace the whole risc method with a single mom instruction. Copying the original risc code into the instrucitons to_risc
also adding some very basic tests
2019-08-11 14:31:00 +03:00
213938075f fix most mom and risc apart
apart from things involving builtn, which is not yet conceptually solved (as it codes risc, not mom)
2019-08-10 21:30:00 +03:00
d5f89a4979 compile from mom compiler to risc 2019-08-10 12:42:47 +03:00
5994cd3276 starting on mom to risc
some docs too
2019-08-08 12:19:27 +03:00
Torsten Rüger
66c2adda20 Copy risc compiler stuff to mom
Start to separate the layers. 
wip, just checkin in to see the following changes better
2019-08-06 18:33:27 +03:00
ab87806d08 fixes #26
only recurse to 1k
then come up for air an go again.
Should allow for 1M objects on a 2k stack (previously exceptions at 3.6k)
2019-07-28 16:42:40 +03:00
14c965360d a basic interpret command for cli
part of the benchmark effort
determine amount of objects
2019-07-25 21:23:55 +03:00
Torsten Ruger
9bd2195a15 use cattr instead of metaclass
so we can catch it in the compiler, like also attr
define cattr in adapter, no change in tests !
2019-03-04 19:56:53 +02:00
Torsten Ruger
3db7707614 Get a basic MetaClass going
Does get created in new, but not in boot.
Also not yet used in compiling
2019-02-16 23:24:16 +02:00
Torsten Ruger
a89301d623 finally creating less objects in the test
basic still #23 , now applied
Basic size of 20, interpreter gets 50 and the full set is 1024

Tests run more than twice as fast!!
2019-02-09 12:44:35 +02:00
Torsten Ruger
74b790250a use class variables instead of globals
and set page size from new argument
This closes #23 , though ripples will follow
2019-02-08 23:03:08 +02:00
Torsten Ruger
8d3a1954fa close #21
Mostly replaced Fixnum with integer
also in the rx-file dependency
all travis and testing with 2.4+
2019-02-07 18:24:35 +02:00
Torsten Ruger
4d30727811 test and last fixes for previous commits (allocalte ints)
mostly just fixing the additional instructions
close #20
2018-11-24 22:40:22 +02:00
Torsten Ruger
e6615d0a6a remove the old add_new_int and calls
last orrurences in syscalls replaced, 2 variants
- exit does not actually need the int
- all else, preallocate the int beforehand and in syscall assume the reg name (integer_tmp)

test missing
2018-11-22 09:16:56 +02:00
Torsten Ruger
a8d1f070f1 replace add_new_int in most builtins
in integer and word mainly, replaced with allocate_int
(tests pending as syscalls are undone)
2018-11-21 20:29:22 +02:00