Commit Graph

52 Commits

Author SHA1 Message Date
Nathan Braswell
08125551af Now computes closed_variables for each lambda, placeholder closure_struct type stuff. 2016-02-22 16:18:55 -05:00
Nathan Braswell
34f11b0874 Plain lambdas implemented, but not fully tested. No closures yet 2016-02-21 17:22:16 -05:00
Nathan Braswell
1795f1b4f1 Implemented function values when type inferenced (not explicit types or lambdas) 2016-02-20 21:02:41 -05:00
Nathan Braswell
b073d5806b Templated methods work now, even explicit instantiation (along with implicit) 2016-02-20 02:36:35 -05:00
Nathan Braswell
c5dda4b7ec Was decorating this by accident, fixed. 32 Kalypso tests passing 2016-02-17 13:59:10 -05:00
Nathan Braswell
21034a820f Finished implementing simple_passthrough 2016-02-17 13:37:48 -05:00
Nathan Braswell
5a4d139d18 Just decided not to support defs in Kalypso (or Cephelapod, but it's already kinda implemented there) and ported some more tests over 2016-02-16 13:43:47 -05:00
Nathan Braswell
a898104f8a Fix multiline strings with single quotes in Cephelapod and implement both types in Kalypso, one new test should pass for both 2016-02-15 23:12:56 -05:00
Nathan Braswell
815c213270 Make --parse-only not highlight and redo format of parse errors so that kraken.vim syntax checking works, fix naming to allow multiple instantiations of object templates, fix so that template type replacements go through to bodies of methods of object templates in the fourth_pass 2016-02-15 16:31:01 -05:00
Nathan Braswell
3da140cc5e Object templates starting to work, don't do methods properly yet, likely all have the same name 2016-02-13 16:56:37 -05:00
Nathan Braswell
2ac1639b5e A little more work on function inference, fixed both Cephalapod and Kalypso's c generators to not accidentally emit this-> for a method param if the object has a var of the same name, which allows us to disambiguate with object_name::member = param_with_same_name, behavior I had not thought of but would have worked out of the box if not for this accidental bug 2016-02-09 02:59:38 -05:00
Nathan Braswell
11eba1ba07 Port many tests and fix small bugs revealed in Kalypso (passes 24/72) - tests have also revealed more extensive features not yet implemented, and I seem to have messed up a test or two so that the C++ version also fails a couple more (it's at 59/71 now). Will investigate 2016-02-07 16:22:55 -05:00
Nathan Braswell
70ebefcc25 Initial explicit function template working (chained may also not work, need to check) 2016-02-01 05:35:08 -05:00
Nathan Braswell
447f0c83b1 Hopefully got self-hosted objects fully working 2016-01-31 19:29:08 -05:00
Nathan Braswell
f72930beef Also destruct parameters after copy_constructing them 2016-01-31 00:10:54 -05:00
Nathan Braswell
ab5250bd0b Copy construct into function calls 2016-01-30 23:59:21 -05:00
Nathan Braswell
83450aa833 Added copy_constructing when returning 2016-01-30 22:04:37 -05:00
Nathan Braswell
6f1f31a212 Added automatic distructor calling for going out of scope, found out that += is broken (and just comes through as =) 2016-01-29 22:46:09 -05:00
Nathan Braswell
17d4371d5c Implemented init position calls 2016-01-28 20:51:40 -05:00
Nathan Braswell
42b942737b Prepping for introducing automatic copy_construct and destruct and tests for them, fixed using - to negate numbers 2016-01-28 12:55:51 -05:00
Nathan Braswell
b06c329236 Changed over to code_triple to set up for automatic object methods 2016-01-27 18:56:44 -05:00
Nathan Braswell
693134b499 defer should now would properly with break and continue 2016-01-25 22:48:41 -05:00
Nathan Braswell
9973bc3bd3 First run at defer working - won't properly do every scope out of a loop yet, likely other problems 2016-01-25 13:48:27 -05:00
Nathan Braswell
135305fb76 The piping for defer, though not the correct stacks in c_generator. Bedtime though 2016-01-25 02:53:00 -05:00
Nathan Braswell
83a76c36de continue and break statements 2016-01-24 17:31:41 -05:00
Nathan Braswell
105a969a00 Method calls from within method from same object 2016-01-24 01:49:14 -05:00
Nathan Braswell
c022758b9c Make this varaiable/keyword work 2016-01-24 01:02:56 -05:00
Nathan Braswell
23cad3ad1b Access to member variables from inside methods now work 2016-01-23 20:39:06 -05:00
Nathan Braswell
961feb6fa5 Method calls will actually work now, but only as regular functions would, though there is a little in place to work towards true methods in the future 2016-01-23 05:33:56 -05:00
Nathan Braswell
c6554ce463 Whoops, some fixes from yesterday 2016-01-23 00:33:47 -05:00
Nathan Braswell
fc343ceea8 This commit has its date moved to earlier to reflect when the work was completed, as I forgot to actually make the commit until 2 minutes after midnight :/. It partially fixed a bug where ADT requires equality for its member types but won't use templated equality to be if there isn't a regular equality operator (not templated) it always returns false 2016-01-22 23:59:59 -05:00
Nathan Braswell
8c5bf36433 Object member access works now 2016-01-21 19:16:29 -05:00
Nathan Braswell
d43f36f1d9 type_def variables and methods are not parsed in ast_transformation, and kind-of generated in c_generator, but access and whatnot not supported yet 2016-01-21 12:54:21 -05:00
Nathan Braswell
c943d591e0 Trivial objects working, fixed adt prefixing bug 2016-01-21 03:18:02 -05:00
Nathan Braswell
4ebb8bf107 Finished poset implementation added to the standard library, starting on getting types added and sorting, etc 2016-01-20 13:50:40 -05:00
Nathan Braswell
162cc98f30 Added for loops 2016-01-19 11:47:09 -05:00
Nathan Braswell
ca85edaeee added while loops and fixed unary operators (including correct precrement and decrement) 2016-01-19 03:16:16 -05:00
Nathan Braswell
4493dfd861 added if_statements 2016-01-19 02:06:30 -05:00
Nathan Braswell
458299abe1 Add perentheses to all operators so that acutal perentheses are preserved and the kraken order of operations is maintained 2016-01-18 18:10:07 -05:00
Nathan Braswell
ac34a550d5 Basic math expressions will passthrough now 2016-01-18 18:04:34 -05:00
Nathan Braswell
bffedcf2fd Add the abitlity to assign during a declaration and type inference if no type given but there is an expression 2016-01-17 01:10:09 -05:00
Nathan Braswell
9c41c2fd12 Added in assignment and variable usage 2016-01-16 22:14:59 -05:00
Nathan Braswell
bf570f027c Variable declaration added now 2016-01-15 19:10:52 -05:00
Nathan Braswell
fe6818edfc working towards function overloading, the emitted functions and function calls are now decorated and we're starting to get the prereqs for function lookup 2016-01-13 21:09:28 -05:00
Nathan Braswell
cb5b072b58 First Hello World! Now have to clean it up a little bit (correct assignment passthrough, add in the newline) 2016-01-12 00:21:01 -05:00
Nathan Braswell
4c569f4f8c Added function calls, printing out of pointers 2016-01-11 23:41:09 -05:00
Nathan Braswell
5db0365a63 Added indirection to types in prep for function calls, full passthrough, and the first real hello world 2016-01-10 18:26:31 -05:00
Nathan Braswell
16aa01a76e Placeholder/passthrough for all the expressions, simple value node passthrough (though if Kraken has the same literal syntax as C it'll work for a while, though it really needs to encode the type... 2016-01-08 00:33:05 -05:00
Nathan Braswell
daae39fe19 Added in the beginnings of pass three which can parse and emit statements and code blocks 2016-01-07 02:52:22 -05:00
Nathan Braswell
337bc424ee Added in basic transformation and generation for functions 2016-01-06 02:46:42 -05:00