summaryrefslogtreecommitdiff
path: root/main.c
diff options
context:
space:
mode:
Diffstat (limited to 'main.c')
-rw-r--r--main.c10
1 files changed, 5 insertions, 5 deletions
diff --git a/main.c b/main.c
index 930d947..f7d5b02 100644
--- a/main.c
+++ b/main.c
@@ -8,14 +8,14 @@
/*
@TODO:
-do we need the possibility that IdentSlot.decl is NULL?
+test:
+ _ := 5;
+ _ := 6;
+don't allow use of the identifier _, e.g. puti(_+5) ok for fields i think
use
- use with struct members (e.g. SuperPoint ::= struct { use p: Point; })
maybe change to #define check(x) do { if_unlikely(x) return 0; } while (0);
always use pointers in cgen'd non-range for loops (sometimes also indices)
-test:
- _ := 5;
- _ := 6;
is there a problem where we can get TYPE_UNKNOWN in cgen, triggering an assert(0)?
-simple example, but maybe try other stuff: x := #C("5");
-also make sure you can't do x:#C("5");
@@ -25,9 +25,9 @@ allow `use ???;` if an error has already occurred
if something gets included into a namespace, and its typing fails, the namespace should still be of type namespace, not ???
make sure you can do a[i] where a is &[5]int or &[]char or something
do we consistently handle x := &some_array_or_slice; x.len
+&void
simplify eval macros with val_to_u/i64
#if should not create a block
-&void
&&, ||
start making a standard library... (printf; stringbuilder would be nice to have)
switch