aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorAndrew Chambers <[email protected]>2021-10-16 13:57:23 +1300
committerAndrew Chambers <[email protected]>2021-10-16 13:57:23 +1300
commit1a70575220391014cda8f7459979a6ca0bcd3787 (patch)
tree33f23403fd0399708a25d8ebfdc5f71686b4f0b4
parente426b0ba589b6780546e712354dfd6e9880a5ee5 (diff)
Improve comments and readme.
-rw-r--r--README.md9
-rw-r--r--parse.c8
2 files changed, 11 insertions, 6 deletions
diff --git a/README.md b/README.md
index 4c2ec6e..b3c72f1 100644
--- a/README.md
+++ b/README.md
@@ -33,10 +33,17 @@ cc -O2 *.c -o minias
# Roadmap
+Essential features:
+
- [x] Self host, feature complete for goals.
- [ ] Two pass jump relaxing.
- [ ] Immediate relaxing.
-- [ ] Bonus features.
+
+Bonus features:
+
+- Test every opcode with all variants in our test suite.
+- Keep simplifying while keeping the code tidy.
+- Parser that doesn't depend on peg/leg.
# Notes
diff --git a/parse.c b/parse.c
index 8704298..a9d7b70 100644
--- a/parse.c
+++ b/parse.c
@@ -8,11 +8,9 @@ static const Parsev *internparsev(Parsev *p) {
for equality, even on pointer values, this works because the
pointers themselves are also interned.
- This simplicity somes with one big cost - Parsev variants with padding
- can trigger a false positive on valgrind. It should still safe
- because reading these undefined bytes do not change the behavior of the
- program. The best fix is still to avoid the padding bytes in the Parsev
- variant layout using a tool such as 'pahole'.
+ This simplicity comes with one big cost - Parsev variants with padding
+ can trigger a false positive on valgrind. It should still be safe,
+ but the best fix is still to avoid the padding bytes in the Parsev variants.
*/
size_t idx;
const Parsev *interned;