nexoniarz/Finch
Watch
1
0
Fork
You've already forked Finch
 0
mirror of https://github.com/Nexoniarz/Finch.git synced 2026-11-03 08:54:25 +00:00
Code Issues Projects Releases Packages Wiki Activity

enums and match (in the compiler written in Finch)

Browse source 
enum Color { Red, Green, Blue }   enum Status { Ok = 200, NotFound = 404 }

Color.Red, Color(n) (checked), int(c), == and != only, printed by name, usable as
map keys and struct fields, with methods (fn Color.next()).

match value { Red, Green -> ... Blue -> ... else -> ... } as a statement, or as an
expression that gives a value. Works on enums, whole numbers, chars, text and bools;
a match on an enum has to handle every variant or end with else, and says which are
missing. A bare variant name (Red) is enough inside a match on an enum.

Only the self-hosted compiler has them; tests/new holds their tests, tests/run stays what
the C++ compiler also runs. boot.sh: 31 passed.
This commit is contained in:
Nexoniarz 2026-10-25 20:49:04 +01:00
commit 609ceffe58
15 changed files with 576 additions and 18 deletions
Download patch file Download diff file Expand all files Collapse all files

322
boot/gen.fch
View file

@ -16,6 +16,13 @@ struct Var {
// A scope's variables, and the defers that run when it ends, newest first.
// A defer sees only the variables declared before it: `orders` says how many.
struct EnumDef {
str module
str name
[]str variants
[]int values
}
struct Scope {
[]Var vars
[]ast.Node defers
@ -33,6 +40,7 @@ struct FnSig {
str name
str recv // a method of this struct ("" for a function)
str recvmod // and the module of that struct: "C" for a struct from a header
bool recvEnum // the receiver is an enum
bool fallible // -> T!: returns { i1 ok, T value, str error }
bool mutself // the method changes self
str ll
@ -96,6 +104,7 @@ struct Gen {
str module
[]FnSig sigs
[]StructDef structs
[]EnumDef enums
[]str done
[]str mm // names of methods that change self, for the mutation analysis
int nvars // variables declared so far in this function
@ -273,6 +282,15 @@ fn findStruct(ptr[Gen] g, str module, str name) -> int {
return -1
}
fn findEnum(ptr[Gen] g, str module, str name) -> int {
for i in 0..g.enums.len {
if g.enums[i].module == module && g.enums[i].name == name {
return i
}
}
return -1
}
fn progOf(ptr[Gen] g, str module) -> int {
for i in 0..g.progs.len {
if g.progs[i].module == module {
@ -320,6 +338,9 @@ fn resolve(ptr[Gen] g, ast.Type t, ast.Node at) -> ast.Type {
}
module = t.module
}
if findEnum(g, module, t.name) >= 0 {
return ast.Type(kind: "enum", name: t.name, module: module)
}
if findStruct(g, module, t.name) < 0 {
// a bare name may be a struct from a C header
if t.module.len == 0 && findStruct(g, "C", t.name) >= 0 {
@ -360,6 +381,9 @@ fn llty(ptr[Gen] g, ast.Type t) -> str {
if k == "fixed" {
return "[" + str(t.count) + " x " + llty(g, t.elem[0]) + "]"
}
if k == "enum" {
return "i32"
}
if k == "ptr" || k == "null" {
return "ptr"
}
@ -389,7 +413,7 @@ fn nodeKey(ast.Node n) -> str {
fn zeroOf(ast.Type t) -> str {
k := t.kind
if isInt(t) || k == "char" {
if isInt(t) || k == "char" || k == "enum" {
return "0"
}
if k == "bool" {
@ -425,6 +449,9 @@ fn typeKey(ast.Type t) -> str {
if t.kind == "map" {
return "map_" + typeKey(t.elem[0]) + "_" + typeKey(t.elem[1])
}
if t.kind == "enum" {
return "E_" + modName(t.module) + "_" + t.name
}
if t.kind == "struct" {
return "S_" + modName(t.module) + "_" + t.name
}
@ -751,6 +778,31 @@ fn formatFn(ptr[Gen] g, ast.Type t) -> str {
emitFormat(g, "%buf", load(g, llty(g, et), ep), et, true)
loopEnd(g, l)
bufText(g, "%buf", "]")
} else if t.kind == "enum" {
// the variant's name; a number that is no variant shows as Color(7)
ed := g.enums[findEnum(g, t.module, t.name)]
v := load(g, "i32", "%a")
def := newLabel(g, "enum.other")
done := newLabel(g, "enum.done")
[]str cases
[]str labels
for k in 0..ed.variants.len {
labels.push(newLabel(g, "enum.v"))
cases.push("i32 " + str(ed.values[k]) + ", label %" + labels[k])
}
emit(g, "switch i32 " + v + ", label %" + def + " [ " + cases.join(" ") + " ]")
g.term = true
for k in 0..ed.variants.len {
label(g, labels[k])
bufText(g, "%buf", ed.variants[k])
br(g, done)
}
label(g, def)
bufText(g, "%buf", ed.name + "(")
emit(g, "call void @finch_buf_int(ptr %buf, i64 " + op(g, "sext i32 " + v + " to i64") + ")")
bufText(g, "%buf", ")")
br(g, done)
label(g, done)
} else if t.kind == "map" {
// {"a": 1, "b": 2}
first := alloca(g, "i1")
@ -1038,6 +1090,32 @@ fn declare(ptr[Gen] g) {
g.structs.push(StructDef(module: g.module, name: s.name, prog: pi, index: si))
}
}
for pi in 0..g.progs.len {
g.module = g.progs[pi].module
for ei in 0..g.progs[pi].enums.len {
e := g.progs[pi].enums[ei]
if findEnum(g, g.module, e.name) >= 0 || findStruct(g, g.module, e.name) >= 0 {
fail(g, e, "'" + e.name + "' is defined twice")
}
ed := EnumDef(module: g.module, name: e.name)
next := 0
for v in e.kids {
if v.text.len > 0 {
next = int(v.text)
}
if ed.variants.contains(v.name) {
fail(g, v, "the enum " + e.name + " has the variant " + v.name + " twice")
}
if ed.values.contains(next) {
fail(g, v, "two variants of " + e.name + " have the number " + str(next) + " (give each its own)")
}
ed.variants.push(v.name)
ed.values.push(next)
next += 1
}
g.enums.push(ed)
}
}
for ni in 0..g.ci.structOrder.len {
g.structs.push(StructDef(module: "C", name: g.ci.structOrder[ni], isC: true, prog: -1, index: ni))
}
@ -1049,7 +1127,13 @@ fn declare(ptr[Gen] g) {
for fi in 0..g.progs[pi].fns.len {
f := g.progs[pi].fns[fi]
recvmod := g.module
if f.recv.len > 0 {
recvEnum := false
if f.recv.len > 0 && findEnum(g, g.module, f.recv) >= 0 {
recvEnum = true
if findMethod(g, recvmod, f.recv, f.name) >= 0 {
fail(g, f, "the method " + f.recv + "." + f.name + " is defined twice")
}
} else if f.recv.len > 0 {
si := findStruct(g, g.module, f.recv)
if si < 0 && findStruct(g, "C", f.recv) >= 0 {
si = findStruct(g, "C", f.recv) // a method on a struct from C
@ -1072,7 +1156,7 @@ fn declare(ptr[Gen] g) {
fail(g, f, "'" + f.name + "' is a built-in function, pick another name")
}
}
sig := FnSig(module: g.module, name: f.name, recv: f.recv, recvmod: recvmod, fallible: f.fallible, prog: pi, index: fi)
sig := FnSig(module: g.module, name: f.name, recv: f.recv, recvmod: recvmod, recvEnum: recvEnum, fallible: f.fallible, prog: pi, index: fi)
sig.ll = "@finch." + modName(g.module) + "." + f.name
if f.recv.len > 0 {
sig.ll = "@finch." + modName(g.module) + "." + f.recv + "." + f.name
@ -1208,7 +1292,11 @@ fn define(ptr[Gen] g, int si) {
if g.cur.recv.len > 0 {
// self is the caller's struct itself, reached through its address
params.push("ptr %self")
g.scopes[0].vars.push(Var(name: "self", ty: ast.Type(kind: "struct", name: g.cur.recv, module: g.cur.recvmod), slot: "%self", readonly: false))
selfKind := "struct"
if g.cur.recvEnum {
selfKind = "enum"
}
g.scopes[0].vars.push(Var(name: "self", ty: ast.Type(kind: selfKind, name: g.cur.recv, module: g.cur.recvmod), slot: "%self", readonly: false))
}
for k in 0..g.cur.params.len {
pt := g.cur.params[k]
@ -1266,6 +1354,8 @@ fn stmt(ptr[Gen] g, ast.Node s) {
} else {
expr(g, s.kids[0])
}
} else if k == "match" {
matchGen(g, s, false)
} else if k == "if" {
ifStmt(g, s)
} else if k == "while" {
@ -1914,6 +2004,139 @@ fn mapMethod(ptr[Gen] g, ast.Node m, str mp, ast.Type t) -> Val {
return Val()
}
// ---------- match ----------
// A pattern as a value of the subject's type. For an enum, a bare variant name (Red) works too.
fn patternValue(ptr[Gen] g, ast.Node pat, ast.Type st) -> Val {
if st.kind == "enum" && pat.kind == "var" && lookup(g, pat.name) < 0 {
ed := g.enums[findEnum(g, st.module, st.name)]
k := ed.variants.find(pat.name)
if k < 0 {
fail(g, pat, "'" + pat.name + "' is not a variant of " + ed.name + " (it has: " + ed.variants.join(", ") + ")")
}
return Val(v: str(ed.values[k]), ty: st)
}
return coerce(g, exprWant(g, pat, st), st, pat, "this pattern")
}
// match value { A, B -> ... C -> ... else -> ... }, as a statement or (asValue) as an expression.
// The patterns are tried in order; for an enum the arms have to cover every variant, or end with else.
fn matchGen(ptr[Gen] g, ast.Node e, bool asValue) -> Val {
subj := expr(g, e.kids[0])
st := subj.ty
if st.kind == "void" || st.kind == "null" {
fail(g, e.kids[0], "there is nothing to match here")
}
end := newLabel(g, "match.end")
hasElse := false
reached := false // does any arm go on to the end?
[]str covered
[]str inVals
[]str inLabels
resultTy := T("void")
arms := e.kids.len - 1
for ai in 0..arms {
arm := e.kids[1 + ai]
if hasElse {
fail(g, arm, "this arm can never be used: 'else' already takes everything that is left")
}
body := newLabel(g, "match.arm")
next := newLabel(g, "match.next")
if arm.flag {
hasElse = true
br(g, body)
} else {
pats := arm.kids[0].kids
for pi in 0..pats.len {
pv := patternValue(g, pats[pi], st)
if covered.contains(pv.v) && st.kind == "enum" {
fail(g, pats[pi], "this variant is already handled by an earlier arm")
}
covered.push(pv.v)
hit := compare(g, "==", subj, pv, pats[pi])
if pi == pats.len - 1 {
condBr(g, hit.v, body, next)
} else {
more := newLabel(g, "match.or")
condBr(g, hit.v, body, more)
label(g, more)
}
}
}
label(g, body)
node := arm.kids[1]
if asValue {
v := expr(g, node)
if ai == 0 || resultTy.kind == "void" {
if v.ty.kind == "void" || v.ty.kind == "null" {
fail(g, node, "an arm of a match that gives a value has to give one")
}
resultTy = v.ty
}
v = coerce(g, v, resultTy, node, "this arm (the arms of a match have to give the same type)")
out := own(g, v)
from := newLabel(g, "match.from")
br(g, from)
label(g, from)
inVals.push(out)
inLabels.push(from)
br(g, end)
reached = true
} else {
if node.kind == "block" {
block(g, node, true)
} else {
stmt(g, node)
}
if !g.term {
reached = true
}
br(g, end)
}
if !arm.flag {
label(g, next)
}
}
if !hasElse {
if st.kind == "enum" {
ed := g.enums[findEnum(g, st.module, st.name)]
[]str missing
for k in 0..ed.variants.len {
if !covered.contains(str(ed.values[k])) {
missing.push(ed.variants[k])
}
}
if missing.len > 0 {
fail(g, e, "this match doesn't handle " + missing.join(", ") + " (add them, or an 'else' arm)")
}
} else {
fail(g, e, "a match on " + ast.show(st) + " needs an 'else' arm for everything that is not listed")
}
// every variant is handled; only a number that was never a variant gets here
panicIf(g, "true", "no arm of this match fits the value", e)
emit(g, "unreachable")
g.term = true
}
label(g, end)
if !reached {
emit(g, "unreachable")
g.term = true
return Val(ty: T("void"))
}
if asValue {
return Val(v: op(g, "phi " + llty(g, resultTy) + " " + phiList(inVals, inLabels)), ty: resultTy, fresh: owning(g, resultTy))
}
return Val(ty: T("void"))
}
fn phiList([]str vals, []str labels) -> str {
[]str parts
for i in 0..vals.len {
parts.push("[ " + vals[i] + ", %" + labels[i] + " ]")
}
return parts.join(", ")
}
// ---------- errors as values ----------
struct Fallible {
@ -2068,6 +2291,9 @@ fn expr(ptr[Gen] g, ast.Node e) -> Val {
}
return Val(v: load(g, llty(g, r.pl.ty), r.pl.addr), ty: r.pl.ty)
}
if k == "match" {
return matchGen(g, e, true)
}
if k == "try" {
return tryExpr(g, e)
}
@ -2263,6 +2489,16 @@ fn compare(ptr[Gen] g, str o, Val l, Val r, ast.Node at) -> Val {
preds := ["oeq", "une", "olt", "ole", "ogt", "oge"]
return Val(v: op(g, "fcmp " + preds[ops.find(o)] + " " + llty(g, l.ty) + " " + l.v + ", " + r.v), ty: T("bool"))
}
if k == "enum" {
if !eq {
fail(g, at, "enums can only be compared with == and != (to go through them in order, give the variants numbers and use int(...))")
}
pred := "eq"
if o == "!=" {
pred = "ne"
}
return Val(v: op(g, "icmp " + pred + " i32 " + l.v + ", " + r.v), ty: T("bool"))
}
if k == "bool" && !eq {
badOperands(g, o, l, r, at)
}
@ -2420,7 +2656,35 @@ fn llFloat(str text) -> str {
return s
}
// Which enum does `m.kids[0]` name, if it names one? Color, or shapes.Color. -1 if neither.
fn enumNamed(ptr[Gen] g, ast.Node n) -> int {
if n.kind == "var" && lookup(g, n.name) < 0 {
return findEnum(g, g.module, n.name)
}
if n.kind == "member" && n.kids[0].kind == "var" && lookup(g, n.kids[0].name) < 0 && imports(g, n.kids[0].name) {
return findEnum(g, n.kids[0].name, n.name)
}
return -1
}
// Color.Red
fn enumValue(ptr[Gen] g, ast.Node m, int ei) -> Val {
ed := g.enums[ei]
k := ed.variants.find(m.name)
if k < 0 {
fail(g, m, "the enum " + ed.name + " has no variant '" + m.name + "' (it has: " + ed.variants.join(", ") + ")")
}
return Val(v: str(ed.values[k]), ty: ast.Type(kind: "enum", name: ed.name, module: ed.module))
}
fn member(ptr[Gen] g, ast.Node m, bool write) -> Ref {
ei := enumNamed(g, m.kids[0])
if ei >= 0 {
if write {
fail(g, m, "a variant of an enum can't be changed")
}
return Ref(isPlace: false, val: enumValue(g, m, ei))
}
obj := ref(g, m.kids[0], write)
t := obj.val.ty
if obj.isPlace {
@ -2655,6 +2919,9 @@ fn convert(ptr[Gen] g, ast.Node c, ast.Type to) -> Val {
r := op(g, "call double @finch_str_to_float(ptr " + spill(g, v) + ", ptr " + file + ", i64 " + str(c.line) + ")")
return Val(v: floatCast(g, r, T("float"), to), ty: to)
}
if isInt(to) && from.kind == "enum" {
return Val(v: intCast(g, v.v, T("i32"), to), ty: to)
}
if isInt(to) || to.kind == "char" {
if isInt(from) || from.kind == "char" {
return Val(v: intCast(g, v.v, from, to), ty: to)
@ -2964,6 +3231,35 @@ fn cThunk(ptr[Gen] g, int fi, ast.Node at) -> str {
return name
}
// Color(n): a number as an enum. A literal is checked while compiling, anything else when it runs.
fn enumFrom(ptr[Gen] g, ast.Node c, int ei) -> Val {
ed := g.enums[ei]
checkArgs(g, c, 0, 1, ed.name)
v := expr(g, c.kids[0])
if !isInt(v.ty) {
fail(g, c.kids[0], ed.name + "(...) turns a whole number into the enum, but this is " + ast.show(v.ty))
}
t := ast.Type(kind: "enum", name: ed.name, module: ed.module)
if v.lit {
if !ed.values.contains(int(v.v)) {
fail(g, c.kids[0], v.v + " is not the number of any variant of " + ed.name)
}
return Val(v: v.v, ty: t)
}
wide := intCast(g, v.v, v.ty, T("int"))
known := ""
for k in 0..ed.values.len {
hit := op(g, "icmp eq i64 " + wide + ", " + str(ed.values[k]))
if k == 0 {
known = hit
} else {
known = op(g, "or i1 " + known + ", " + hit)
}
}
panicIf(g, op(g, "xor i1 " + known + ", true"), "this number is not a variant of " + ed.name, c)
return Val(v: op(g, "trunc i64 " + wide + " to i32"), ty: t)
}
// ---------- calling C ----------
// How C extends a number narrower than 32 bits: plain `char` is signed on x86.
@ -3173,7 +3469,7 @@ fn call(ptr[Gen] g, ast.Node c) -> Val {
} else if k == "bool" {
z := op(g, "zext i1 " + v.v + " to i32")
emit(g, "call void @finch_str_from_bool(ptr " + out + ", i32 " + z + ")")
} else if k == "bool" || k == "char" || k == "array" || k == "map" || k == "struct" {
} else if k == "bool" || k == "char" || k == "array" || k == "map" || k == "struct" || k == "enum" {
store(g, "{ ptr, i64, i64 }", "zeroinitializer", out)
emitFormat(g, out, v.v, v.ty, false)
} else {
@ -3286,6 +3582,10 @@ fn call(ptr[Gen] g, ast.Node c) -> Val {
if si >= 0 {
return construct(g, si, c.kids, c.names, c)
}
ei := findEnum(g, g.module, n)
if ei >= 0 {
return enumFrom(g, c, ei)
}
if g.ci.fns.has(n) {
return callC(g, c, g.ci.fns[n])
}
@ -3411,6 +3711,10 @@ fn method(ptr[Gen] g, ast.Node m) -> Val {
if fi >= 0 {
return callFn(g, fi, m, 1, mn + "." + m.name, "", "")
}
mei := findEnum(g, mn, m.name)
if mei >= 0 {
return enumFrom(g, m, mei)
}
si := findStruct(g, mn, m.name)
if si >= 0 {
[]ast.Node args
@ -3468,20 +3772,20 @@ fn method(ptr[Gen] g, ast.Node m) -> Val {
panicIf(g, op(g, "icmp eq ptr " + pv + ", null"), "called ." + m.name + "() on a null pointer", m)
st = t.elem[0]
selfAddr = pv
} else if t.kind == "struct" {
} else if t.kind == "struct" || t.kind == "enum" {
if r.isPlace {
selfAddr = r.pl.addr
} else {
selfAddr = spill(g, r.val)
}
}
if st.kind == "struct" {
if st.kind == "struct" || st.kind == "enum" {
fi := findMethod(g, st.module, st.name, m.name)
if fi < 0 {
fail(g, m, "the struct " + ast.show(st) + " has no method '" + m.name + "'")
fail(g, m, "the " + st.kind + " " + ast.show(st) + " has no method '" + m.name + "'")
}
// a method that changes self can't be called on a temporary or a read-only value
if g.sigs[fi].mutself && t.kind == "struct" {
if g.sigs[fi].mutself && (t.kind == "struct" || t.kind == "enum") {
if !r.isPlace {
fail(g, m, m.name + "() would change a temporary " + ast.show(st) + "; store it in a variable first")
}