spl/stage2/utils/source.spl

// Expect retcode = 0
// 源文件与位置映射：fname + text 借用调用方存储；行起始表由 Source 拥有。
// 提供字节偏移 <-> (行,列)（行列均为 1-based、按字节计），以及取整行文本。
// Sources 是按 fname 索引的多文件集合。

type .{ .Error, .Status, .make } = @import("error.spl");
type .{ .ByteView, .FindResult } = @import("byte_view.spl");
type .{ .Vec } = @import("vec.spl");
type .{ .Map, .hash_kind_view } = @import("map.spl");
type .{ .hash_view, .eq_view } = @import("hash.spl");

type Position = struct {
    line: usize, // 1-based
    col: usize,  // 1-based，按字节
}

type Source = @self;

fname: ByteView,    // 借用
text: ByteView,     // 借用
line_starts: Vec,   // usize：每行起始字节偏移（含首行 0）

type SourceResult = enum {
    Ok: Source,
    Err: Error,
}

type AddResult = enum {
    Ok: usize,
    Err: Error,
}

fn blank(): Source {
    ret Source {
        .fname = ByteView.empty(),
        .text = ByteView.empty(),
        .line_starts = Vec.blank(@sizeof(usize), 8),
    };
}

fn new(fname: ByteView, text: ByteView): SourceResult {
    var starts = Vec.blank(@sizeof(usize), 8);
    var zero: usize = 0;
    match Vec.push(&starts, &zero as *u8) {
        .Ok => {},
        .Err(error) => { ret .Err(error); },
    }

    var i: usize = 0;
    while i < text.len {
        if text.ptr[i] == '\n' {
            var next = i + 1;
            match Vec.push(&starts, &next as *u8) {
                .Ok => {},
                .Err(error) => { Vec.drop(&starts); ret .Err(error); },
            }
        }
        i = i + 1;
    }

    ret .Ok(Source { .fname = fname, .text = text, .line_starts = starts });
}

fn line_count(self: Source): usize {
    ret self.line_starts.len;
}

fn line_start_offset(self: Source, line: usize): usize {
    if line == 0 || line > self.line_starts.len { ret ~(0 as usize); }
    match Vec.get(&self.line_starts, line - 1) {
        .Ok(ptr) => { ret (ptr as *usize).*; },
        .Err(error) => { ret ~(0 as usize); },
    }
    ret ~(0 as usize);
}

// offset 所在行（0-based）；超出范围夹到末行。
fn line_index_at(self: Source, offset: usize): usize {
    var off = offset;
    if off > self.text.len { off = self.text.len; }
    var lo: usize = 0;
    var hi: usize = self.line_starts.len;
    while lo + 1 < hi {
        var mid = (lo + hi) / 2;
        var start = line_start_offset(self, mid + 1); // 1-based -> 内部
        if start != ~(0 as usize) && start <= off { lo = mid; } else { hi = mid; }
    }
    ret lo;
}

fn line_col(self: Source, offset: usize): Position {
    var off = offset;
    if off > self.text.len { off = self.text.len; }
    var line0 = line_index_at(self, off);
    var start = line_start_offset(self, line0 + 1); // 1-based
    if start == ~(0 as usize) { start = 0; }
    ret Position { .line = line0 + 1, .col = off - start + 1 };
}

// line 为 1-based；返回整行（去掉行尾 '\n' 与 '\r'）。
fn line_text(self: Source, line: usize): ByteView {
    var n = self.line_starts.len;
    if line == 0 || line > n { ret ByteView.empty(); }
    var start = line_start_offset(self, line);
    if start == ~(0 as usize) { ret ByteView.empty(); }
    var end = self.text.len;
    if line < n {
        var next = line_start_offset(self, line + 1);
        if next != ~(0 as usize) { end = next; }
    }
    while end > start && (self.text.ptr[end - 1] == '\n' || self.text.ptr[end - 1] == '\r') {
        end = end - 1;
    }
    var r = ByteView.slice(self.text, start, end);
    match r {
        .Ok(v) => { ret v; },
        .Err(error) => { ret ByteView.empty(); },
    }
    ret ByteView.empty();
}

fn drop(self: *Source) {
    Vec.drop(&self.line_starts);
}

// ---- 多文件集合 ----

type Sources = struct {
    items: Vec,   // Source
    by_name: Map, // ByteView(fname) -> usize
}

fn set_new(): Sources {
    ret Sources {
        .items = Vec.blank(@sizeof(Source), 8),
        .by_name = Map.blank(@sizeof(ByteView), @alignof(ByteView), @sizeof(usize),
                             @alignof(usize), hash_kind_view()),
    };
}

fn set_add(set: *Sources, fname: ByteView, text: ByteView): AddResult {
    var src = new(fname, text);
    var s = blank();
    match src {
        .Ok(v) => { s = v; },
        .Err(error) => { ret .Err(error); },
    }

    var index = set.items.len;
    match Vec.push(&set.items, &s as *u8) {
        .Ok => {},
        .Err(error) => { drop(&s); ret .Err(error); },
    }
    match Map.put(&set.by_name, &fname as *u8, &index as *u8) {
        .Ok => {},
        .Err(error) => {
            // 回滚最后一项
            set.items.len = index;
            drop(&s);
            ret .Err(error);
        },
    }
    ret .Ok(index);
}

fn set_at(set: *Sources, index: usize): *Source {
    match Vec.get(&set.items, index) {
        .Ok(ptr) => { ret ptr as *Source; },
        .Err(error) => { ret null; },
    }
    ret null;
}

fn set_find(set: *Sources, fname: ByteView): FindResult {
    var out: usize = 0;
    match Map.get(&set.by_name, &fname as *u8, &out as *u8) {
        .Found => { ret .Found(out); },
        .NotFound => {},
    }
    ret .NotFound;
}

fn set_drop(set: *Sources) {
    var i: usize = 0;
    while i < set.items.len {
        var s = set_at(set, i);
        if s != null { drop(s); }
        i = i + 1;
    }
    Vec.drop(&set.items);
    Map.drop(&set.by_name);
}

// ---- 测试辅助 ----

fn view(text: []u8): ByteView {
    ret ByteView.from_slice(text);
}

fn test_single() {
    var text = "line1\nline2\n\nline4";
    var r = new(view("a.spl"), view(text));
    var src = blank();
    match r {
        .Ok(v) => { src = v; },
        .Err(error) => { @assert(false, "source new"); },
    }

    @assert(line_count(src) == 4, "line count");

    var p0 = line_col(src, 0);
    @assert(p0.line == 1 && p0.col == 1, "start");
    var p6 = line_col(src, 6);
    @assert(p6.line == 2 && p6.col == 1, "line2 start");
    var p8 = line_col(src, 8);
    @assert(p8.line == 2 && p8.col == 3, "line2 col3");
    var p12 = line_col(src, 12);
    @assert(p12.line == 3 && p12.col == 1, "empty line3");
    var p13 = line_col(src, 13);
    @assert(p13.line == 4 && p13.col == 1, "line4 start");
    var p18 = line_col(src, 18);
    @assert(p18.line == 4 && p18.col == 6, "past end clamps");

    var l1 = line_text(src, 1);
    var l3 = line_text(src, 3);
    var l4 = line_text(src, 4);
    @assert(ByteView.eq(l1, view("line1")), "line1");
    @assert(ByteView.eq(l3, view("")), "line3 empty");
    @assert(ByteView.eq(l4, view("line4")), "line4");

    drop(&src);
}

fn test_sources() {
    var set = set_new();
    var r1 = set_add(&set, view("a.spl"), view("x\ny"));
    var r2 = set_add(&set, view("b.spl"), view("z"));
    var a: usize = 0;
    var b: usize = 0;
    match r1 { .Ok(v) => { a = v; }, .Err(e) => { @assert(false, "add a"); } }
    match r2 { .Ok(v) => { b = v; }, .Err(e) => { @assert(false, "add b"); } }
    @assert(a == 0 && b == 1, "indices");

    match set_find(&set, view("b.spl")) {
        .Found(i) => { @assert(i == 1, "find b"); },
        .NotFound => { @assert(false, "b missing"); },
    }
    match set_find(&set, view("c.spl")) {
        .Found(i) => { @assert(false, "c unexpected"); },
        .NotFound => {},
    }

    var sb = set_at(&set, b);
    @assert(sb != null && line_count(sb.*) == 1, "at b");
    set_drop(&set);
}

fn main() {
    test_single();
    test_sources();
}