Compare commits
3 commits
1a47ad0ad2
...
8e12b7396a
| Author | SHA1 | Date | |
|---|---|---|---|
| 8e12b7396a | |||
| bb4a5e0bf3 | |||
| 3f7953be74 |
1 changed files with 223 additions and 145 deletions
368
src/srf.zig
368
src/srf.zig
|
|
@ -87,11 +87,9 @@ pub const Value = union(enum) {
|
|||
// }
|
||||
// }
|
||||
pub fn parse(allocator: std.mem.Allocator, str: []const u8, state: *RecordIterator.State, delimiter: u8) ParseError!ValueWithMetaData {
|
||||
const debug = str.len > 2 and str[0] == '1' and str[1] == '1';
|
||||
if (debug) log.debug("parsing {s}", .{str});
|
||||
const type_val_sep_raw = std.mem.indexOfScalar(u8, str, ':');
|
||||
if (type_val_sep_raw == null) {
|
||||
try parseError(allocator, "no type data or value after key", state.*);
|
||||
try parseError(allocator, "no type data or value after key", state);
|
||||
return ParseError.ParseFailed;
|
||||
}
|
||||
|
||||
|
|
@ -121,7 +119,7 @@ pub const Value = union(enum) {
|
|||
state.partial_line_column += total_chars;
|
||||
const Decoder = std.base64.standard.Decoder;
|
||||
const size = Decoder.calcSizeForSlice(val) catch {
|
||||
try parseError(allocator, "error parsing base64 value", state.*);
|
||||
try parseError(allocator, "error parsing base64 value", state);
|
||||
return .{
|
||||
.item_value = null,
|
||||
.error_parsing = true,
|
||||
|
|
@ -130,7 +128,7 @@ pub const Value = union(enum) {
|
|||
const data = try allocator.alloc(u8, size);
|
||||
errdefer allocator.free(data);
|
||||
Decoder.decode(data, val) catch {
|
||||
try parseError(allocator, "error parsing base64 value", state.*);
|
||||
try parseError(allocator, "error parsing base64 value", state);
|
||||
allocator.free(data);
|
||||
return .{
|
||||
.item_value = null,
|
||||
|
|
@ -151,7 +149,7 @@ pub const Value = union(enum) {
|
|||
state.partial_line_column += total_chars;
|
||||
const val_trimmed = std.mem.trim(u8, val, &std.ascii.whitespace);
|
||||
const number = std.fmt.parseFloat(@FieldType(Value, "number"), val_trimmed) catch {
|
||||
try parseError(allocator, "error parsing numeric value", state.*);
|
||||
try parseError(allocator, "error parsing numeric value", state);
|
||||
return .{
|
||||
.item_value = null,
|
||||
.error_parsing = true,
|
||||
|
|
@ -173,7 +171,7 @@ pub const Value = union(enum) {
|
|||
if (std.mem.eql(u8, "false", val_trimmed)) break :blk false;
|
||||
if (std.mem.eql(u8, "true", val_trimmed)) break :blk true;
|
||||
|
||||
try parseError(allocator, "error parsing boolean value", state.*);
|
||||
try parseError(allocator, "error parsing boolean value", state);
|
||||
return .{
|
||||
.item_value = null,
|
||||
.error_parsing = true,
|
||||
|
|
@ -200,18 +198,16 @@ pub const Value = union(enum) {
|
|||
state.partial_line_column += total_metadata_chars;
|
||||
const size = std.fmt.parseInt(usize, trimmed_meta, 0) catch {
|
||||
log.debug("parseInt fail, trimmed_data: '{s}'", .{trimmed_meta});
|
||||
try parseError(allocator, "unrecognized metadata for key", state.*);
|
||||
try parseError(allocator, "unrecognized metadata for key", state);
|
||||
return .{
|
||||
.item_value = null,
|
||||
.error_parsing = true,
|
||||
};
|
||||
};
|
||||
if (debug) log.debug("found fixed string size {d}. State {f}", .{ size, state });
|
||||
// Update again for number of bytes. All failures beyond this point are
|
||||
// fatal, so this is safe.
|
||||
state.column += size;
|
||||
state.partial_line_column += size;
|
||||
if (debug) log.debug("New state {f}", .{state});
|
||||
|
||||
// If we are being asked specifically for bytes, we no longer care about
|
||||
// delimiters. We just want raw bytes. This might adjust our line/column
|
||||
|
|
@ -220,41 +216,29 @@ pub const Value = union(enum) {
|
|||
if (rest_of_data.len >= size) {
|
||||
// We fit on this line, everything is "normal"
|
||||
const val = rest_of_data[0..size];
|
||||
if (debug) log.debug("val {s}", .{val});
|
||||
return .{
|
||||
.item_value = .{ .string = val },
|
||||
};
|
||||
}
|
||||
// This is not enough, we need more data from the reader
|
||||
log.debug("item value includes newlines {f}", .{state});
|
||||
// We need to advance the reader, so we need a copy of what we have so fa
|
||||
const start = try dupe(allocator, state.options, rest_of_data);
|
||||
defer allocator.free(start);
|
||||
const buf = try allocator.alloc(u8, size);
|
||||
errdefer allocator.free(buf);
|
||||
@memcpy(buf[0..rest_of_data.len], rest_of_data);
|
||||
// add back the newline we are skipping
|
||||
buf[rest_of_data.len] = '\n';
|
||||
// We won't do a parseError here. If we have an allocation error, read
|
||||
// error, or end of stream, all of these are fatal. Our reader is currently
|
||||
// past the newline, so we have to remove a character from size to account.
|
||||
const end = try state.reader.readAlloc(allocator, size - rest_of_data.len - 1);
|
||||
try state.reader.readSliceAll(buf[rest_of_data.len + 1 ..]);
|
||||
// However, we want to be past the end of the *next* newline too (in long
|
||||
// format mode)
|
||||
if (delimiter == '\n') state.reader.toss(1);
|
||||
defer allocator.free(end);
|
||||
// This \n is because the reader state will have advanced beyond the next newline, so end
|
||||
// really should start with the newline. This only applies to long mode, because otherwise the
|
||||
// entire record is a single line
|
||||
const final = try std.mem.concat(allocator, u8, &.{ start, "\n", end });
|
||||
// const final = if (delimiter == '\n')
|
||||
// try std.mem.concat(allocator, u8, &.{ start, "\n", end })
|
||||
// else
|
||||
// try std.mem.concat(allocator, u8, &.{ start, end });
|
||||
errdefer allocator.free(final);
|
||||
// log.debug("full val: {s}", .{final});
|
||||
std.debug.assert(final.len == size);
|
||||
// Because we've now advanced the line, we need to reset everything
|
||||
state.line += std.mem.count(u8, final, "\n");
|
||||
state.column = final.len - std.mem.lastIndexOf(u8, final, "\n").?;
|
||||
state.line += std.mem.count(u8, buf, "\n");
|
||||
state.column = buf.len - std.mem.lastIndexOf(u8, buf, "\n").?;
|
||||
state.partial_line_column = state.column;
|
||||
return .{
|
||||
.item_value = .{ .string = final },
|
||||
.item_value = .{ .string = buf },
|
||||
.reader_advanced = true,
|
||||
};
|
||||
}
|
||||
|
|
@ -266,6 +250,63 @@ pub const Field = struct {
|
|||
value: ?Value,
|
||||
};
|
||||
|
||||
fn coerce(name: []const u8, comptime T: type, val: ?Value) !T {
|
||||
// Here's the deduplicated set of field types that coerce needs to handle:
|
||||
// Direct from SRF values:
|
||||
// Need parsing from string:
|
||||
// - Date, ?Date -- Date.parse(string)
|
||||
//
|
||||
// Won't work with Record.to(T) generically:
|
||||
// - []const OptionContract -- nested sub-records (OptionsChain has calls/puts arrays)
|
||||
// - ?[]const Holding, ?[]const SectorWeight -- nested sub-records in EtfProfile
|
||||
//
|
||||
const ti = @typeInfo(T);
|
||||
if (val == null and ti != .optional)
|
||||
return error.NullValueCannotBeAssignedToNonNullField;
|
||||
|
||||
// []const u8 is classified as a pointer
|
||||
switch (ti) {
|
||||
.optional => |o| if (val) |_|
|
||||
return try coerce(name, o.child, val)
|
||||
else
|
||||
return null,
|
||||
.pointer => |p| {
|
||||
// We don't have an allocator, so the only thing we can do
|
||||
// here is manage []const u8 or []u8
|
||||
if (p.size != .slice or p.child != u8)
|
||||
return error.CoercionNotPossible;
|
||||
if (val.? != .string and val.? != .bytes)
|
||||
return error.CoercionNotPossible;
|
||||
if (val.? == .string)
|
||||
return val.?.string;
|
||||
return val.?.bytes;
|
||||
},
|
||||
.type, .void, .noreturn => return error.CoercionNotPossible,
|
||||
.comptime_float, .comptime_int, .undefined, .null, .error_union => return error.CoercionNotPossible,
|
||||
.error_set, .@"fn", .@"opaque", .frame => return error.CoercionNotPossible,
|
||||
.@"anyframe", .vector, .enum_literal => return error.CoercionNotPossible,
|
||||
.int => return @as(T, @intFromFloat(val.?.number)),
|
||||
.float => return @as(T, @floatCast(val.?.number)),
|
||||
.bool => return val.?.boolean,
|
||||
.@"enum" => return std.meta.stringToEnum(T, val.?.string).?,
|
||||
.array => return error.NotImplemented,
|
||||
.@"struct", .@"union" => {
|
||||
if (std.meta.hasMethod(T, "srfParse")) {
|
||||
if (val.? == .string)
|
||||
return T.srfParse(val.?.string) catch |e| {
|
||||
log.err(
|
||||
"custom parse of value {s} failed : {}",
|
||||
.{ val.?.string, e },
|
||||
);
|
||||
return error.CustomParseFailed;
|
||||
};
|
||||
}
|
||||
return error.CoercionNotPossible;
|
||||
},
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
// A record has a list of fields, with no assumptions regarding duplication,
|
||||
// etc. This is for parsing speed, but also for more flexibility in terms of
|
||||
// use cases. One can make a defacto array out of this structure by having
|
||||
|
|
@ -289,63 +330,6 @@ pub const Record = struct {
|
|||
return null;
|
||||
}
|
||||
|
||||
fn coerce(name: []const u8, comptime T: type, val: ?Value) !T {
|
||||
// Here's the deduplicated set of field types that coerce needs to handle:
|
||||
// Direct from SRF values:
|
||||
// Need parsing from string:
|
||||
// - Date, ?Date -- Date.parse(string)
|
||||
//
|
||||
// Won't work with Record.to(T) generically:
|
||||
// - []const OptionContract -- nested sub-records (OptionsChain has calls/puts arrays)
|
||||
// - ?[]const Holding, ?[]const SectorWeight -- nested sub-records in EtfProfile
|
||||
//
|
||||
const ti = @typeInfo(T);
|
||||
if (val == null and ti != .optional)
|
||||
return error.NullValueCannotBeAssignedToNonNullField;
|
||||
|
||||
// []const u8 is classified as a pointer
|
||||
switch (ti) {
|
||||
.optional => |o| if (val) |_|
|
||||
return try coerce(name, o.child, val)
|
||||
else
|
||||
return null,
|
||||
.pointer => |p| {
|
||||
// We don't have an allocator, so the only thing we can do
|
||||
// here is manage []const u8 or []u8
|
||||
if (p.size != .slice or p.child != u8)
|
||||
return error.CoercionNotPossible;
|
||||
if (val.? != .string and val.? != .bytes)
|
||||
return error.CoercionNotPossible;
|
||||
if (val.? == .string)
|
||||
return val.?.string;
|
||||
return val.?.bytes;
|
||||
},
|
||||
.type, .void, .noreturn => return error.CoercionNotPossible,
|
||||
.comptime_float, .comptime_int, .undefined, .null, .error_union => return error.CoercionNotPossible,
|
||||
.error_set, .@"fn", .@"opaque", .frame => return error.CoercionNotPossible,
|
||||
.@"anyframe", .vector, .enum_literal => return error.CoercionNotPossible,
|
||||
.int => return @as(T, @intFromFloat(val.?.number)),
|
||||
.float => return @as(T, @floatCast(val.?.number)),
|
||||
.bool => return val.?.boolean,
|
||||
.@"enum" => return std.meta.stringToEnum(T, val.?.string).?,
|
||||
.array => return error.NotImplemented,
|
||||
.@"struct", .@"union" => {
|
||||
if (std.meta.hasMethod(T, "srfParse")) {
|
||||
if (val.? == .string)
|
||||
return T.srfParse(val.?.string) catch |e| {
|
||||
log.err(
|
||||
"custom parse of value {s} failed : {}",
|
||||
.{ val.?.string, e },
|
||||
);
|
||||
return error.CustomParseFailed;
|
||||
};
|
||||
}
|
||||
return error.CoercionNotPossible;
|
||||
},
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
fn maxFields(comptime T: type) usize {
|
||||
const ti = @typeInfo(T);
|
||||
if (ti != .@"union") return std.meta.fields(T).len;
|
||||
|
|
@ -605,6 +589,7 @@ pub const RecordIterator = struct {
|
|||
|
||||
field_delimiter: u8 = ',',
|
||||
end_of_record_reached: bool = false,
|
||||
field_iterator: ?FieldIterator = null,
|
||||
|
||||
/// Takes the next line, trimming leading whitespace and ignoring comments
|
||||
/// Directives (comments starting with #!) are preserved
|
||||
|
|
@ -630,6 +615,11 @@ pub const RecordIterator = struct {
|
|||
// TODO: we need to capture the fieldIterator here and make sure it's run
|
||||
// to the ground to keep our state intact
|
||||
const state = self.state;
|
||||
if (state.field_iterator) |f| {
|
||||
// We need to finish the fields on the previous record
|
||||
while (try f.next()) |_| {}
|
||||
state.field_iterator = null;
|
||||
}
|
||||
if (state.current_line == null) {
|
||||
if (state.options.diagnostics) |d|
|
||||
if (d.errors.items.len > 0) return ParseError.ParseFailed;
|
||||
|
|
@ -646,12 +636,12 @@ pub const RecordIterator = struct {
|
|||
if (state.current_line == null) return self.next();
|
||||
}
|
||||
// non-blank line, but we could have an eof marker
|
||||
if (try Directive.parse(self.arena.allocator(), state.current_line.?, state.*)) |d| {
|
||||
if (try Directive.parse(self.arena.allocator(), state.current_line.?, state)) |d| {
|
||||
switch (d) {
|
||||
.eof => {
|
||||
// there needs to be an eof then
|
||||
if (state.nextLine()) |_| {
|
||||
try parseError(self.arena.allocator(), "Data found after #!eof", state.*);
|
||||
try parseError(self.arena.allocator(), "Data found after #!eof", state);
|
||||
return ParseError.ParseFailed; // this is terminal
|
||||
} else {
|
||||
state.eof_found = true;
|
||||
|
|
@ -660,7 +650,7 @@ pub const RecordIterator = struct {
|
|||
}
|
||||
},
|
||||
else => {
|
||||
try parseError(self.arena.allocator(), "Directive found after data started", state.*);
|
||||
try parseError(self.arena.allocator(), "Directive found after data started", state);
|
||||
state.current_line = state.nextLine();
|
||||
// TODO: This runs the risk of a malicious file creating
|
||||
// a stackoverflow by using many non-eof directives
|
||||
|
|
@ -669,7 +659,8 @@ pub const RecordIterator = struct {
|
|||
}
|
||||
}
|
||||
state.end_of_record_reached = false;
|
||||
return .{ .ri = self };
|
||||
state.field_iterator = .{ .ri = self };
|
||||
return state.field_iterator.?;
|
||||
}
|
||||
|
||||
pub const FieldIterator = struct {
|
||||
|
|
@ -677,6 +668,7 @@ pub const RecordIterator = struct {
|
|||
|
||||
pub fn next(self: FieldIterator) !?Field {
|
||||
const state = self.ri.state;
|
||||
const aa = self.ri.arena.allocator();
|
||||
// Main parsing. We already have the first line of data, which could
|
||||
// be a record (compact format) or a key/value pair (long format)
|
||||
|
||||
|
|
@ -686,12 +678,12 @@ pub const RecordIterator = struct {
|
|||
if (state.end_of_record_reached) return null;
|
||||
// non-blank line, but we could have an eof marker
|
||||
// TODO: deduplicate this code
|
||||
if (try Directive.parse(self.ri.arena.allocator(), state.current_line.?, state.*)) |d| {
|
||||
if (try Directive.parse(aa, state.current_line.?, state)) |d| {
|
||||
switch (d) {
|
||||
.eof => {
|
||||
// there needs to be an eof then
|
||||
if (state.nextLine()) |_| {
|
||||
try parseError(self.ri.arena.allocator(), "Data found after #!eof", state.*);
|
||||
try parseError(aa, "Data found after #!eof", state);
|
||||
return ParseError.ParseFailed; // this is terminal
|
||||
} else {
|
||||
state.eof_found = true;
|
||||
|
|
@ -700,7 +692,7 @@ pub const RecordIterator = struct {
|
|||
}
|
||||
},
|
||||
else => {
|
||||
try parseError(self.ri.arena.allocator(), "Directive found after data started", state.*);
|
||||
try parseError(aa, "Directive found after data started", state);
|
||||
state.current_line = state.nextLine();
|
||||
// TODO: This runs the risk of a malicious file creating
|
||||
// a stackoverflow by using many non-eof directives
|
||||
|
|
@ -717,7 +709,7 @@ pub const RecordIterator = struct {
|
|||
state.column += key.len + 1;
|
||||
state.partial_line_column += key.len + 1;
|
||||
const value = try Value.parse(
|
||||
self.ri.arena.allocator(),
|
||||
aa,
|
||||
it.rest(),
|
||||
state,
|
||||
state.field_delimiter,
|
||||
|
|
@ -725,7 +717,7 @@ pub const RecordIterator = struct {
|
|||
|
||||
var field: ?Field = null;
|
||||
if (!value.error_parsing) {
|
||||
field = .{ .key = try dupe(self.ri.arena.allocator(), state.options, key), .value = value.item_value };
|
||||
field = .{ .key = try dupe(aa, state.options, key), .value = value.item_value };
|
||||
}
|
||||
|
||||
if (value.reader_advanced and state.field_delimiter == ',') {
|
||||
|
|
@ -774,8 +766,84 @@ pub const RecordIterator = struct {
|
|||
}
|
||||
return field;
|
||||
}
|
||||
};
|
||||
|
||||
/// Coerce Record to a type. Does not handle fields with arrays
|
||||
pub fn to(self: FieldIterator, comptime T: type) !T {
|
||||
const ti = @typeInfo(T);
|
||||
|
||||
switch (ti) {
|
||||
.@"struct" => {
|
||||
// What is this magic? The FieldEnum creates a type (an enum)
|
||||
// where each enum member has the name of a field in the struct
|
||||
//
|
||||
// So... struct { a: u8, b: u8 } will yield enum { a, b }
|
||||
const FieldEnum = std.meta.FieldEnum(T);
|
||||
// Then...EnumFieldStruct will create a struct from this, where
|
||||
// each enum value becomes a field. We will specify the field
|
||||
// type, and the default value. Combining these two calls gets
|
||||
// us a struct with all the same field names, but we get a chance
|
||||
// to make all the fields boolean, so we can use it to track
|
||||
// which fields have been set
|
||||
var found: std.enums.EnumFieldStruct(FieldEnum, bool, false) = .{};
|
||||
// SAFETY: all fields updated below or error is returned
|
||||
var obj: T = undefined;
|
||||
|
||||
while (try self.next()) |f| {
|
||||
inline for (std.meta.fields(T)) |type_field| {
|
||||
// To replicate the behavior of the record version of to,
|
||||
// we need to only take the first version of the field,
|
||||
// so if it's specified twice in the data, we will ignore
|
||||
// all but the first instance
|
||||
if (std.mem.eql(u8, f.key, type_field.name) and
|
||||
!@field(found, type_field.name))
|
||||
{
|
||||
@field(obj, type_field.name) =
|
||||
try coerce(type_field.name, type_field.type, f.value);
|
||||
// Now account for this in our magic found struct...
|
||||
@field(found, type_field.name) = true;
|
||||
}
|
||||
}
|
||||
}
|
||||
// Fill in the defaults for remaining fields. Throw if anything
|
||||
// is missing both value (from above) and default (from here)
|
||||
inline for (std.meta.fields(T)) |type_field| {
|
||||
if (!@field(found, type_field.name)) {
|
||||
// We did not find this field above...revert to default value
|
||||
if (type_field.default_value_ptr) |ptr| {
|
||||
@field(obj, type_field.name) = @as(*const type_field.type, @ptrCast(@alignCast(ptr))).*;
|
||||
} else {
|
||||
log.debug("Record could not be coerced. Field {s} not found on srf data, and no default value exists on the type", .{type_field.name});
|
||||
return error.FieldNotFoundOnFieldWithoutDefaultValue;
|
||||
}
|
||||
}
|
||||
}
|
||||
return obj;
|
||||
},
|
||||
.@"union" => {
|
||||
const active_tag_name = if (@hasDecl(T, "srf_tag_field"))
|
||||
T.srf_tag_field
|
||||
else
|
||||
"active_tag";
|
||||
const first_try = try self.next();
|
||||
if (first_try == null) return error.ActiveTagFieldNotFound;
|
||||
const f = first_try.?;
|
||||
if (!std.mem.eql(u8, f.key, active_tag_name))
|
||||
return error.ActiveTagNotFirstField; // required here, but not on the Record version of to
|
||||
if (f.value == null or f.value.? != .string)
|
||||
return error.ActiveTagValueMustBeAString;
|
||||
const active_tag = f.value.?.string;
|
||||
inline for (std.meta.fields(T)) |field_type| {
|
||||
if (std.mem.eql(u8, active_tag, field_type.name)) {
|
||||
return @unionInit(T, field_type.name, try self.to(field_type.type));
|
||||
}
|
||||
}
|
||||
return error.ActiveTagDoesNotExist;
|
||||
},
|
||||
else => @compileError("Deserialization not supported on " ++ @tagName(ti) ++ " types"),
|
||||
}
|
||||
return error.CoercionNotPossible;
|
||||
}
|
||||
};
|
||||
pub fn deinit(self: RecordIterator) void {
|
||||
const child_allocator = self.arena.child_allocator;
|
||||
self.arena.deinit();
|
||||
|
|
@ -809,7 +877,7 @@ const Directive = union(enum) {
|
|||
eof,
|
||||
expires: i64,
|
||||
|
||||
pub fn parse(allocator: std.mem.Allocator, str: []const u8, state: RecordIterator.State) ParseError!?Directive {
|
||||
pub fn parse(allocator: std.mem.Allocator, str: []const u8, state: *RecordIterator.State) ParseError!?Directive {
|
||||
if (!std.mem.startsWith(u8, str, "#!")) return null;
|
||||
// strip any comments off
|
||||
var it = std.mem.splitScalar(u8, str[2..], '#');
|
||||
|
|
@ -962,8 +1030,7 @@ pub const RecordFormatter = struct {
|
|||
};
|
||||
|
||||
pub const Parsed = struct {
|
||||
// TODO: rip this down and return an array from parse
|
||||
records: std.ArrayList(Record),
|
||||
records: []Record,
|
||||
arena: *std.heap.ArenaAllocator,
|
||||
expires: ?i64,
|
||||
|
||||
|
|
@ -974,35 +1041,29 @@ pub const Parsed = struct {
|
|||
}
|
||||
};
|
||||
|
||||
/// parse function. Prefer iterator over this function. Note that this function will
|
||||
/// change soon
|
||||
/// parse function
|
||||
pub fn parse(reader: *std.Io.Reader, allocator: std.mem.Allocator, options: ParseOptions) ParseError!Parsed {
|
||||
var records = std.ArrayList(Record).empty;
|
||||
var it = try iterator(reader, allocator, options);
|
||||
errdefer it.deinit();
|
||||
const aa = it.arena.allocator();
|
||||
var field_count: usize = 1;
|
||||
while (try it.next()) |fi| {
|
||||
var al = std.ArrayList(Field).empty;
|
||||
var al = try std.ArrayList(Field).initCapacity(aa, field_count);
|
||||
while (try fi.next()) |f| {
|
||||
const val = if (f.value != null)
|
||||
switch (f.value.?) {
|
||||
.string => Value{ .string = try aa.dupe(u8, f.value.?.string) },
|
||||
.bytes => Value{ .bytes = try aa.dupe(u8, f.value.?.bytes) },
|
||||
else => f.value,
|
||||
}
|
||||
else
|
||||
f.value;
|
||||
try al.append(aa, .{
|
||||
.key = try aa.dupe(u8, f.key),
|
||||
.value = val,
|
||||
.key = f.key,
|
||||
.value = f.value,
|
||||
});
|
||||
}
|
||||
// assume that most records are same number of fields
|
||||
field_count = @max(field_count, al.items.len);
|
||||
try records.append(aa, .{
|
||||
.fields = try al.toOwnedSlice(aa),
|
||||
});
|
||||
}
|
||||
return .{
|
||||
.records = records,
|
||||
.records = try records.toOwnedSlice(aa),
|
||||
.arena = it.arena,
|
||||
.expires = it.expires,
|
||||
};
|
||||
|
|
@ -1032,16 +1093,16 @@ pub fn iterator(reader: *std.Io.Reader, allocator: std.mem.Allocator, options: P
|
|||
};
|
||||
const first_line = it.state.nextLine() orelse return ParseError.ParseFailed;
|
||||
|
||||
if (try Directive.parse(aa, first_line, it.state.*)) |d| {
|
||||
if (d != .magic) try parseError(aa, "Magic header not found on first line", it.state.*);
|
||||
} else try parseError(aa, "Magic header not found on first line", it.state.*);
|
||||
if (try Directive.parse(aa, first_line, it.state)) |d| {
|
||||
if (d != .magic) try parseError(aa, "Magic header not found on first line", it.state);
|
||||
} else try parseError(aa, "Magic header not found on first line", it.state);
|
||||
|
||||
// Loop through the header material and configure our main parsing
|
||||
it.state.current_line = blk: {
|
||||
while (it.state.nextLine()) |line| {
|
||||
if (try Directive.parse(aa, line, it.state.*)) |d| {
|
||||
if (try Directive.parse(aa, line, it.state)) |d| {
|
||||
switch (d) {
|
||||
.magic => try parseError(aa, "Found a duplicate magic header", it.state.*),
|
||||
.magic => try parseError(aa, "Found a duplicate magic header", it.state),
|
||||
.long_format => it.state.field_delimiter = '\n',
|
||||
.compact_format => it.state.field_delimiter = ',', // what if we have both?
|
||||
.require_eof => it.state.require_eof = true,
|
||||
|
|
@ -1049,7 +1110,7 @@ pub fn iterator(reader: *std.Io.Reader, allocator: std.mem.Allocator, options: P
|
|||
.eof => {
|
||||
// there needs to be an eof then
|
||||
if (it.state.nextLine()) |_| {
|
||||
try parseError(aa, "Data found after #!eof", it.state.*);
|
||||
try parseError(aa, "Data found after #!eof", it.state);
|
||||
return ParseError.ParseFailed; // this is terminal
|
||||
} else return it;
|
||||
},
|
||||
|
|
@ -1066,7 +1127,7 @@ inline fn dupe(allocator: std.mem.Allocator, options: ParseOptions, data: []cons
|
|||
return try allocator.dupe(u8, data);
|
||||
return data;
|
||||
}
|
||||
inline fn parseError(allocator: std.mem.Allocator, message: []const u8, state: RecordIterator.State) ParseError!void {
|
||||
inline fn parseError(allocator: std.mem.Allocator, message: []const u8, state: *RecordIterator.State) ParseError!void {
|
||||
log.debug("Parse error. Parse state {f}, message: {s}", .{ state, message });
|
||||
if (state.options.diagnostics) |d| {
|
||||
try d.addError(allocator, .{
|
||||
|
|
@ -1093,9 +1154,9 @@ test "long format single record, no eof" {
|
|||
var reader = std.Io.Reader.fixed(data);
|
||||
const records = try parse(&reader, allocator, .{});
|
||||
defer records.deinit();
|
||||
try std.testing.expectEqual(@as(usize, 1), records.records.items.len);
|
||||
try std.testing.expectEqual(@as(usize, 1), records.records.items[0].fields.len);
|
||||
const kvps = records.records.items[0].fields;
|
||||
try std.testing.expectEqual(@as(usize, 1), records.records.len);
|
||||
try std.testing.expectEqual(@as(usize, 1), records.records[0].fields.len);
|
||||
const kvps = records.records[0].fields;
|
||||
try std.testing.expectEqualStrings("key", kvps[0].key);
|
||||
try std.testing.expectEqualStrings("string value, with any data except a \\n. an optional string length between the colons", kvps[0].value.?.string);
|
||||
}
|
||||
|
|
@ -1115,7 +1176,7 @@ test "long format from README - generic data structures, first record only" {
|
|||
var reader = std.Io.Reader.fixed(data);
|
||||
const records = try parse(&reader, allocator, .{});
|
||||
defer records.deinit();
|
||||
try std.testing.expectEqual(@as(usize, 1), records.records.items.len);
|
||||
try std.testing.expectEqual(@as(usize, 1), records.records.len);
|
||||
}
|
||||
|
||||
test "long format from README - generic data structures" {
|
||||
|
|
@ -1147,8 +1208,8 @@ test "long format from README - generic data structures" {
|
|||
var reader = std.Io.Reader.fixed(data);
|
||||
const records = try parse(&reader, allocator, .{});
|
||||
defer records.deinit();
|
||||
try std.testing.expectEqual(@as(usize, 2), records.records.items.len);
|
||||
const first = records.records.items[0];
|
||||
try std.testing.expectEqual(@as(usize, 2), records.records.len);
|
||||
const first = records.records[0];
|
||||
try std.testing.expectEqual(@as(usize, 6), first.fields.len);
|
||||
try std.testing.expectEqualStrings("key", first.fields[0].key);
|
||||
try std.testing.expectEqualStrings("string value, with any data except a \\n. an optional string length between the colons", first.fields[0].value.?.string);
|
||||
|
|
@ -1163,7 +1224,7 @@ test "long format from README - generic data structures" {
|
|||
try std.testing.expectEqualStrings("boolean value", first.fields[5].key);
|
||||
try std.testing.expect(!first.fields[5].value.?.boolean);
|
||||
|
||||
const second = records.records.items[1];
|
||||
const second = records.records[1];
|
||||
try std.testing.expectEqual(@as(usize, 5), second.fields.len);
|
||||
try std.testing.expectEqualStrings("key", second.fields[0].key);
|
||||
try std.testing.expectEqualStrings("this is the second record", second.fields[0].value.?.string);
|
||||
|
|
@ -1190,8 +1251,8 @@ test "compact format from README - generic data structures" {
|
|||
// We want "parse" and "parseLeaky" probably. Second parameter is a diagnostics
|
||||
const records = try parse(&reader, allocator, .{});
|
||||
defer records.deinit();
|
||||
try std.testing.expectEqual(@as(usize, 2), records.records.items.len);
|
||||
const first = records.records.items[0];
|
||||
try std.testing.expectEqual(@as(usize, 2), records.records.len);
|
||||
const first = records.records[0];
|
||||
try std.testing.expectEqual(@as(usize, 6), first.fields.len);
|
||||
try std.testing.expectEqualStrings("key", first.fields[0].key);
|
||||
try std.testing.expectEqualStrings("string value must have a length between colons or end with a comma", first.fields[0].value.?.string);
|
||||
|
|
@ -1206,7 +1267,7 @@ test "compact format from README - generic data structures" {
|
|||
try std.testing.expectEqualStrings("boolean value", first.fields[5].key);
|
||||
try std.testing.expect(!first.fields[5].value.?.boolean);
|
||||
|
||||
const second = records.records.items[1];
|
||||
const second = records.records[1];
|
||||
try std.testing.expectEqual(@as(usize, 1), second.fields.len);
|
||||
try std.testing.expectEqualStrings("key", second.fields[0].key);
|
||||
try std.testing.expectEqualStrings("this is the second record", second.fields[0].value.?.string);
|
||||
|
|
@ -1273,7 +1334,7 @@ test "format all the things" {
|
|||
var formatted_reader = std.Io.Reader.fixed(formatted);
|
||||
const parsed = try parse(&formatted_reader, std.testing.allocator, .{});
|
||||
defer parsed.deinit();
|
||||
try std.testing.expectEqualDeep(records, parsed.records.items);
|
||||
try std.testing.expectEqualDeep(records, parsed.records);
|
||||
|
||||
const compact = try std.fmt.bufPrint(
|
||||
&buf,
|
||||
|
|
@ -1290,7 +1351,7 @@ test "format all the things" {
|
|||
var compact_reader = std.Io.Reader.fixed(compact);
|
||||
const parsed_compact = try parse(&compact_reader, std.testing.allocator, .{});
|
||||
defer parsed_compact.deinit();
|
||||
try std.testing.expectEqualDeep(records, parsed_compact.records.items);
|
||||
try std.testing.expectEqualDeep(records, parsed_compact.records);
|
||||
|
||||
const expected_expires: i64 = 1772589213;
|
||||
const compact_expires = try std.fmt.bufPrint(
|
||||
|
|
@ -1309,7 +1370,7 @@ test "format all the things" {
|
|||
var expires_reader = std.Io.Reader.fixed(compact_expires);
|
||||
const parsed_expires = try parse(&expires_reader, std.testing.allocator, .{});
|
||||
defer parsed_expires.deinit();
|
||||
try std.testing.expectEqualDeep(records, parsed_expires.records.items);
|
||||
try std.testing.expectEqualDeep(records, parsed_expires.records);
|
||||
try std.testing.expectEqual(expected_expires, parsed_expires.expires.?);
|
||||
}
|
||||
test "serialize/deserialize" {
|
||||
|
|
@ -1355,17 +1416,34 @@ test "serialize/deserialize" {
|
|||
const parsed = try parse(&compact_reader, std.testing.allocator, .{});
|
||||
defer parsed.deinit();
|
||||
|
||||
const rec1 = try parsed.records.items[0].to(Data);
|
||||
const rec1 = try parsed.records[0].to(Data);
|
||||
try std.testing.expectEqualStrings("bar", rec1.foo);
|
||||
try std.testing.expectEqual(@as(u8, 42), rec1.bar);
|
||||
try std.testing.expectEqual(@as(RecType, .foo), rec1.qux);
|
||||
const rec4 = try parsed.records.items[3].to(Data);
|
||||
const rec4 = try parsed.records[3].to(Data);
|
||||
try std.testing.expectEqualStrings("bar", rec4.foo);
|
||||
try std.testing.expectEqual(@as(u8, 42), rec4.bar);
|
||||
try std.testing.expectEqual(@as(RecType, .bar), rec4.qux.?);
|
||||
try std.testing.expectEqual(true, rec4.b);
|
||||
try std.testing.expectEqual(@as(f32, 6.9), rec4.f);
|
||||
|
||||
// Now we'll do it with the iterator version
|
||||
var it_reader = std.Io.Reader.fixed(compact);
|
||||
const ri = try iterator(&it_reader, std.testing.allocator, .{});
|
||||
defer ri.deinit();
|
||||
const rec1_it = try (try ri.next()).?.to(Data);
|
||||
try std.testing.expectEqualStrings("bar", rec1_it.foo);
|
||||
try std.testing.expectEqual(@as(u8, 42), rec1_it.bar);
|
||||
try std.testing.expectEqual(@as(RecType, .foo), rec1_it.qux);
|
||||
_ = try ri.next();
|
||||
_ = try ri.next();
|
||||
const rec4_it = try (try ri.next()).?.to(Data);
|
||||
try std.testing.expectEqualStrings("bar", rec4_it.foo);
|
||||
try std.testing.expectEqual(@as(u8, 42), rec4_it.bar);
|
||||
try std.testing.expectEqual(@as(RecType, .bar), rec4_it.qux.?);
|
||||
try std.testing.expectEqual(true, rec4_it.b);
|
||||
try std.testing.expectEqual(@as(f32, 6.9), rec4_it.f);
|
||||
|
||||
const alloc = std.testing.allocator;
|
||||
var owned_record_1 = try Record.from(Data, alloc, rec1);
|
||||
defer owned_record_1.deinit();
|
||||
|
|
@ -1443,9 +1521,9 @@ test "unions" {
|
|||
const parsed = try parse(&compact_reader, std.testing.allocator, .{});
|
||||
defer parsed.deinit();
|
||||
|
||||
const rec1 = try parsed.records.items[0].to(MixedData);
|
||||
const rec1 = try parsed.records[0].to(MixedData);
|
||||
try std.testing.expectEqualDeep(data[0], rec1);
|
||||
const rec2 = try parsed.records.items[1].to(MixedData);
|
||||
const rec2 = try parsed.records[1].to(MixedData);
|
||||
try std.testing.expectEqualDeep(data[1], rec2);
|
||||
}
|
||||
test "enums" {
|
||||
|
|
@ -1488,9 +1566,9 @@ test "enums" {
|
|||
const parsed = try parse(&compact_reader, std.testing.allocator, .{});
|
||||
defer parsed.deinit();
|
||||
|
||||
const rec1 = try parsed.records.items[0].to(Data);
|
||||
const rec1 = try parsed.records[0].to(Data);
|
||||
try std.testing.expectEqualDeep(data[0], rec1);
|
||||
const rec2 = try parsed.records.items[1].to(Data);
|
||||
const rec2 = try parsed.records[1].to(Data);
|
||||
try std.testing.expectEqualDeep(data[1], rec2);
|
||||
|
||||
const missing_tag =
|
||||
|
|
@ -1501,10 +1579,10 @@ test "enums" {
|
|||
var mt_reader = std.Io.Reader.fixed(missing_tag);
|
||||
const mt_parsed = try parse(&mt_reader, std.testing.allocator, .{});
|
||||
defer mt_parsed.deinit();
|
||||
const mt_rec1 = try mt_parsed.records.items[0].to(Data);
|
||||
const mt_rec1 = try mt_parsed.records[0].to(Data);
|
||||
try std.testing.expect(mt_rec1.data_type == null);
|
||||
|
||||
const mt_rec1_dt2 = try mt_parsed.records.items[0].to(Data2);
|
||||
const mt_rec1_dt2 = try mt_parsed.records[0].to(Data2);
|
||||
try std.testing.expect(mt_rec1_dt2.data_type == .bar);
|
||||
}
|
||||
test "compact format length-prefixed string as last field" {
|
||||
|
|
@ -1520,8 +1598,8 @@ test "compact format length-prefixed string as last field" {
|
|||
var reader = std.Io.Reader.fixed(data);
|
||||
const records = try parse(&reader, allocator, .{});
|
||||
defer records.deinit();
|
||||
try std.testing.expectEqual(@as(usize, 1), records.records.items.len);
|
||||
const rec = records.records.items[0];
|
||||
try std.testing.expectEqual(@as(usize, 1), records.records.len);
|
||||
const rec = records.records[0];
|
||||
try std.testing.expectEqual(@as(usize, 2), rec.fields.len);
|
||||
try std.testing.expectEqualStrings("name", rec.fields[0].key);
|
||||
try std.testing.expectEqualStrings("alice", rec.fields[0].value.?.string);
|
||||
|
|
|
|||
Loading…
Add table
Reference in a new issue