1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
//! Module for handling random access text
use std::fmt;
use std::str::FromStr;
use std::collections::HashMap;
use super::{FileImage,Records,TextConversion,Error};
use crate::{STDRESULT,DYNERR};
/// See if `fimg` can be interpreted as records, and if so return the record length.
/// The converter is only used to decide whether we should expect postive or negative ASCII.
/// It is always assumed the record separator is a run of ascii null.
/// This basically relies on finding a run of contiguous records with
/// clean runs of separators between every record.
fn deduce_record_length(fimg: &FileImage,converter: &impl TextConversion) -> Option<usize> {
let mut l = usize::MAX; // current estimate of record length
let mut num = 0; // number of contiguous records found
let mut pos = 0; // position in tentative record
let mut in_fields = false;
let lims = match converter.to_utf8(&[10]) {
None => [129,255],
Some(s) if s.as_str()=="\u{0000}" => [129,255],
_ => [1,127]
};
for k in fimg.chunks.keys() {
let Some(chunk) = fimg.chunks.get(k) else { return None; };
for b in chunk {
if in_fields && *b == 0 {
in_fields = false;
pos += 1;
} else if in_fields && *b > lims[0] && *b < lims[1] {
pos += 1;
} else if !in_fields && *b > lims[0] && *b < lims[1] {
in_fields = true;
num += 1;
if pos > 0 {
if l == usize::MAX {
l = pos;
} else if pos < l && l % pos == 0 {
l = pos;
} else if pos < l && l % pos != 0 {
return None;
} else if pos > l && pos % l != 0 {
return None;
}
}
pos = 1; // pointer is after character we have found
} else if !in_fields && *b == 0 {
pos += 1;
} else {
return None;
}
}
}
if num > 3 && l > 2 && l <= 0xffff {
Some(l)
} else {
None
}
}
impl Records {
pub fn new(record_len: usize) -> Self {
Self {
record_len,
map: HashMap::new()
}
}
/// see if the slice is a JSON string with fimg_type = rec
pub fn test(dat: &[u8]) -> bool {
match str::from_utf8(dat) {
Ok(s) => match json::parse(s) {
Ok(parsed) => {
match parsed["fimg_type"].as_str() {
Some(val) => val == "rec",
None => false
}
},
_ => false
},
_ => false
}
}
/// add a string as record number `num`, fields should be separated by LF or CRLF.
pub fn add_record(&mut self,num: usize,fields: &str) {
self.map.insert(num,fields.to_string());
}
/// Derive records from file image, this should find any real record, but may also find spurious ones.
/// This is due to fundamental non-invertibility of the A2 file system's random access storage pattern.
/// This routine assumes ASCII null terminates any record. If `record_length` is 0 an analysis is
/// triggered that tries to match a random access text pattern against rules deduced from the `converter` object.
/// If the matching succeeds the Records are returned, if not an error is returned.
pub fn from_fimg(fimg: &FileImage,mut record_length: usize,converter: impl TextConversion) -> Result<Records,DYNERR> {
// TODO: allow for record separators other than ascii null
if record_length == 0 {
record_length = match deduce_record_length(fimg,&converter) {
Some(l) => l,
None => return Err(Box::new(Error::FileFormat)) // don't want a log message here
};
}
if record_length < 2 || record_length > 0xffff {
log::error!("refusing record length {}",record_length);
return Err(Box::new(Error::FileFormat));
}
let mut ans = Records::new(record_length);
let mut list: Vec<usize> = Vec::new();
// add record index for each starting record boundary that falls within a chunk
let chunk_len = fimg.chunk_len;
for c in fimg.chunks.keys() {
let start_rec = c*chunk_len/record_length + match c*chunk_len%record_length { x if x>0 => 1, _ => 0 };
let end_rec = (c+1)*chunk_len/record_length + match (c+1)*chunk_len%record_length { x if x>0 => 1, _ => 0 };
for r in start_rec..end_rec {
list.push(r);
}
}
// add only records with complete data
for r in list {
let start_chunk = r*record_length/chunk_len;
let end_chunk = 1 + (r+1)*record_length/chunk_len;
let start_offset = r*record_length%chunk_len;
let mut bytes: Vec<u8> = Vec::new();
let mut complete = true;
for chunk_num in start_chunk..end_chunk {
match fimg.chunks.get(&chunk_num) {
Some(chunk) => {
for i in chunk {
bytes.push(*i);
}
},
_ => complete = false
}
}
if complete && start_offset < bytes.len() {
let actual_end = usize::min(start_offset+record_length,bytes.len());
if let Some(long_str) = converter.to_utf8(&bytes[start_offset..actual_end].to_vec()) {
if let Some(partial) = long_str.split("\u{0000}").next() {
if partial.len()>0 {
ans.map.insert(r,partial.to_string());
}
} else {
if long_str.len()>0 {
ans.map.insert(r,long_str);
}
}
}
}
}
return Ok(ans);
}
/// Update a file image's data using the records, this is usually done before writing to a disk image.
/// This will set the file image's eof, but no other metadata.
pub fn update_fimg(&self,ans: &mut FileImage,require_first: bool,converter: impl TextConversion,clear: bool) -> STDRESULT {
if self.record_len < 2 || self.record_len > 0xffff {
log::error!("refusing record length {}",self.record_len);
return Err(Box::new(Error::FileFormat));
}
if clear {
ans.chunks = HashMap::new();
}
let chunk_len = ans.chunk_len;
let mut eof: usize = 0;
// always need to have the first chunk referenced on ProDOS
if require_first {
ans.chunks.insert(0,vec![0;chunk_len]);
}
let mut next_buf = |old_chunk: usize,old_buf: Option<Vec<u8>>| -> Vec<u8> {
let mut chunk = old_chunk;
if let Some(old) = old_buf {
ans.chunks.insert(chunk,old);
chunk += 1;
}
match ans.chunks.get_mut(&chunk) {
Some(v) => v.clone(),
None => Vec::new()
}
};
// now insert the actual records, first chunk can always be overwritten
for (rec_num,fields) in &self.map {
match converter.from_utf8(fields) {
Some(data_bytes) => {
if data_bytes.len() > self.record_len {
log::warn!("record {} is too long and will corrupt other records",rec_num);
//log::warn!("truncating record {} to maximum {}",rec_num,self.record_len);
//data_bytes = data_bytes[0..self.record_len].to_vec();
}
let mut chunk = self.record_len * rec_num / chunk_len;
let mut offset = self.record_len * rec_num % chunk_len;
let mut buf = next_buf(chunk,None);
for i in 0..data_bytes.len() {
if offset >= buf.len() {
for _j in buf.len()..offset {
buf.push(0);
}
buf.push(data_bytes[i]);
} else {
buf[offset] = data_bytes[i];
}
offset += 1;
if offset >= chunk_len || i+1 == data_bytes.len() {
eof = usize::max(chunk*chunk_len + buf.len(),eof);
buf = next_buf(chunk,Some(buf));
chunk += 1;
offset = 0;
}
}
},
None => return Err(Box::new(std::fmt::Error))
}
}
ans.set_eof(eof);
return Ok(());
}
/// Get records from the JSON string representation
pub fn from_json(json_str: &str) -> Result<Records,DYNERR> {
match json::parse(json_str) {
Ok(parsed) => {
let maybe_type = parsed["fimg_type"].as_str();
let maybe_len = parsed["record_length"].as_usize();
if let (Some(typ),Some(len)) = (maybe_type,maybe_len) {
if typ=="rec" {
let mut records: HashMap<usize,String> = HashMap::new();
let map_obj = &parsed["records"];
if map_obj.entries().len()==0 {
log::error!("no object entries in json records");
return Err(Box::new(Error::FileImageFormat));
}
for (key,lines) in map_obj.entries() {
if let Ok(num) = usize::from_str(key) {
let mut fields = String::new();
for maybe_field in lines.members() {
if let Some(line) = maybe_field.as_str() {
fields = fields + line + "\n";
} else {
log::error!("record is not a string");
return Err(Box::new(Error::FileImageFormat));
}
}
records.insert(num,fields);
} else {
log::error!("key is not a number");
return Err(Box::new(Error::FileImageFormat));
}
}
return Ok(Self {
record_len: len,
map: records
});
} else {
log::error!("json metadata type mismatch");
return Err(Box::new(Error::FileImageFormat));
}
}
log::error!("json records missing metadata");
Err(Box::new(Error::FileImageFormat))
},
Err(_e) => Err(Box::new(Error::FileImageFormat))
}
}
/// Put records into the JSON string representation, if indent=0 use unpretty form
pub fn to_json(&self,indent: Option<u16>) -> String {
let mut json_map = json::JsonValue::new_object();
for (r,l) in &self.map {
let mut json_array = json::JsonValue::new_array();
for line in l.lines() {
json_array.push(line).expect("error while building JSON array");
}
json_map[r.to_string()] = json_array;
}
let ans = json::object! {
fimg_type: "rec",
record_length: self.record_len,
records: json_map
};
if let Some(spaces) = indent {
return json::stringify_pretty(ans, spaces);
} else {
return json::stringify(ans);
}
}
}
/// Allows the records to be displayed to the console using `println!`. This also
/// derives `to_string`, so the structure can be converted to `String`.
impl fmt::Display for Records {
fn fmt(&self,f: &mut fmt::Formatter<'_>) -> fmt::Result {
for (idx,fields) in &self.map {
write!(f,"Record {}",idx).expect("format error");
for field in fields.lines() {
write!(f," {}",field).expect("format error");
}
}
write!(f,"Record Count = {}",self.map.len())
}
}