]>
Commit | Line | Data |
---|---|---|
4a35548b MS |
1 | /* |
2 | * This file is part of the libsigrok project. | |
3 | * | |
4 | * Copyright (C) 2013 Marc Schink <sigrok-dev@marcschink.de> | |
5 | * | |
6 | * This program is free software: you can redistribute it and/or modify | |
7 | * it under the terms of the GNU General Public License as published by | |
8 | * the Free Software Foundation, either version 3 of the License, or | |
9 | * (at your option) any later version. | |
10 | * | |
11 | * This program is distributed in the hope that it will be useful, | |
12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
14 | * GNU General Public License for more details. | |
15 | * | |
16 | * You should have received a copy of the GNU General Public License | |
17 | * along with this program. If not, see <http://www.gnu.org/licenses/>. | |
18 | */ | |
19 | ||
6ec6c43b | 20 | #include <config.h> |
4a35548b MS |
21 | #include <stdlib.h> |
22 | #include <string.h> | |
23 | #include <glib.h> | |
c1aae900 | 24 | #include <libsigrok/libsigrok.h> |
4a35548b MS |
25 | #include "libsigrok-internal.h" |
26 | ||
3544f848 | 27 | #define LOG_PREFIX "input/csv" |
4a35548b MS |
28 | |
29 | /* | |
30 | * The CSV input module has the following options: | |
31 | * | |
32 | * single-column: Specifies the column number which stores the sample data for | |
33 | * single column mode and enables single column mode. Multi | |
34 | * column mode is used if this parameter is omitted. | |
35 | * | |
ba7dd8bb UH |
36 | * numchannels: Specifies the number of channels to use. In multi column mode |
37 | * the number of channels are the number of columns and in single | |
4a35548b | 38 | * column mode the number of bits (LSB first) beginning at |
ba7dd8bb | 39 | * 'first-channel'. |
4a35548b MS |
40 | * |
41 | * delimiter: Specifies the delimiter for columns. Must be at least one | |
42 | * character. Comma is used as default delimiter. | |
43 | * | |
44 | * format: Specifies the format of the sample data in single column mode. | |
45 | * Available formats are: 'bin', 'hex' and 'oct'. The binary | |
46 | * format is used by default. This option has no effect in multi | |
47 | * column mode. | |
48 | * | |
49 | * comment: Specifies the prefix character(s) for comments. No prefix | |
50 | * characters are used by default which disables removing of | |
51 | * comments. | |
52 | * | |
53 | * samplerate: Samplerate which the sample data was captured with. Default | |
54 | * value is 0. | |
55 | * | |
ba7dd8bb UH |
56 | * first-channel: Column number of the first channel in multi column mode and |
57 | * position of the bit for the first channel in single column mode. | |
4a35548b MS |
58 | * Default value is 0. |
59 | * | |
60 | * header: Determines if the first line should be treated as header | |
ba7dd8bb UH |
61 | * and used for channel names in multi column mode. Empty header |
62 | * names will be replaced by the channel number. If enabled in | |
4a35548b MS |
63 | * single column mode the first line will be skipped. Usage of |
64 | * header is disabled by default. | |
65 | * | |
66 | * startline: Line number to start processing sample data. Must be greater | |
67 | * than 0. The default line number to start processing is 1. | |
68 | */ | |
69 | ||
ccff468b GS |
70 | /* |
71 | * TODO | |
72 | * | |
73 | * - Determine how the text line handling can get improved, regarding | |
74 | * all of robustness and flexibility and correctness. | |
75 | * - The current implementation splits on "any run of CR and LF". Which | |
76 | * translates to: Line numbers are wrong in the presence of empty | |
77 | * lines in the input stream. | |
78 | * - The current implementation insists in the presence of end-of-line | |
79 | * markers on _every_ line in the input stream. "Incomplete" text | |
80 | * files that are so typical on the Windows platform get rejected as | |
81 | * invalid. | |
82 | * - Dropping support for CR style end-of-line markers could improve | |
83 | * the situation a lot. Code could search for and split on LF, and | |
84 | * trim optional trailing CR. This would result in proper support | |
85 | * for CRLF (Windows) as well as LF (Unix), and allow for correct | |
86 | * line number counts. | |
87 | * - When support for CR-only line termination cannot get dropped, | |
88 | * then the current implementation is inappropriate. Currently the | |
89 | * input stream is scanned for the first occurance of either of the | |
90 | * supported termination styles (which is good). For the remaining | |
91 | * session a consistent encoding of the text lines is assumed (which | |
92 | * is acceptable). Potential absence of the terminator for the last | |
93 | * line is orthogonal, and can get handled by a "force" flag when | |
94 | * the end() routine calls the process_buffer() routine. | |
95 | * - When line numbers need to be correct and reliable, _and_ the full | |
96 | * set of previously supported line termination sequences are required, | |
97 | * and potentially more are to get added for improved compatibility | |
98 | * with more platforms or generators, then the current approach of | |
99 | * splitting on runs of termination characters needs to get replaced, | |
100 | * by the more expensive approach to scan for and count the initially | |
101 | * determined termination sequence. | |
102 | * | |
103 | * - Add support for analog input data? (optional) | |
104 | * - Needs a syntax first for user specs which channels (columns) are | |
105 | * logic and which are analog. May need heuristics(?) to guess from | |
106 | * input data in the absence of user provided specs. | |
107 | */ | |
108 | ||
4a35548b MS |
109 | /* Single column formats. */ |
110 | enum { | |
111 | FORMAT_BIN, | |
112 | FORMAT_HEX, | |
113 | FORMAT_OCT | |
114 | }; | |
115 | ||
116 | struct context { | |
41d214f6 BV |
117 | gboolean started; |
118 | ||
4a35548b MS |
119 | /* Current selected samplerate. */ |
120 | uint64_t samplerate; | |
121 | ||
ba7dd8bb | 122 | /* Number of channels. */ |
6433156c | 123 | unsigned int num_channels; |
4a35548b MS |
124 | |
125 | /* Column delimiter character(s). */ | |
126 | GString *delimiter; | |
127 | ||
128 | /* Comment prefix character(s). */ | |
129 | GString *comment; | |
130 | ||
d9251a2c | 131 | /* Termination character(s) used in current stream. */ |
41d214f6 BV |
132 | char *termination; |
133 | ||
4a35548b MS |
134 | /* Determines if sample data is stored in multiple columns. */ |
135 | gboolean multi_column_mode; | |
136 | ||
137 | /* Column number of the sample data in single column mode. */ | |
6433156c | 138 | unsigned int single_column; |
4a35548b MS |
139 | |
140 | /* | |
141 | * Number of the first column to parse. Equivalent to the number of the | |
ba7dd8bb | 142 | * first channel in multi column mode and the single column number in |
4a35548b MS |
143 | * single column mode. |
144 | */ | |
6433156c | 145 | unsigned int first_column; |
4a35548b MS |
146 | |
147 | /* | |
ba7dd8bb UH |
148 | * Column number of the first channel in multi column mode and position of |
149 | * the bit for the first channel in single column mode. | |
4a35548b | 150 | */ |
6433156c | 151 | unsigned int first_channel; |
4a35548b MS |
152 | |
153 | /* Line number to start processing. */ | |
6433156c | 154 | size_t start_line; |
4a35548b MS |
155 | |
156 | /* | |
157 | * Determines if the first line should be treated as header and used for | |
ba7dd8bb | 158 | * channel names in multi column mode. |
4a35548b MS |
159 | */ |
160 | gboolean header; | |
161 | ||
162 | /* Format sample data is stored in single column mode. */ | |
163 | int format; | |
164 | ||
165 | /* Size of the sample buffer. */ | |
6433156c | 166 | size_t sample_buffer_size; |
4a35548b MS |
167 | |
168 | /* Buffer to store sample data. */ | |
169 | uint8_t *sample_buffer; | |
170 | ||
4a35548b | 171 | /* Current line number. */ |
6433156c | 172 | size_t line_number; |
4a35548b MS |
173 | }; |
174 | ||
41d214f6 | 175 | static void strip_comment(char *buf, const GString *prefix) |
4a35548b MS |
176 | { |
177 | char *ptr; | |
178 | ||
179 | if (!prefix->len) | |
180 | return; | |
181 | ||
41d214f6 BV |
182 | if ((ptr = strstr(buf, prefix->str))) |
183 | *ptr = '\0'; | |
4a35548b MS |
184 | } |
185 | ||
41d214f6 | 186 | static int parse_binstr(const char *str, struct context *inc) |
4a35548b MS |
187 | { |
188 | gsize i, j, length; | |
189 | ||
190 | length = strlen(str); | |
191 | ||
192 | if (!length) { | |
6433156c | 193 | sr_err("Column %u in line %zu is empty.", inc->single_column, |
41d214f6 | 194 | inc->line_number); |
4a35548b MS |
195 | return SR_ERR; |
196 | } | |
197 | ||
198 | /* Clear buffer in order to set bits only. */ | |
41d214f6 | 199 | memset(inc->sample_buffer, 0, (inc->num_channels + 7) >> 3); |
4a35548b | 200 | |
41d214f6 | 201 | i = inc->first_channel; |
4a35548b | 202 | |
41d214f6 | 203 | for (j = 0; i < length && j < inc->num_channels; i++, j++) { |
4a35548b | 204 | if (str[length - i - 1] == '1') { |
41d214f6 | 205 | inc->sample_buffer[j / 8] |= (1 << (j % 8)); |
4a35548b | 206 | } else if (str[length - i - 1] != '0') { |
6433156c | 207 | sr_err("Invalid value '%s' in column %u in line %zu.", |
41d214f6 | 208 | str, inc->single_column, inc->line_number); |
4a35548b MS |
209 | return SR_ERR; |
210 | } | |
211 | } | |
212 | ||
213 | return SR_OK; | |
214 | } | |
215 | ||
41d214f6 | 216 | static int parse_hexstr(const char *str, struct context *inc) |
4a35548b MS |
217 | { |
218 | gsize i, j, k, length; | |
219 | uint8_t value; | |
220 | char c; | |
221 | ||
222 | length = strlen(str); | |
223 | ||
224 | if (!length) { | |
6433156c | 225 | sr_err("Column %u in line %zu is empty.", inc->single_column, |
41d214f6 | 226 | inc->line_number); |
4a35548b MS |
227 | return SR_ERR; |
228 | } | |
229 | ||
230 | /* Clear buffer in order to set bits only. */ | |
41d214f6 | 231 | memset(inc->sample_buffer, 0, (inc->num_channels + 7) >> 3); |
4a35548b MS |
232 | |
233 | /* Calculate the position of the first hexadecimal digit. */ | |
41d214f6 | 234 | i = inc->first_channel / 4; |
4a35548b | 235 | |
41d214f6 | 236 | for (j = 0; i < length && j < inc->num_channels; i++) { |
4a35548b MS |
237 | c = str[length - i - 1]; |
238 | ||
239 | if (!g_ascii_isxdigit(c)) { | |
6433156c | 240 | sr_err("Invalid value '%s' in column %u in line %zu.", |
41d214f6 | 241 | str, inc->single_column, inc->line_number); |
4a35548b MS |
242 | return SR_ERR; |
243 | } | |
244 | ||
245 | value = g_ascii_xdigit_value(c); | |
246 | ||
41d214f6 | 247 | k = (inc->first_channel + j) % 4; |
4a35548b | 248 | |
41d214f6 | 249 | for (; j < inc->num_channels && k < 4; k++) { |
4a35548b | 250 | if (value & (1 << k)) |
41d214f6 | 251 | inc->sample_buffer[j / 8] |= (1 << (j % 8)); |
4a35548b MS |
252 | |
253 | j++; | |
254 | } | |
255 | } | |
256 | ||
257 | return SR_OK; | |
258 | } | |
259 | ||
41d214f6 | 260 | static int parse_octstr(const char *str, struct context *inc) |
4a35548b MS |
261 | { |
262 | gsize i, j, k, length; | |
263 | uint8_t value; | |
264 | char c; | |
265 | ||
266 | length = strlen(str); | |
267 | ||
268 | if (!length) { | |
6433156c | 269 | sr_err("Column %u in line %zu is empty.", inc->single_column, |
41d214f6 | 270 | inc->line_number); |
4a35548b MS |
271 | return SR_ERR; |
272 | } | |
273 | ||
274 | /* Clear buffer in order to set bits only. */ | |
41d214f6 | 275 | memset(inc->sample_buffer, 0, (inc->num_channels + 7) >> 3); |
4a35548b MS |
276 | |
277 | /* Calculate the position of the first octal digit. */ | |
41d214f6 | 278 | i = inc->first_channel / 3; |
4a35548b | 279 | |
41d214f6 | 280 | for (j = 0; i < length && j < inc->num_channels; i++) { |
4a35548b MS |
281 | c = str[length - i - 1]; |
282 | ||
283 | if (c < '0' || c > '7') { | |
6433156c | 284 | sr_err("Invalid value '%s' in column %u in line %zu.", |
41d214f6 | 285 | str, inc->single_column, inc->line_number); |
4a35548b MS |
286 | return SR_ERR; |
287 | } | |
288 | ||
289 | value = g_ascii_xdigit_value(c); | |
290 | ||
41d214f6 | 291 | k = (inc->first_channel + j) % 3; |
4a35548b | 292 | |
41d214f6 | 293 | for (; j < inc->num_channels && k < 3; k++) { |
4a35548b | 294 | if (value & (1 << k)) |
41d214f6 | 295 | inc->sample_buffer[j / 8] |= (1 << (j % 8)); |
4a35548b MS |
296 | |
297 | j++; | |
298 | } | |
299 | } | |
300 | ||
301 | return SR_OK; | |
302 | } | |
303 | ||
41d214f6 | 304 | static char **parse_line(char *buf, struct context *inc, int max_columns) |
4a35548b MS |
305 | { |
306 | const char *str, *remainder; | |
307 | GSList *list, *l; | |
308 | char **columns; | |
309 | char *column; | |
310 | gsize n, k; | |
311 | ||
312 | n = 0; | |
313 | k = 0; | |
314 | list = NULL; | |
315 | ||
41d214f6 BV |
316 | remainder = buf; |
317 | str = strstr(remainder, inc->delimiter->str); | |
4a35548b MS |
318 | |
319 | while (str && max_columns) { | |
41d214f6 | 320 | if (n >= inc->first_column) { |
4a35548b MS |
321 | column = g_strndup(remainder, str - remainder); |
322 | list = g_slist_prepend(list, g_strstrip(column)); | |
323 | ||
324 | max_columns--; | |
325 | k++; | |
326 | } | |
327 | ||
41d214f6 BV |
328 | remainder = str + inc->delimiter->len; |
329 | str = strstr(remainder, inc->delimiter->str); | |
4a35548b MS |
330 | n++; |
331 | } | |
332 | ||
41d214f6 | 333 | if (buf[0] && max_columns && n >= inc->first_column) { |
4a35548b MS |
334 | column = g_strdup(remainder); |
335 | list = g_slist_prepend(list, g_strstrip(column)); | |
336 | k++; | |
337 | } | |
338 | ||
339 | if (!(columns = g_try_new(char *, k + 1))) | |
340 | return NULL; | |
341 | ||
342 | columns[k--] = NULL; | |
343 | ||
344 | for (l = list; l; l = l->next) | |
345 | columns[k--] = l->data; | |
346 | ||
347 | g_slist_free(list); | |
348 | ||
349 | return columns; | |
350 | } | |
351 | ||
41d214f6 | 352 | static int parse_multi_columns(char **columns, struct context *inc) |
4a35548b MS |
353 | { |
354 | gsize i; | |
df0db9fd | 355 | char *column; |
4a35548b MS |
356 | |
357 | /* Clear buffer in order to set bits only. */ | |
41d214f6 | 358 | memset(inc->sample_buffer, 0, (inc->num_channels + 7) >> 3); |
4a35548b | 359 | |
41d214f6 | 360 | for (i = 0; i < inc->num_channels; i++) { |
df0db9fd GS |
361 | column = columns[i]; |
362 | if (column[0] == '1') { | |
41d214f6 | 363 | inc->sample_buffer[i / 8] |= (1 << (i % 8)); |
df0db9fd | 364 | } else if (!strlen(column)) { |
4a35548b | 365 | sr_err("Column %zu in line %zu is empty.", |
41d214f6 | 366 | inc->first_channel + i, inc->line_number); |
4a35548b | 367 | return SR_ERR; |
df0db9fd | 368 | } else if (column[0] != '0') { |
4a35548b | 369 | sr_err("Invalid value '%s' in column %zu in line %zu.", |
df0db9fd | 370 | column, inc->first_channel + i, |
41d214f6 | 371 | inc->line_number); |
4a35548b MS |
372 | return SR_ERR; |
373 | } | |
374 | } | |
375 | ||
376 | return SR_OK; | |
377 | } | |
378 | ||
41d214f6 | 379 | static int parse_single_column(const char *column, struct context *inc) |
4a35548b MS |
380 | { |
381 | int res; | |
382 | ||
383 | res = SR_ERR; | |
384 | ||
0c5f2abc | 385 | switch (inc->format) { |
4a35548b | 386 | case FORMAT_BIN: |
41d214f6 | 387 | res = parse_binstr(column, inc); |
4a35548b MS |
388 | break; |
389 | case FORMAT_HEX: | |
41d214f6 | 390 | res = parse_hexstr(column, inc); |
4a35548b MS |
391 | break; |
392 | case FORMAT_OCT: | |
41d214f6 | 393 | res = parse_octstr(column, inc); |
4a35548b MS |
394 | break; |
395 | } | |
396 | ||
397 | return res; | |
398 | } | |
399 | ||
400 | static int send_samples(const struct sr_dev_inst *sdi, uint8_t *buffer, | |
41d214f6 | 401 | gsize buffer_size, gsize count) |
4a35548b | 402 | { |
4a35548b MS |
403 | struct sr_datafeed_packet packet; |
404 | struct sr_datafeed_logic logic; | |
41d214f6 | 405 | int res; |
4a35548b MS |
406 | gsize i; |
407 | ||
408 | packet.type = SR_DF_LOGIC; | |
409 | packet.payload = &logic; | |
410 | logic.unitsize = buffer_size; | |
411 | logic.length = buffer_size; | |
412 | logic.data = buffer; | |
413 | ||
414 | for (i = 0; i < count; i++) { | |
df0db9fd GS |
415 | res = sr_session_send(sdi, &packet); |
416 | if (res != SR_OK) | |
4a35548b MS |
417 | return res; |
418 | } | |
419 | ||
420 | return SR_OK; | |
421 | } | |
422 | ||
41d214f6 | 423 | static int init(struct sr_input *in, GHashTable *options) |
4a35548b | 424 | { |
41d214f6 BV |
425 | struct context *inc; |
426 | const char *s; | |
4a35548b | 427 | |
aac29cc1 | 428 | in->sdi = g_malloc0(sizeof(struct sr_dev_inst)); |
41d214f6 | 429 | in->priv = inc = g_malloc0(sizeof(struct context)); |
4a35548b | 430 | |
41d214f6 BV |
431 | inc->single_column = g_variant_get_int32(g_hash_table_lookup(options, "single-column")); |
432 | inc->multi_column_mode = inc->single_column == 0; | |
4a35548b | 433 | |
41d214f6 | 434 | inc->num_channels = g_variant_get_int32(g_hash_table_lookup(options, "numchannels")); |
4a35548b | 435 | |
41d214f6 BV |
436 | inc->delimiter = g_string_new(g_variant_get_string( |
437 | g_hash_table_lookup(options, "delimiter"), NULL)); | |
438 | if (inc->delimiter->len == 0) { | |
439 | sr_err("Delimiter must be at least one character."); | |
440 | return SR_ERR_ARG; | |
4a35548b MS |
441 | } |
442 | ||
41d214f6 BV |
443 | s = g_variant_get_string(g_hash_table_lookup(options, "format"), NULL); |
444 | if (!g_ascii_strncasecmp(s, "bin", 3)) { | |
445 | inc->format = FORMAT_BIN; | |
446 | } else if (!g_ascii_strncasecmp(s, "hex", 3)) { | |
447 | inc->format = FORMAT_HEX; | |
448 | } else if (!g_ascii_strncasecmp(s, "oct", 3)) { | |
449 | inc->format = FORMAT_OCT; | |
450 | } else { | |
451 | sr_err("Invalid format: '%s'", s); | |
452 | return SR_ERR_ARG; | |
4a35548b MS |
453 | } |
454 | ||
41d214f6 BV |
455 | inc->comment = g_string_new(g_variant_get_string( |
456 | g_hash_table_lookup(options, "comment"), NULL)); | |
457 | if (g_string_equal(inc->comment, inc->delimiter)) { | |
458 | /* That's never going to work. Likely the result of the user | |
459 | * setting the delimiter to ; -- the default comment. Clearing | |
460 | * the comment setting will work in that case. */ | |
461 | g_string_truncate(inc->comment, 0); | |
4a35548b MS |
462 | } |
463 | ||
6e8d95a5 | 464 | inc->samplerate = g_variant_get_uint64(g_hash_table_lookup(options, "samplerate")); |
4a35548b | 465 | |
41d214f6 | 466 | inc->first_channel = g_variant_get_int32(g_hash_table_lookup(options, "first-channel")); |
4a35548b | 467 | |
41d214f6 | 468 | inc->header = g_variant_get_boolean(g_hash_table_lookup(options, "header")); |
4a35548b | 469 | |
41d214f6 BV |
470 | inc->start_line = g_variant_get_int32(g_hash_table_lookup(options, "startline")); |
471 | if (inc->start_line < 1) { | |
6433156c | 472 | sr_err("Invalid start line %zu.", inc->start_line); |
41d214f6 | 473 | return SR_ERR_ARG; |
4a35548b MS |
474 | } |
475 | ||
41d214f6 BV |
476 | if (inc->multi_column_mode) |
477 | inc->first_column = inc->first_channel; | |
4a35548b | 478 | else |
41d214f6 | 479 | inc->first_column = inc->single_column; |
4a35548b | 480 | |
41d214f6 | 481 | if (!inc->multi_column_mode && !inc->num_channels) { |
ba7dd8bb | 482 | sr_err("Number of channels needs to be specified in single column mode."); |
41d214f6 | 483 | return SR_ERR_ARG; |
4a35548b MS |
484 | } |
485 | ||
41d214f6 BV |
486 | return SR_OK; |
487 | } | |
4a35548b | 488 | |
492dfa90 GS |
489 | static const char *delim_set = "\r\n"; |
490 | ||
329733d9 | 491 | static const char *get_line_termination(GString *buf) |
41d214f6 | 492 | { |
329733d9 | 493 | const char *term; |
4a35548b | 494 | |
41d214f6 BV |
495 | term = NULL; |
496 | if (g_strstr_len(buf->str, buf->len, "\r\n")) | |
497 | term = "\r\n"; | |
498 | else if (memchr(buf->str, '\n', buf->len)) | |
499 | term = "\n"; | |
500 | else if (memchr(buf->str, '\r', buf->len)) | |
501 | term = "\r"; | |
4a35548b | 502 | |
41d214f6 BV |
503 | return term; |
504 | } | |
4a35548b | 505 | |
41d214f6 BV |
506 | static int initial_parse(const struct sr_input *in, GString *buf) |
507 | { | |
508 | struct context *inc; | |
41d214f6 | 509 | GString *channel_name; |
6433156c DE |
510 | unsigned int num_columns, i; |
511 | size_t line_number, l; | |
41d214f6 | 512 | int ret; |
df0db9fd | 513 | char **lines, *line, **columns, *column; |
41d214f6 BV |
514 | |
515 | ret = SR_OK; | |
516 | inc = in->priv; | |
517 | columns = NULL; | |
518 | ||
519 | line_number = 0; | |
492dfa90 | 520 | lines = g_strsplit_set(buf->str, delim_set, 0); |
41d214f6 BV |
521 | for (l = 0; lines[l]; l++) { |
522 | line_number++; | |
df0db9fd | 523 | line = lines[l]; |
41d214f6 BV |
524 | if (inc->start_line > line_number) { |
525 | sr_spew("Line %zu skipped.", line_number); | |
4a35548b MS |
526 | continue; |
527 | } | |
df0db9fd | 528 | if (line[0] == '\0') { |
41d214f6 BV |
529 | sr_spew("Blank line %zu skipped.", line_number); |
530 | continue; | |
531 | } | |
df0db9fd GS |
532 | strip_comment(line, inc->comment); |
533 | if (line[0] == '\0') { | |
41d214f6 | 534 | sr_spew("Comment-only line %zu skipped.", line_number); |
4a35548b MS |
535 | continue; |
536 | } | |
537 | ||
41d214f6 BV |
538 | /* Reached first proper line. */ |
539 | break; | |
540 | } | |
541 | if (!lines[l]) { | |
542 | /* Not enough data for a proper line yet. */ | |
60107497 | 543 | ret = SR_ERR_NA; |
41d214f6 | 544 | goto out; |
4a35548b MS |
545 | } |
546 | ||
547 | /* | |
548 | * In order to determine the number of columns parse the current line | |
549 | * without limiting the number of columns. | |
550 | */ | |
df0db9fd GS |
551 | columns = parse_line(line, inc, -1); |
552 | if (!columns) { | |
41d214f6 BV |
553 | sr_err("Error while parsing line %zu.", line_number); |
554 | ret = SR_ERR; | |
555 | goto out; | |
4a35548b | 556 | } |
4a35548b MS |
557 | num_columns = g_strv_length(columns); |
558 | ||
559 | /* Ensure that the first column is not out of bounds. */ | |
560 | if (!num_columns) { | |
6433156c | 561 | sr_err("Column %u in line %zu is out of bounds.", |
41d214f6 BV |
562 | inc->first_column, line_number); |
563 | ret = SR_ERR; | |
564 | goto out; | |
4a35548b MS |
565 | } |
566 | ||
41d214f6 | 567 | if (inc->multi_column_mode) { |
4a35548b | 568 | /* |
ba7dd8bb | 569 | * Detect the number of channels in multi column mode |
4a35548b MS |
570 | * automatically if not specified. |
571 | */ | |
41d214f6 BV |
572 | if (!inc->num_channels) { |
573 | inc->num_channels = num_columns; | |
6433156c | 574 | sr_dbg("Number of auto-detected channels: %u.", |
41d214f6 | 575 | inc->num_channels); |
4a35548b MS |
576 | } |
577 | ||
578 | /* | |
ba7dd8bb | 579 | * Ensure that the number of channels does not exceed the number |
4a35548b MS |
580 | * of columns in multi column mode. |
581 | */ | |
41d214f6 | 582 | if (num_columns < inc->num_channels) { |
ba7dd8bb | 583 | sr_err("Not enough columns for desired number of channels in line %zu.", |
41d214f6 BV |
584 | line_number); |
585 | ret = SR_ERR; | |
586 | goto out; | |
4a35548b MS |
587 | } |
588 | } | |
589 | ||
41d214f6 BV |
590 | channel_name = g_string_sized_new(64); |
591 | for (i = 0; i < inc->num_channels; i++) { | |
df0db9fd GS |
592 | column = columns[i]; |
593 | if (inc->header && inc->multi_column_mode && column[0] != '\0') | |
594 | g_string_assign(channel_name, column); | |
4a35548b | 595 | else |
6433156c | 596 | g_string_printf(channel_name, "%u", i); |
5e23fcab | 597 | sr_channel_new(in->sdi, i, SR_CHANNEL_LOGIC, TRUE, channel_name->str); |
4a35548b | 598 | } |
41d214f6 | 599 | g_string_free(channel_name, TRUE); |
4a35548b MS |
600 | |
601 | /* | |
602 | * Calculate the minimum buffer size to store the sample data of the | |
ba7dd8bb | 603 | * channels. |
4a35548b | 604 | */ |
41d214f6 BV |
605 | inc->sample_buffer_size = (inc->num_channels + 7) >> 3; |
606 | inc->sample_buffer = g_malloc(inc->sample_buffer_size); | |
4a35548b | 607 | |
41d214f6 BV |
608 | out: |
609 | if (columns) | |
610 | g_strfreev(columns); | |
611 | g_strfreev(lines); | |
4a35548b | 612 | |
41d214f6 | 613 | return ret; |
4a35548b MS |
614 | } |
615 | ||
4439363a GS |
616 | /* |
617 | * Gets called from initial_receive(), which runs until the end-of-line | |
618 | * encoding of the input stream could get determined. Assumes that this | |
619 | * routine receives enough buffered initial input data to either see the | |
620 | * BOM when there is one, or that no BOM will follow when a text line | |
621 | * termination sequence was seen. Silently drops the UTF-8 BOM sequence | |
622 | * from the input buffer if one was seen. Does not care to protect | |
623 | * against multiple execution or dropping the BOM multiple times -- | |
624 | * there should be at most one in the input stream. | |
625 | */ | |
626 | static void initial_bom_check(const struct sr_input *in) | |
627 | { | |
628 | static const char *utf8_bom = "\xef\xbb\xbf"; | |
629 | ||
630 | if (in->buf->len < strlen(utf8_bom)) | |
631 | return; | |
632 | if (strncmp(in->buf->str, utf8_bom, strlen(utf8_bom)) != 0) | |
633 | return; | |
634 | g_string_erase(in->buf, 0, strlen(utf8_bom)); | |
635 | } | |
636 | ||
41d214f6 | 637 | static int initial_receive(const struct sr_input *in) |
4a35548b | 638 | { |
41d214f6 BV |
639 | struct context *inc; |
640 | GString *new_buf; | |
641 | int len, ret; | |
329733d9 UH |
642 | char *p; |
643 | const char *termination; | |
4a35548b | 644 | |
4439363a GS |
645 | initial_bom_check(in); |
646 | ||
41d214f6 | 647 | inc = in->priv; |
4a35548b | 648 | |
df0db9fd GS |
649 | termination = get_line_termination(in->buf); |
650 | if (!termination) | |
41d214f6 | 651 | /* Don't have a full line yet. */ |
d0181813 | 652 | return SR_ERR_NA; |
4a35548b | 653 | |
df0db9fd GS |
654 | p = g_strrstr_len(in->buf->str, in->buf->len, termination); |
655 | if (!p) | |
41d214f6 | 656 | /* Don't have a full line yet. */ |
d0181813 | 657 | return SR_ERR_NA; |
41d214f6 BV |
658 | len = p - in->buf->str - 1; |
659 | new_buf = g_string_new_len(in->buf->str, len); | |
660 | g_string_append_c(new_buf, '\0'); | |
4a35548b | 661 | |
41d214f6 BV |
662 | inc->termination = g_strdup(termination); |
663 | ||
664 | if (in->buf->str[0] != '\0') | |
665 | ret = initial_parse(in, new_buf); | |
666 | else | |
667 | ret = SR_OK; | |
668 | ||
669 | g_string_free(new_buf, TRUE); | |
670 | ||
671 | return ret; | |
672 | } | |
673 | ||
7f4c3a62 | 674 | static int process_buffer(struct sr_input *in, gboolean is_eof) |
41d214f6 BV |
675 | { |
676 | struct sr_datafeed_packet packet; | |
677 | struct sr_datafeed_meta meta; | |
678 | struct sr_config *src; | |
679 | struct context *inc; | |
680 | gsize num_columns; | |
681 | uint64_t samplerate; | |
682 | int max_columns, ret, l; | |
df0db9fd | 683 | char *p, **lines, *line, **columns; |
41d214f6 | 684 | |
41d214f6 | 685 | inc = in->priv; |
d0181813 | 686 | if (!inc->started) { |
bee2b016 | 687 | std_session_send_df_header(in->sdi); |
41d214f6 BV |
688 | |
689 | if (inc->samplerate) { | |
690 | packet.type = SR_DF_META; | |
691 | packet.payload = &meta; | |
692 | samplerate = inc->samplerate; | |
693 | src = sr_config_new(SR_CONF_SAMPLERATE, g_variant_new_uint64(samplerate)); | |
694 | meta.config = g_slist_append(NULL, src); | |
695 | sr_session_send(in->sdi, &packet); | |
c01378c9 | 696 | g_slist_free(meta.config); |
41d214f6 BV |
697 | sr_config_free(src); |
698 | } | |
d0181813 BV |
699 | |
700 | inc->started = TRUE; | |
4a35548b MS |
701 | } |
702 | ||
f9b74861 GS |
703 | /* Limit the number of columns to parse. */ |
704 | if (inc->multi_column_mode) | |
705 | max_columns = inc->num_channels; | |
706 | else | |
707 | max_columns = 1; | |
708 | ||
4555d3bd GS |
709 | /* |
710 | * Consider empty input non-fatal. Keep accumulating input until | |
711 | * at least one full text line has become available. Grab the | |
712 | * maximum amount of accumulated data that consists of full text | |
713 | * lines, and process what has been received so far, leaving not | |
714 | * yet complete lines for the next invocation. | |
7f4c3a62 GS |
715 | * |
716 | * Enforce that all previously buffered data gets processed in | |
717 | * the "EOF" condition. Do not insist in the presence of the | |
718 | * termination sequence for the last line (may often be missing | |
719 | * on Windows). A present termination sequence will just result | |
720 | * in the "execution of an empty line", and does not harm. | |
4555d3bd GS |
721 | */ |
722 | if (!in->buf->len) | |
723 | return SR_OK; | |
7f4c3a62 GS |
724 | if (is_eof) { |
725 | p = in->buf->str + in->buf->len; | |
726 | } else { | |
727 | p = g_strrstr_len(in->buf->str, in->buf->len, inc->termination); | |
728 | if (!p) | |
729 | return SR_ERR; | |
730 | *p = '\0'; | |
731 | p += strlen(inc->termination); | |
732 | } | |
41d214f6 | 733 | g_strstrip(in->buf->str); |
4a35548b | 734 | |
18078d05 | 735 | ret = SR_OK; |
492dfa90 | 736 | lines = g_strsplit_set(in->buf->str, delim_set, 0); |
41d214f6 BV |
737 | for (l = 0; lines[l]; l++) { |
738 | inc->line_number++; | |
df0db9fd GS |
739 | line = lines[l]; |
740 | if (line[0] == '\0') { | |
41d214f6 | 741 | sr_spew("Blank line %zu skipped.", inc->line_number); |
4a35548b MS |
742 | continue; |
743 | } | |
744 | ||
745 | /* Remove trailing comment. */ | |
df0db9fd GS |
746 | strip_comment(line, inc->comment); |
747 | if (line[0] == '\0') { | |
41d214f6 | 748 | sr_spew("Comment-only line %zu skipped.", inc->line_number); |
4a35548b MS |
749 | continue; |
750 | } | |
751 | ||
160691b9 JS |
752 | /* Skip the header line, its content was used as the channel names. */ |
753 | if (inc->header) { | |
754 | sr_spew("Header line %zu skipped.", inc->line_number); | |
755 | inc->header = FALSE; | |
756 | continue; | |
757 | } | |
758 | ||
df0db9fd GS |
759 | columns = parse_line(line, inc, max_columns); |
760 | if (!columns) { | |
41d214f6 | 761 | sr_err("Error while parsing line %zu.", inc->line_number); |
4a35548b MS |
762 | return SR_ERR; |
763 | } | |
4a35548b | 764 | num_columns = g_strv_length(columns); |
4a35548b | 765 | if (!num_columns) { |
6433156c | 766 | sr_err("Column %u in line %zu is out of bounds.", |
41d214f6 | 767 | inc->first_column, inc->line_number); |
4a35548b | 768 | g_strfreev(columns); |
4a35548b MS |
769 | return SR_ERR; |
770 | } | |
4a35548b | 771 | /* |
ba7dd8bb | 772 | * Ensure that the number of channels does not exceed the number |
4a35548b MS |
773 | * of columns in multi column mode. |
774 | */ | |
41d214f6 | 775 | if (inc->multi_column_mode && num_columns < inc->num_channels) { |
ba7dd8bb | 776 | sr_err("Not enough columns for desired number of channels in line %zu.", |
41d214f6 | 777 | inc->line_number); |
4a35548b | 778 | g_strfreev(columns); |
4a35548b MS |
779 | return SR_ERR; |
780 | } | |
781 | ||
41d214f6 BV |
782 | if (inc->multi_column_mode) |
783 | ret = parse_multi_columns(columns, inc); | |
4a35548b | 784 | else |
41d214f6 BV |
785 | ret = parse_single_column(columns[0], inc); |
786 | if (ret != SR_OK) { | |
4a35548b | 787 | g_strfreev(columns); |
4a35548b MS |
788 | return SR_ERR; |
789 | } | |
790 | ||
4a35548b | 791 | /* Send sample data to the session bus. */ |
41d214f6 BV |
792 | ret = send_samples(in->sdi, inc->sample_buffer, |
793 | inc->sample_buffer_size, 1); | |
794 | if (ret != SR_OK) { | |
4a35548b | 795 | sr_err("Sending samples failed."); |
4a35548b MS |
796 | return SR_ERR; |
797 | } | |
41d214f6 BV |
798 | g_strfreev(columns); |
799 | } | |
800 | g_strfreev(lines); | |
241c386a | 801 | g_string_erase(in->buf, 0, p - in->buf->str); |
41d214f6 | 802 | |
7066fd46 | 803 | return ret; |
41d214f6 BV |
804 | } |
805 | ||
7066fd46 | 806 | static int receive(struct sr_input *in, GString *buf) |
41d214f6 BV |
807 | { |
808 | struct context *inc; | |
7066fd46 BV |
809 | int ret; |
810 | ||
811 | g_string_append_len(in->buf, buf->str, buf->len); | |
41d214f6 BV |
812 | |
813 | inc = in->priv; | |
7066fd46 | 814 | if (!inc->termination) { |
df0db9fd GS |
815 | ret = initial_receive(in); |
816 | if (ret == SR_ERR_NA) | |
7066fd46 BV |
817 | /* Not enough data yet. */ |
818 | return SR_OK; | |
819 | else if (ret != SR_OK) | |
820 | return SR_ERR; | |
821 | ||
822 | /* sdi is ready, notify frontend. */ | |
823 | in->sdi_ready = TRUE; | |
41d214f6 | 824 | return SR_OK; |
7066fd46 BV |
825 | } |
826 | ||
7f4c3a62 | 827 | ret = process_buffer(in, FALSE); |
7066fd46 BV |
828 | |
829 | return ret; | |
830 | } | |
831 | ||
832 | static int end(struct sr_input *in) | |
833 | { | |
834 | struct context *inc; | |
7066fd46 | 835 | int ret; |
41d214f6 | 836 | |
7066fd46 | 837 | if (in->sdi_ready) |
7f4c3a62 | 838 | ret = process_buffer(in, TRUE); |
7066fd46 BV |
839 | else |
840 | ret = SR_OK; | |
841 | ||
842 | inc = in->priv; | |
3be42bc2 | 843 | if (inc->started) |
bee2b016 | 844 | std_session_send_df_end(in->sdi); |
4a35548b | 845 | |
7066fd46 BV |
846 | return ret; |
847 | } | |
848 | ||
d5cc282f | 849 | static void cleanup(struct sr_input *in) |
7066fd46 BV |
850 | { |
851 | struct context *inc; | |
852 | ||
853 | inc = in->priv; | |
854 | ||
41d214f6 BV |
855 | if (inc->delimiter) |
856 | g_string_free(inc->delimiter, TRUE); | |
857 | ||
858 | if (inc->comment) | |
859 | g_string_free(inc->comment, TRUE); | |
4a35548b | 860 | |
b1f83103 UH |
861 | g_free(inc->termination); |
862 | g_free(inc->sample_buffer); | |
4a35548b MS |
863 | } |
864 | ||
ad93bfb0 SA |
865 | static int reset(struct sr_input *in) |
866 | { | |
867 | struct context *inc = in->priv; | |
868 | ||
869 | cleanup(in); | |
870 | inc->started = FALSE; | |
871 | g_string_truncate(in->buf, 0); | |
872 | ||
873 | return SR_OK; | |
874 | } | |
875 | ||
41d214f6 BV |
876 | static struct sr_option options[] = { |
877 | { "single-column", "Single column", "Enable/specify single column", NULL, NULL }, | |
878 | { "numchannels", "Max channels", "Number of channels", NULL, NULL }, | |
879 | { "delimiter", "Delimiter", "Column delimiter", NULL, NULL }, | |
880 | { "format", "Format", "Numeric format", NULL, NULL }, | |
881 | { "comment", "Comment", "Comment prefix character", NULL, NULL }, | |
882 | { "samplerate", "Samplerate", "Samplerate used during capture", NULL, NULL }, | |
883 | { "first-channel", "First channel", "Column number of first channel", NULL, NULL }, | |
884 | { "header", "Header", "Treat first line as header with channel names", NULL, NULL }, | |
885 | { "startline", "Start line", "Line number at which to start processing samples", NULL, NULL }, | |
06ad20be | 886 | ALL_ZERO |
41d214f6 BV |
887 | }; |
888 | ||
2c240774 | 889 | static const struct sr_option *get_options(void) |
41d214f6 BV |
890 | { |
891 | if (!options[0].def) { | |
892 | options[0].def = g_variant_ref_sink(g_variant_new_int32(0)); | |
893 | options[1].def = g_variant_ref_sink(g_variant_new_int32(0)); | |
894 | options[2].def = g_variant_ref_sink(g_variant_new_string(",")); | |
895 | options[3].def = g_variant_ref_sink(g_variant_new_string("bin")); | |
896 | options[4].def = g_variant_ref_sink(g_variant_new_string(";")); | |
6e8d95a5 | 897 | options[5].def = g_variant_ref_sink(g_variant_new_uint64(0)); |
41d214f6 BV |
898 | options[6].def = g_variant_ref_sink(g_variant_new_int32(0)); |
899 | options[7].def = g_variant_ref_sink(g_variant_new_boolean(FALSE)); | |
900 | options[8].def = g_variant_ref_sink(g_variant_new_int32(1)); | |
901 | } | |
902 | ||
903 | return options; | |
904 | } | |
905 | ||
d4c93774 | 906 | SR_PRIV struct sr_input_module input_csv = { |
4a35548b | 907 | .id = "csv", |
41d214f6 BV |
908 | .name = "CSV", |
909 | .desc = "Comma-separated values", | |
c7bc82ff | 910 | .exts = (const char*[]){"csv", NULL}, |
41d214f6 | 911 | .options = get_options, |
4a35548b | 912 | .init = init, |
41d214f6 | 913 | .receive = receive, |
7066fd46 | 914 | .end = end, |
41d214f6 | 915 | .cleanup = cleanup, |
ad93bfb0 | 916 | .reset = reset, |
4a35548b | 917 | }; |