]>
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> | |
e53f32d2 | 5 | * Copyright (C) 2019 Gerhard Sittig <gerhard.sittig@gmx.net> |
4a35548b MS |
6 | * |
7 | * This program is free software: you can redistribute it and/or modify | |
8 | * it under the terms of the GNU General Public License as published by | |
9 | * the Free Software Foundation, either version 3 of the License, or | |
10 | * (at your option) any later version. | |
11 | * | |
12 | * This program is distributed in the hope that it will be useful, | |
13 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
14 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
15 | * GNU General Public License for more details. | |
16 | * | |
17 | * You should have received a copy of the GNU General Public License | |
18 | * along with this program. If not, see <http://www.gnu.org/licenses/>. | |
19 | */ | |
20 | ||
e05f1827 GS |
21 | #include "config.h" |
22 | ||
23 | #include <glib.h> | |
4a35548b MS |
24 | #include <stdlib.h> |
25 | #include <string.h> | |
e05f1827 | 26 | |
c1aae900 | 27 | #include <libsigrok/libsigrok.h> |
4a35548b | 28 | #include "libsigrok-internal.h" |
f6dcb320 | 29 | #include "scpi.h" /* String un-quote for channel name from header line. */ |
4a35548b | 30 | |
3544f848 | 31 | #define LOG_PREFIX "input/csv" |
4a35548b | 32 | |
9a4fd01a | 33 | #define CHUNK_SIZE (4 * 1024 * 1024) |
cd59e6ec | 34 | |
4a35548b MS |
35 | /* |
36 | * The CSV input module has the following options: | |
37 | * | |
72903e9d GS |
38 | * column_formats: Specifies the data formats and channel counts for the |
39 | * input file's text columns. Accepts a comma separated list of tuples | |
40 | * with: an optional column repeat count ('*' as a wildcard meaning | |
41 | * "all remaining columns", only applicable to the last field), a format | |
42 | * specifying character ('x' hexadecimal, 'o' octal, 'b' binary, 'l' | |
43 | * single-bit logic), and an optional bit count (translating to: logic | |
44 | * channels communicated in that column). This "column_formats" option | |
45 | * is most versatile, other forms of specifying the column layout only | |
46 | * exist for backwards compatibility. | |
4a35548b | 47 | * |
72903e9d GS |
48 | * single_column: Specifies the column number which contains the logic data |
49 | * for single-column mode. All logic data is taken from several bits | |
50 | * which all are kept within that one column. Only exists for backwards | |
51 | * compatibility, see "column_formats" for more flexibility. | |
4a35548b | 52 | * |
72903e9d GS |
53 | * first_column: Specifies the number of the first column with logic data |
54 | * in simple multi-column mode. Only exists for backwards compatibility, | |
55 | * see "column_formats" for more flexibility. | |
4a35548b | 56 | * |
72903e9d GS |
57 | * logic_channels: Specifies the number of logic channels. Is required in |
58 | * simple single-column mode. Is optional in simple multi-column mode | |
59 | * (and defaults to all remaining columns). Only exists for backwards | |
60 | * compatibility, see "column_formats" for more flexibility. | |
4a35548b | 61 | * |
72903e9d GS |
62 | * single_format: Specifies the format of the input text in simple single- |
63 | * column mode. Available formats are: 'bin' (default), 'hex' and 'oct'. | |
64 | * Simple multi-column mode always uses single-bit data per column. | |
65 | * Only exists for backwards compatibility, see "column_formats" for | |
66 | * more flexibility. | |
4a35548b | 67 | * |
72903e9d GS |
68 | * start_line: Specifies at which line to start processing the input file. |
69 | * Allows to skip leading lines which neither are header nor data lines. | |
70 | * By default all of the input file gets processed. | |
4a35548b | 71 | * |
72903e9d GS |
72 | * header: Boolean option, controls whether the first processed line is used |
73 | * to determine channel names. Off by default. Generic channel names are | |
74 | * used in the absence of header line content. | |
4a35548b | 75 | * |
72903e9d GS |
76 | * samplerate: Specifies the samplerate of the input data. Defaults to 0. |
77 | * User specs take precedence over data which optionally gets derived | |
78 | * from input data. | |
4a35548b | 79 | * |
72903e9d GS |
80 | * column_separator: Specifies the sequence which separates the text file |
81 | * columns. Cannot be empty. Defaults to comma. | |
82 | * | |
83 | * comment_leader: Specifies the sequence which starts comments that run | |
84 | * up to the end of the current text line. Can be empty to disable | |
85 | * comment support. Defaults to semicolon. | |
86 | * | |
87 | * Typical examples of using these options: | |
88 | * - ... -I csv:column_formats=*l ... | |
89 | * All columns are single-bit logic data. Identical to the previous | |
90 | * multi-column mode (the default when no options were given at all). | |
91 | * - ... -I csv:column_formats=3-,*l ... | |
92 | * Ignore the first three columns, get single-bit logic data from all | |
93 | * remaining lines (multi-column mode with first-column above 1). | |
94 | * - ... -I csv:column_formats=3-,4l,x8 ... | |
95 | * Ignore the first three columns, get single-bit logic data from the | |
96 | * next four columns, then eight-bit data in hex format from the next | |
97 | * column. More columns may follow in the input text but won't get | |
98 | * processed. (Mix of previous multi-column as well as single-column | |
99 | * modes.) | |
100 | * - ... -I csv:column_formats=4x8,b16,5l ... | |
101 | * Get eight-bit data in hex format from the first four columns, then | |
102 | * sixteen-bit data in binary format, then five times single-bit data. | |
103 | * - ... -I csv:single_column=2:single_format=bin:logic_channels=8 ... | |
104 | * Get eight logic bits in binary format from column 2. (Simple | |
105 | * single-column mode, corresponds to the "-,b8" format.) | |
106 | * - ... -I csv:first_column=6:logic_channels=4 ... | |
107 | * Get four single-bit logic channels from columns 6 to 9 respectively. | |
108 | * (Simple multi-column mode, corresponds to the "5-,4b" format.) | |
109 | * - ... -I csv:start_line=20:header=yes:... | |
110 | * Skip the first 19 text lines. Use line 20 to derive channel names. | |
111 | * Data starts at line 21. | |
4a35548b MS |
112 | */ |
113 | ||
ccff468b GS |
114 | /* |
115 | * TODO | |
116 | * | |
ccff468b | 117 | * - Add support for analog input data? (optional) |
5a971176 GS |
118 | * - Extend the set of supported column types. Just grab a double |
119 | * value from floating point format input text. | |
120 | * - Optionally get precision ('digits') from the column's format spec? | |
121 | * From the position which is "bit count" for logic channels? | |
122 | * - Optionally get sample rate from timestamp column. Just best-effort | |
123 | * approach, not necessarily reliable. Users can always specify rates. | |
124 | * - Add a test suite for input modules in general, and CSV in specific? | |
125 | * Becomes more important with the multitude of options and their | |
126 | * interaction. Could cover edge cases (BOM presence, line termination | |
127 | * absence, etc) and auto-stuff as well (channel names, channel counts, | |
128 | * samplerates, etc). | |
ccff468b GS |
129 | */ |
130 | ||
4a35548b | 131 | /* Single column formats. */ |
ad6a2bee | 132 | enum single_col_format { |
e53f32d2 GS |
133 | FORMAT_NONE, /* Ignore this column. */ |
134 | FORMAT_BIN, /* Bin digits for a set of bits (or just one bit). */ | |
135 | FORMAT_HEX, /* Hex digits for a set of bits. */ | |
136 | FORMAT_OCT, /* Oct digits for a set of bits. */ | |
137 | }; | |
138 | ||
139 | static const char *col_format_text[] = { | |
140 | [FORMAT_NONE] = "unknown", | |
141 | [FORMAT_BIN] = "binary", | |
142 | [FORMAT_HEX] = "hexadecimal", | |
143 | [FORMAT_OCT] = "octal", | |
144 | }; | |
145 | ||
1a920e33 GS |
146 | static const char col_format_char[] = { |
147 | [FORMAT_NONE] = '?', | |
148 | [FORMAT_BIN] = 'b', | |
149 | [FORMAT_HEX] = 'x', | |
150 | [FORMAT_OCT] = 'o', | |
151 | }; | |
152 | ||
e53f32d2 GS |
153 | struct column_details { |
154 | size_t col_nr; | |
155 | enum single_col_format text_format; | |
156 | size_t channel_offset; | |
157 | size_t channel_count; | |
4a35548b MS |
158 | }; |
159 | ||
160 | struct context { | |
41d214f6 BV |
161 | gboolean started; |
162 | ||
4a35548b MS |
163 | /* Current selected samplerate. */ |
164 | uint64_t samplerate; | |
246aca5f | 165 | gboolean samplerate_sent; |
4a35548b | 166 | |
836fac9c GS |
167 | /* Number of logic channels. */ |
168 | size_t logic_channels; | |
4a35548b | 169 | |
836fac9c | 170 | /* Column delimiter (actually separator), comment leader, EOL sequence. */ |
4a35548b | 171 | GString *delimiter; |
4a35548b | 172 | GString *comment; |
41d214f6 BV |
173 | char *termination; |
174 | ||
1a920e33 GS |
175 | /* Format specs for input columns, and processing state. */ |
176 | size_t column_seen_count; | |
177 | const char *column_formats; | |
e53f32d2 GS |
178 | size_t column_want_count; |
179 | struct column_details *column_details; | |
180 | ||
4a35548b | 181 | /* Line number to start processing. */ |
6433156c | 182 | size_t start_line; |
4a35548b MS |
183 | |
184 | /* | |
185 | * Determines if the first line should be treated as header and used for | |
ba7dd8bb | 186 | * channel names in multi column mode. |
4a35548b | 187 | */ |
de8fe3b5 GS |
188 | gboolean use_header; |
189 | gboolean header_seen; | |
4a35548b | 190 | |
cd59e6ec GS |
191 | size_t sample_unit_size; /**!< Byte count for a single sample. */ |
192 | uint8_t *sample_buffer; /**!< Buffer for a single sample. */ | |
4a35548b | 193 | |
cd59e6ec GS |
194 | uint8_t *datafeed_buffer; /**!< Queue for datafeed submission. */ |
195 | size_t datafeed_buf_size; | |
196 | size_t datafeed_buf_fill; | |
4a35548b | 197 | |
4a35548b | 198 | /* Current line number. */ |
6433156c | 199 | size_t line_number; |
affaf540 GS |
200 | |
201 | /* List of previously created sigrok channels. */ | |
202 | GSList *prev_sr_channels; | |
4a35548b MS |
203 | }; |
204 | ||
626c388a GS |
205 | /* |
206 | * Primitive operations to handle sample sets: | |
207 | * - Keep a buffer for datafeed submission, capable of holding many | |
208 | * samples (reduces call overhead, improves throughput). | |
209 | * - Have a "current sample set" pointer reference one position in that | |
210 | * large samples buffer. | |
211 | * - Clear the current sample set before text line inspection, then set | |
212 | * the bits which are found active in the current line of text input. | |
213 | * Phrase the API such that call sites can be kept simple. Advance to | |
214 | * the next sample set between lines, flush the larger buffer as needed | |
215 | * (when it is full, or upon EOF). | |
216 | */ | |
217 | ||
218 | static void clear_logic_samples(struct context *inc) | |
219 | { | |
220 | inc->sample_buffer = &inc->datafeed_buffer[inc->datafeed_buf_fill]; | |
221 | memset(inc->sample_buffer, 0, inc->sample_unit_size); | |
222 | } | |
223 | ||
224 | static void set_logic_level(struct context *inc, size_t ch_idx, int on) | |
225 | { | |
226 | size_t byte_idx, bit_idx; | |
227 | uint8_t bit_mask; | |
228 | ||
836fac9c | 229 | if (ch_idx >= inc->logic_channels) |
626c388a GS |
230 | return; |
231 | if (!on) | |
232 | return; | |
233 | ||
234 | byte_idx = ch_idx / 8; | |
235 | bit_idx = ch_idx % 8; | |
236 | bit_mask = 1 << bit_idx; | |
237 | inc->sample_buffer[byte_idx] |= bit_mask; | |
238 | } | |
239 | ||
240 | static int flush_logic_samples(const struct sr_input *in) | |
241 | { | |
242 | struct context *inc; | |
243 | struct sr_datafeed_packet packet; | |
246aca5f GS |
244 | struct sr_datafeed_meta meta; |
245 | struct sr_config *src; | |
626c388a GS |
246 | struct sr_datafeed_logic logic; |
247 | int rc; | |
248 | ||
249 | inc = in->priv; | |
250 | if (!inc->datafeed_buf_fill) | |
251 | return SR_OK; | |
252 | ||
246aca5f GS |
253 | if (inc->samplerate && !inc->samplerate_sent) { |
254 | packet.type = SR_DF_META; | |
255 | packet.payload = &meta; | |
5ada72fc | 256 | src = sr_config_new(SR_CONF_SAMPLERATE, g_variant_new_uint64(inc->samplerate)); |
246aca5f GS |
257 | meta.config = g_slist_append(NULL, src); |
258 | sr_session_send(in->sdi, &packet); | |
259 | g_slist_free(meta.config); | |
260 | sr_config_free(src); | |
261 | inc->samplerate_sent = TRUE; | |
262 | } | |
263 | ||
626c388a GS |
264 | memset(&packet, 0, sizeof(packet)); |
265 | memset(&logic, 0, sizeof(logic)); | |
266 | packet.type = SR_DF_LOGIC; | |
267 | packet.payload = &logic; | |
268 | logic.unitsize = inc->sample_unit_size; | |
269 | logic.length = inc->datafeed_buf_fill; | |
270 | logic.data = inc->datafeed_buffer; | |
271 | ||
272 | rc = sr_session_send(in->sdi, &packet); | |
273 | if (rc != SR_OK) | |
274 | return rc; | |
275 | ||
276 | inc->datafeed_buf_fill = 0; | |
277 | return SR_OK; | |
278 | } | |
279 | ||
280 | static int queue_logic_samples(const struct sr_input *in) | |
281 | { | |
282 | struct context *inc; | |
283 | int rc; | |
284 | ||
285 | inc = in->priv; | |
836fac9c GS |
286 | if (!inc->logic_channels) |
287 | return SR_OK; | |
626c388a GS |
288 | |
289 | inc->datafeed_buf_fill += inc->sample_unit_size; | |
290 | if (inc->datafeed_buf_fill == inc->datafeed_buf_size) { | |
291 | rc = flush_logic_samples(in); | |
292 | if (rc != SR_OK) | |
293 | return rc; | |
294 | } | |
295 | return SR_OK; | |
296 | } | |
297 | ||
2142a79b GS |
298 | /* Helpers for "column processing". */ |
299 | ||
300 | static int split_column_format(const char *spec, | |
301 | size_t *column_count, enum single_col_format *format, size_t *bit_count) | |
302 | { | |
303 | size_t count; | |
304 | char *endp, format_char; | |
305 | enum single_col_format format_code; | |
306 | ||
307 | if (!spec || !*spec) | |
308 | return SR_ERR_ARG; | |
309 | ||
1a920e33 | 310 | /* Get the (optional, decimal, default 1) column count. Accept '*'. */ |
2142a79b | 311 | endp = NULL; |
1a920e33 | 312 | if (*spec == '*') { |
5ada72fc | 313 | /* Workaround, strtoul("*") won't always yield expected endp. */ |
1a920e33 GS |
314 | count = 0; |
315 | endp = (char *)&spec[1]; | |
316 | } else { | |
317 | count = strtoul(spec, &endp, 10); | |
318 | } | |
2142a79b GS |
319 | if (!endp) |
320 | return SR_ERR_ARG; | |
321 | if (endp == spec) | |
322 | count = 1; | |
323 | if (column_count) | |
324 | *column_count = count; | |
325 | spec = endp; | |
326 | ||
327 | /* Get the (mandatory, single letter) type spec (-/xob/l). */ | |
328 | format_char = *spec++; | |
329 | switch (format_char) { | |
5ada72fc | 330 | case '-': |
2142a79b GS |
331 | case '/': |
332 | format_char = '-'; | |
333 | format_code = FORMAT_NONE; | |
334 | break; | |
335 | case 'x': | |
336 | format_code = FORMAT_HEX; | |
337 | break; | |
338 | case 'o': | |
339 | format_code = FORMAT_OCT; | |
340 | break; | |
341 | case 'b': | |
342 | case 'l': | |
343 | format_code = FORMAT_BIN; | |
344 | break; | |
345 | default: /* includes NUL */ | |
346 | return SR_ERR_ARG; | |
347 | } | |
348 | if (format) | |
349 | *format = format_code; | |
350 | ||
351 | /* Get the (optional, decimal, default 1) bit count. */ | |
352 | endp = NULL; | |
353 | count = strtoul(spec, &endp, 10); | |
354 | if (!endp) | |
355 | return SR_ERR_ARG; | |
356 | if (endp == spec) | |
357 | count = 1; | |
358 | if (format_char == '-') | |
359 | count = 0; | |
360 | if (format_char == 'l') | |
361 | count = 1; | |
362 | if (bit_count) | |
363 | *bit_count = count; | |
364 | spec = endp; | |
365 | ||
366 | /* Input spec must have been exhausted. */ | |
367 | if (*spec) | |
368 | return SR_ERR_ARG; | |
369 | ||
370 | return SR_OK; | |
371 | } | |
372 | ||
9e7af34e GS |
373 | static int make_column_details_from_format(const struct sr_input *in, |
374 | const char *column_format, char **column_texts) | |
2142a79b | 375 | { |
9e7af34e | 376 | struct context *inc; |
2142a79b GS |
377 | char **formats, *format; |
378 | size_t format_count, column_count, bit_count; | |
1a920e33 | 379 | size_t auto_column_count; |
2142a79b GS |
380 | size_t format_idx, c, b, column_idx, channel_idx; |
381 | enum single_col_format f; | |
382 | struct column_details *detail; | |
9e7af34e GS |
383 | GString *channel_name; |
384 | size_t create_idx; | |
385 | char *column; | |
386 | const char *caption; | |
2142a79b GS |
387 | int ret; |
388 | ||
9e7af34e GS |
389 | inc = in->priv; |
390 | inc->column_seen_count = g_strv_length(column_texts); | |
391 | ||
2142a79b GS |
392 | /* Split the input spec, count involved columns and bits. */ |
393 | formats = g_strsplit(column_format, ",", 0); | |
394 | if (!formats) { | |
395 | sr_err("Cannot parse columns format %s (comma split).", column_format); | |
396 | return SR_ERR_ARG; | |
397 | } | |
398 | format_count = g_strv_length(formats); | |
399 | if (!format_count) { | |
400 | sr_err("Cannot parse columns format %s (field count).", column_format); | |
401 | g_strfreev(formats); | |
402 | return SR_ERR_ARG; | |
403 | } | |
404 | column_count = bit_count = 0; | |
1a920e33 | 405 | auto_column_count = 0; |
2142a79b GS |
406 | for (format_idx = 0; format_idx < format_count; format_idx++) { |
407 | format = formats[format_idx]; | |
408 | ret = split_column_format(format, &c, &f, &b); | |
409 | sr_dbg("fmt %s -> %zu cols, %s fmt, %zu bits, rc %d", format, c, col_format_text[f], b, ret); | |
410 | if (ret != SR_OK) { | |
411 | sr_err("Cannot parse columns format %s (field split, %s).", column_format, format); | |
412 | g_strfreev(formats); | |
413 | return SR_ERR_ARG; | |
414 | } | |
1a920e33 GS |
415 | if (f && !c) { |
416 | /* User requested "auto-count", must be last format. */ | |
417 | if (formats[format_idx + 1]) { | |
418 | sr_err("Auto column count must be last format field."); | |
419 | g_strfreev(formats); | |
420 | return SR_ERR_ARG; | |
421 | } | |
422 | auto_column_count = inc->column_seen_count - column_count; | |
423 | c = auto_column_count; | |
424 | } | |
2142a79b GS |
425 | column_count += c; |
426 | bit_count += c * b; | |
427 | } | |
428 | sr_dbg("Column format %s -> %zu columns, %zu logic channels.", | |
429 | column_format, column_count, bit_count); | |
430 | ||
9e7af34e | 431 | /* Allocate and fill in "column processing" details. Create channels. */ |
2142a79b | 432 | inc->column_want_count = column_count; |
9e7af34e GS |
433 | if (inc->column_seen_count < inc->column_want_count) { |
434 | sr_err("Insufficient input text width for desired data amount, got %zu but want %zu columns.", | |
435 | inc->column_seen_count, inc->column_want_count); | |
436 | g_strfreev(formats); | |
437 | return SR_ERR_ARG; | |
438 | } | |
2142a79b GS |
439 | inc->column_details = g_malloc0_n(column_count, sizeof(inc->column_details[0])); |
440 | column_idx = channel_idx = 0; | |
9e7af34e | 441 | channel_name = g_string_sized_new(64); |
2142a79b | 442 | for (format_idx = 0; format_idx < format_count; format_idx++) { |
9e7af34e | 443 | /* Process a format field, which can span multiple columns. */ |
2142a79b GS |
444 | format = formats[format_idx]; |
445 | (void)split_column_format(format, &c, &f, &b); | |
1a920e33 GS |
446 | if (f && !c) |
447 | c = auto_column_count; | |
2142a79b | 448 | while (c-- > 0) { |
9e7af34e | 449 | /* Fill in a column's processing details. */ |
2142a79b GS |
450 | detail = &inc->column_details[column_idx++]; |
451 | detail->col_nr = column_idx; | |
452 | detail->text_format = f; | |
453 | if (detail->text_format) { | |
454 | detail->channel_offset = channel_idx; | |
455 | detail->channel_count = b; | |
456 | channel_idx += b; | |
457 | } | |
458 | sr_dbg("detail -> col %zu, fmt %s, ch off/cnt %zu/%zu", | |
459 | detail->col_nr, col_format_text[detail->text_format], | |
460 | detail->channel_offset, detail->channel_count); | |
9e7af34e GS |
461 | if (!detail->text_format) |
462 | continue; | |
463 | /* | |
464 | * Create channels with appropriate names. Optionally | |
465 | * use text from a header line (when requested by the | |
466 | * user). In the absence of header text, channels are | |
467 | * assigned rather generic names. | |
468 | * | |
469 | * Manipulation of the column's caption (when a header | |
470 | * line is seen) is acceptable, because this header | |
471 | * line won't get processed another time. | |
472 | */ | |
473 | column = column_texts[detail->col_nr - 1]; | |
474 | if (inc->use_header && column && *column) | |
475 | caption = sr_scpi_unquote_string(column); | |
476 | else | |
477 | caption = NULL; | |
478 | if (!caption || !*caption) | |
479 | caption = NULL; | |
480 | for (create_idx = 0; create_idx < detail->channel_count; create_idx++) { | |
481 | if (caption && detail->channel_count == 1) { | |
482 | g_string_assign(channel_name, caption); | |
483 | } else if (caption) { | |
484 | g_string_printf(channel_name, "%s[%zu]", | |
485 | caption, create_idx); | |
486 | } else { | |
487 | g_string_printf(channel_name, "%zu", | |
488 | detail->channel_offset + create_idx); | |
489 | } | |
490 | sr_channel_new(in->sdi, detail->channel_offset + create_idx, | |
491 | SR_CHANNEL_LOGIC, TRUE, channel_name->str); | |
492 | } | |
2142a79b GS |
493 | } |
494 | } | |
495 | inc->logic_channels = channel_idx; | |
9e7af34e | 496 | g_string_free(channel_name, TRUE); |
2142a79b GS |
497 | g_strfreev(formats); |
498 | ||
499 | return SR_OK; | |
500 | } | |
501 | ||
e53f32d2 GS |
502 | static const struct column_details *lookup_column_details(struct context *inc, size_t nr) |
503 | { | |
504 | if (!inc || !inc->column_details) | |
505 | return NULL; | |
506 | if (!nr || nr > inc->column_want_count) | |
507 | return NULL; | |
508 | return &inc->column_details[nr - 1]; | |
509 | } | |
510 | ||
19267272 GS |
511 | /* |
512 | * Primitive operations for text input: Strip comments off text lines. | |
513 | * Split text lines into columns. Process input text for individual | |
514 | * columns. | |
515 | */ | |
516 | ||
41d214f6 | 517 | static void strip_comment(char *buf, const GString *prefix) |
4a35548b MS |
518 | { |
519 | char *ptr; | |
520 | ||
521 | if (!prefix->len) | |
522 | return; | |
523 | ||
b2c4dde2 | 524 | if ((ptr = strstr(buf, prefix->str))) { |
41d214f6 | 525 | *ptr = '\0'; |
b2c4dde2 GS |
526 | g_strstrip(buf); |
527 | } | |
4a35548b MS |
528 | } |
529 | ||
19267272 | 530 | /** |
e53f32d2 | 531 | * @brief Splits a text line into a set of columns. |
19267272 | 532 | * |
e53f32d2 | 533 | * @param[in] buf The input text line to split. |
19267272 GS |
534 | * @param[in] inc The input module's context. |
535 | * | |
e53f32d2 | 536 | * @returns An array of strings, representing the columns' text. |
19267272 | 537 | * |
e53f32d2 | 538 | * This routine splits a text line on previously determined separators. |
19267272 | 539 | */ |
e53f32d2 | 540 | static char **split_line(char *buf, struct context *inc) |
4a35548b | 541 | { |
e53f32d2 | 542 | return g_strsplit(buf, inc->delimiter->str, 0); |
4a35548b MS |
543 | } |
544 | ||
19267272 | 545 | /** |
e53f32d2 | 546 | * @brief Parse a multi-bit field into several logic channels. |
19267272 | 547 | * |
e53f32d2 | 548 | * @param[in] column The input text, a run of bin/hex/oct digits. |
19267272 | 549 | * @param[in] inc The input module's context. |
836fac9c | 550 | * @param[in] details The column processing details. |
19267272 GS |
551 | * |
552 | * @retval SR_OK Success. | |
553 | * @retval SR_ERR Invalid input data (empty, or format error). | |
554 | * | |
555 | * This routine modifies the logic levels in the current sample set, | |
e53f32d2 | 556 | * based on the text input and a user provided format spec. |
19267272 | 557 | */ |
836fac9c GS |
558 | static int parse_logic(const char *column, struct context *inc, |
559 | const struct column_details *details) | |
4a35548b | 560 | { |
e53f32d2 GS |
561 | size_t length, ch_rem, ch_idx, ch_inc; |
562 | const char *rdptr; | |
4a35548b | 563 | char c; |
e53f32d2 GS |
564 | gboolean valid; |
565 | const char *type_text; | |
566 | uint8_t bits; | |
567 | ||
e53f32d2 GS |
568 | /* |
569 | * Prepare to read the digits from the text end towards the start. | |
570 | * A digit corresponds to a variable number of channels (depending | |
571 | * on the value's radix). Prepare the mapping of text digits to | |
572 | * (a number of) logic channels. | |
573 | */ | |
574 | length = strlen(column); | |
4a35548b | 575 | if (!length) { |
836fac9c | 576 | sr_err("Column %zu in line %zu is empty.", details->col_nr, |
41d214f6 | 577 | inc->line_number); |
4a35548b MS |
578 | return SR_ERR; |
579 | } | |
e53f32d2 | 580 | rdptr = &column[length]; |
836fac9c GS |
581 | ch_idx = details->channel_offset; |
582 | ch_rem = details->channel_count; | |
4a35548b | 583 | |
e53f32d2 GS |
584 | /* |
585 | * Get another digit and derive up to four logic channels' state from | |
586 | * it. Make sure to not process more bits than the column has channels | |
587 | * associated with it. | |
588 | */ | |
589 | while (rdptr > column && ch_rem) { | |
590 | /* Check for valid digits according to the input radix. */ | |
591 | c = *(--rdptr); | |
836fac9c | 592 | switch (details->text_format) { |
e53f32d2 GS |
593 | case FORMAT_BIN: |
594 | valid = g_ascii_isxdigit(c) && c < '2'; | |
595 | ch_inc = 1; | |
596 | break; | |
597 | case FORMAT_OCT: | |
598 | valid = g_ascii_isxdigit(c) && c < '8'; | |
599 | ch_inc = 3; | |
600 | break; | |
601 | case FORMAT_HEX: | |
602 | valid = g_ascii_isxdigit(c); | |
603 | ch_inc = 4; | |
604 | break; | |
605 | default: | |
606 | valid = FALSE; | |
607 | break; | |
4a35548b | 608 | } |
e53f32d2 | 609 | if (!valid) { |
836fac9c | 610 | type_text = col_format_text[details->text_format]; |
e53f32d2 | 611 | sr_err("Invalid text '%s' in %s type column %zu in line %zu.", |
836fac9c | 612 | column, type_text, details->col_nr, inc->line_number); |
4a35548b | 613 | return SR_ERR; |
e53f32d2 GS |
614 | } |
615 | /* Use the digit's bits for logic channels' data. */ | |
616 | bits = g_ascii_xdigit_value(c); | |
836fac9c | 617 | switch (details->text_format) { |
e53f32d2 GS |
618 | case FORMAT_HEX: |
619 | if (ch_rem >= 4) { | |
620 | ch_rem--; | |
621 | set_logic_level(inc, ch_idx + 3, bits & (1 << 3)); | |
622 | } | |
623 | /* FALLTHROUGH */ | |
624 | case FORMAT_OCT: | |
625 | if (ch_rem >= 3) { | |
626 | ch_rem--; | |
627 | set_logic_level(inc, ch_idx + 2, bits & (1 << 2)); | |
628 | } | |
629 | if (ch_rem >= 2) { | |
630 | ch_rem--; | |
631 | set_logic_level(inc, ch_idx + 1, bits & (1 << 1)); | |
632 | } | |
633 | /* FALLTHROUGH */ | |
634 | case FORMAT_BIN: | |
635 | ch_rem--; | |
636 | set_logic_level(inc, ch_idx + 0, bits & (1 << 0)); | |
637 | break; | |
836fac9c GS |
638 | case FORMAT_NONE: |
639 | /* ShouldNotHappen(TM), but silences compiler warning. */ | |
640 | return SR_ERR; | |
4a35548b | 641 | } |
e53f32d2 | 642 | ch_idx += ch_inc; |
4a35548b | 643 | } |
e53f32d2 GS |
644 | /* |
645 | * TODO Determine whether the availability of extra input data | |
646 | * for unhandled logic channels is worth warning here. In this | |
647 | * implementation users are in control, and can have the more | |
648 | * significant bits ignored (which can be considered a feature | |
649 | * and not really a limitation). | |
650 | */ | |
4a35548b MS |
651 | |
652 | return SR_OK; | |
653 | } | |
654 | ||
836fac9c GS |
655 | /** |
656 | * @brief Parse routine which ignores the input text. | |
657 | * | |
658 | * This routine exists to unify dispatch code paths, mapping input file | |
659 | * columns' data types to their respective parse routines. | |
660 | */ | |
661 | static int parse_ignore(const char *column, struct context *inc, | |
662 | const struct column_details *details) | |
663 | { | |
664 | (void)column; | |
665 | (void)inc; | |
666 | (void)details; | |
667 | return SR_OK; | |
668 | } | |
669 | ||
670 | typedef int (*col_parse_cb)(const char *column, struct context *inc, | |
671 | const struct column_details *details); | |
672 | ||
673 | static const col_parse_cb col_parse_funcs[] = { | |
674 | [FORMAT_NONE] = parse_ignore, | |
675 | [FORMAT_BIN] = parse_logic, | |
676 | [FORMAT_OCT] = parse_logic, | |
677 | [FORMAT_HEX] = parse_logic, | |
678 | }; | |
679 | ||
41d214f6 | 680 | static int init(struct sr_input *in, GHashTable *options) |
4a35548b | 681 | { |
41d214f6 | 682 | struct context *inc; |
1a920e33 | 683 | size_t single_column, first_column, logic_channels; |
41d214f6 | 684 | const char *s; |
836fac9c | 685 | enum single_col_format format; |
1a920e33 | 686 | char format_char; |
4a35548b | 687 | |
836fac9c GS |
688 | in->sdi = g_malloc0(sizeof(*in->sdi)); |
689 | in->priv = inc = g_malloc0(sizeof(*inc)); | |
4a35548b | 690 | |
72903e9d | 691 | single_column = g_variant_get_uint32(g_hash_table_lookup(options, "single_column")); |
72903e9d | 692 | logic_channels = g_variant_get_uint32(g_hash_table_lookup(options, "logic_channels")); |
41d214f6 | 693 | inc->delimiter = g_string_new(g_variant_get_string( |
72903e9d | 694 | g_hash_table_lookup(options, "column_separator"), NULL)); |
836fac9c | 695 | if (!inc->delimiter->len) { |
72903e9d | 696 | sr_err("Column separator cannot be empty."); |
41d214f6 | 697 | return SR_ERR_ARG; |
4a35548b | 698 | } |
72903e9d | 699 | s = g_variant_get_string(g_hash_table_lookup(options, "single_format"), NULL); |
836fac9c GS |
700 | if (g_ascii_strncasecmp(s, "bin", 3) == 0) { |
701 | format = FORMAT_BIN; | |
702 | } else if (g_ascii_strncasecmp(s, "hex", 3) == 0) { | |
703 | format = FORMAT_HEX; | |
704 | } else if (g_ascii_strncasecmp(s, "oct", 3) == 0) { | |
705 | format = FORMAT_OCT; | |
41d214f6 | 706 | } else { |
72903e9d | 707 | sr_err("Invalid single-column format: '%s'", s); |
41d214f6 | 708 | return SR_ERR_ARG; |
4a35548b | 709 | } |
41d214f6 | 710 | inc->comment = g_string_new(g_variant_get_string( |
72903e9d | 711 | g_hash_table_lookup(options, "comment_leader"), NULL)); |
41d214f6 | 712 | if (g_string_equal(inc->comment, inc->delimiter)) { |
e53f32d2 GS |
713 | /* |
714 | * Using the same sequence as comment leader and column | |
72903e9d GS |
715 | * separator won't work. The user probably specified ';' |
716 | * as the column separator but did not adjust the comment | |
e53f32d2 GS |
717 | * leader. Try DWIM, drop comment strippin support here. |
718 | */ | |
72903e9d | 719 | sr_warn("Comment leader and column separator conflict, disabling comment support."); |
41d214f6 | 720 | g_string_truncate(inc->comment, 0); |
4a35548b | 721 | } |
6e8d95a5 | 722 | inc->samplerate = g_variant_get_uint64(g_hash_table_lookup(options, "samplerate")); |
72903e9d | 723 | first_column = g_variant_get_uint32(g_hash_table_lookup(options, "first_column")); |
de8fe3b5 | 724 | inc->use_header = g_variant_get_boolean(g_hash_table_lookup(options, "header")); |
72903e9d | 725 | inc->start_line = g_variant_get_uint32(g_hash_table_lookup(options, "start_line")); |
41d214f6 | 726 | if (inc->start_line < 1) { |
6433156c | 727 | sr_err("Invalid start line %zu.", inc->start_line); |
41d214f6 | 728 | return SR_ERR_ARG; |
4a35548b MS |
729 | } |
730 | ||
e53f32d2 | 731 | /* |
1a920e33 GS |
732 | * Scan flexible, to get prefered format specs which describe |
733 | * the input file's data formats. As well as some simple specs | |
734 | * for backwards compatibility and user convenience. | |
735 | * | |
736 | * This logic ends up with a copy of the format string, either | |
737 | * user provided or internally derived. Actual creation of the | |
738 | * column processing details gets deferred until the first line | |
739 | * of input data was seen. To support automatic determination of | |
740 | * e.g. channel counts from column counts. | |
e53f32d2 | 741 | */ |
72903e9d | 742 | s = g_variant_get_string(g_hash_table_lookup(options, "column_formats"), NULL); |
2142a79b | 743 | if (s && *s) { |
1a920e33 | 744 | inc->column_formats = g_strdup(s); |
72903e9d | 745 | sr_dbg("User specified column_formats: %s.", s); |
1a920e33 GS |
746 | } else if (single_column && logic_channels) { |
747 | format_char = col_format_char[format]; | |
748 | if (single_column == 1) { | |
749 | inc->column_formats = g_strdup_printf("%c%zu", | |
750 | format_char, logic_channels); | |
e53f32d2 | 751 | } else { |
1a920e33 GS |
752 | inc->column_formats = g_strdup_printf("%zu-,%c%zu", |
753 | single_column - 1, | |
754 | format_char, logic_channels); | |
e53f32d2 | 755 | } |
72903e9d | 756 | sr_dbg("Backwards compat single_column, col %zu, fmt %s, bits %zu -> %s.", |
1a920e33 GS |
757 | single_column, col_format_text[format], logic_channels, |
758 | inc->column_formats); | |
759 | } else if (!single_column) { | |
760 | if (first_column > 1) { | |
761 | inc->column_formats = g_strdup_printf("%zu-,%zul", | |
762 | first_column - 1, logic_channels); | |
763 | } else { | |
764 | inc->column_formats = g_strdup_printf("%zul", | |
765 | logic_channels); | |
766 | } | |
767 | sr_dbg("Backwards compat multi-column, col %zu, chans %zu -> %s.", | |
768 | first_column, logic_channels, | |
769 | inc->column_formats); | |
e53f32d2 | 770 | } else { |
72903e9d | 771 | sr_warn("Unknown or unsupported columns layout spec, assuming simple multi-column mode."); |
1a920e33 | 772 | inc->column_formats = g_strdup("*l"); |
4a35548b MS |
773 | } |
774 | ||
41d214f6 BV |
775 | return SR_OK; |
776 | } | |
4a35548b | 777 | |
affaf540 GS |
778 | /* |
779 | * Check the channel list for consistency across file re-import. See | |
780 | * the VCD input module for more details and motivation. | |
781 | */ | |
782 | ||
783 | static void keep_header_for_reread(const struct sr_input *in) | |
784 | { | |
785 | struct context *inc; | |
786 | ||
787 | inc = in->priv; | |
788 | g_slist_free_full(inc->prev_sr_channels, sr_channel_free_cb); | |
789 | inc->prev_sr_channels = in->sdi->channels; | |
790 | in->sdi->channels = NULL; | |
791 | } | |
792 | ||
793 | static int check_header_in_reread(const struct sr_input *in) | |
794 | { | |
795 | struct context *inc; | |
796 | ||
797 | if (!in) | |
798 | return FALSE; | |
799 | inc = in->priv; | |
800 | if (!inc) | |
801 | return FALSE; | |
802 | if (!inc->prev_sr_channels) | |
803 | return TRUE; | |
804 | ||
805 | if (sr_channel_lists_differ(inc->prev_sr_channels, in->sdi->channels)) { | |
806 | sr_err("Channel list change not supported for file re-read."); | |
807 | return FALSE; | |
808 | } | |
809 | g_slist_free_full(in->sdi->channels, sr_channel_free_cb); | |
810 | in->sdi->channels = inc->prev_sr_channels; | |
811 | inc->prev_sr_channels = NULL; | |
812 | ||
813 | return TRUE; | |
814 | } | |
815 | ||
492dfa90 GS |
816 | static const char *delim_set = "\r\n"; |
817 | ||
329733d9 | 818 | static const char *get_line_termination(GString *buf) |
41d214f6 | 819 | { |
329733d9 | 820 | const char *term; |
4a35548b | 821 | |
41d214f6 BV |
822 | term = NULL; |
823 | if (g_strstr_len(buf->str, buf->len, "\r\n")) | |
824 | term = "\r\n"; | |
825 | else if (memchr(buf->str, '\n', buf->len)) | |
826 | term = "\n"; | |
827 | else if (memchr(buf->str, '\r', buf->len)) | |
828 | term = "\r"; | |
4a35548b | 829 | |
41d214f6 BV |
830 | return term; |
831 | } | |
4a35548b | 832 | |
41d214f6 BV |
833 | static int initial_parse(const struct sr_input *in, GString *buf) |
834 | { | |
835 | struct context *inc; | |
9e7af34e | 836 | size_t num_columns; |
e53f32d2 | 837 | size_t line_number, line_idx; |
41d214f6 | 838 | int ret; |
9e7af34e | 839 | char **lines, *line, **columns; |
41d214f6 BV |
840 | |
841 | ret = SR_OK; | |
842 | inc = in->priv; | |
843 | columns = NULL; | |
844 | ||
9e7af34e | 845 | /* Search for the first line to process (header or data). */ |
41d214f6 | 846 | line_number = 0; |
ef0b9935 GS |
847 | if (inc->termination) |
848 | lines = g_strsplit(buf->str, inc->termination, 0); | |
849 | else | |
850 | lines = g_strsplit_set(buf->str, delim_set, 0); | |
e53f32d2 | 851 | for (line_idx = 0; (line = lines[line_idx]); line_idx++) { |
41d214f6 BV |
852 | line_number++; |
853 | if (inc->start_line > line_number) { | |
e53f32d2 | 854 | sr_spew("Line %zu skipped (before start).", line_number); |
4a35548b MS |
855 | continue; |
856 | } | |
df0db9fd | 857 | if (line[0] == '\0') { |
41d214f6 BV |
858 | sr_spew("Blank line %zu skipped.", line_number); |
859 | continue; | |
860 | } | |
df0db9fd GS |
861 | strip_comment(line, inc->comment); |
862 | if (line[0] == '\0') { | |
41d214f6 | 863 | sr_spew("Comment-only line %zu skipped.", line_number); |
4a35548b MS |
864 | continue; |
865 | } | |
866 | ||
41d214f6 BV |
867 | /* Reached first proper line. */ |
868 | break; | |
869 | } | |
e53f32d2 | 870 | if (!line) { |
41d214f6 | 871 | /* Not enough data for a proper line yet. */ |
60107497 | 872 | ret = SR_ERR_NA; |
41d214f6 | 873 | goto out; |
4a35548b MS |
874 | } |
875 | ||
9e7af34e | 876 | /* Get the number of columns in the line. */ |
e53f32d2 | 877 | columns = split_line(line, inc); |
df0db9fd | 878 | if (!columns) { |
41d214f6 BV |
879 | sr_err("Error while parsing line %zu.", line_number); |
880 | ret = SR_ERR; | |
881 | goto out; | |
4a35548b | 882 | } |
4a35548b | 883 | num_columns = g_strv_length(columns); |
4a35548b | 884 | if (!num_columns) { |
e53f32d2 | 885 | sr_err("Error while parsing line %zu.", line_number); |
41d214f6 BV |
886 | ret = SR_ERR; |
887 | goto out; | |
4a35548b | 888 | } |
e53f32d2 GS |
889 | sr_dbg("DIAG Got %zu columns in text line: %s.", num_columns, line); |
890 | ||
1a920e33 | 891 | /* |
9e7af34e GS |
892 | * Interpret the user provided column format specs. This might |
893 | * involve inspection of the now received input text, to support | |
894 | * e.g. automatic detection of channel counts in the absence of | |
895 | * user provided specs. Optionally a header line is used to get | |
896 | * channels' names. | |
897 | * | |
898 | * Check the then created channels for consistency across .reset | |
899 | * and .receive sequences (file re-load). | |
1a920e33 | 900 | */ |
9e7af34e | 901 | ret = make_column_details_from_format(in, inc->column_formats, columns); |
1a920e33 GS |
902 | if (ret != SR_OK) { |
903 | sr_err("Cannot parse columns format using line %zu.", line_number); | |
904 | goto out; | |
4a35548b | 905 | } |
affaf540 GS |
906 | if (!check_header_in_reread(in)) { |
907 | ret = SR_ERR_DATA; | |
908 | goto out; | |
909 | } | |
4a35548b MS |
910 | |
911 | /* | |
9e7af34e | 912 | * Allocate buffer memory for datafeed submission of sample data. |
cd59e6ec GS |
913 | * Calculate the minimum buffer size to store the set of samples |
914 | * of all channels (unit size). Determine a larger buffer size | |
915 | * for datafeed submission that is a multiple of the unit size. | |
626c388a GS |
916 | * Allocate the larger buffer, the "sample buffer" will point |
917 | * to a location within that large buffer later. | |
4a35548b | 918 | */ |
836fac9c | 919 | inc->sample_unit_size = (inc->logic_channels + 7) / 8; |
8bc2fa6d | 920 | inc->datafeed_buf_size = CHUNK_SIZE; |
cd59e6ec GS |
921 | inc->datafeed_buf_size *= inc->sample_unit_size; |
922 | inc->datafeed_buffer = g_malloc(inc->datafeed_buf_size); | |
923 | inc->datafeed_buf_fill = 0; | |
4a35548b | 924 | |
41d214f6 BV |
925 | out: |
926 | if (columns) | |
927 | g_strfreev(columns); | |
928 | g_strfreev(lines); | |
4a35548b | 929 | |
41d214f6 | 930 | return ret; |
4a35548b MS |
931 | } |
932 | ||
4439363a GS |
933 | /* |
934 | * Gets called from initial_receive(), which runs until the end-of-line | |
935 | * encoding of the input stream could get determined. Assumes that this | |
936 | * routine receives enough buffered initial input data to either see the | |
937 | * BOM when there is one, or that no BOM will follow when a text line | |
938 | * termination sequence was seen. Silently drops the UTF-8 BOM sequence | |
939 | * from the input buffer if one was seen. Does not care to protect | |
940 | * against multiple execution or dropping the BOM multiple times -- | |
941 | * there should be at most one in the input stream. | |
942 | */ | |
943 | static void initial_bom_check(const struct sr_input *in) | |
944 | { | |
945 | static const char *utf8_bom = "\xef\xbb\xbf"; | |
946 | ||
947 | if (in->buf->len < strlen(utf8_bom)) | |
948 | return; | |
949 | if (strncmp(in->buf->str, utf8_bom, strlen(utf8_bom)) != 0) | |
950 | return; | |
951 | g_string_erase(in->buf, 0, strlen(utf8_bom)); | |
952 | } | |
953 | ||
41d214f6 | 954 | static int initial_receive(const struct sr_input *in) |
4a35548b | 955 | { |
41d214f6 BV |
956 | struct context *inc; |
957 | GString *new_buf; | |
958 | int len, ret; | |
329733d9 UH |
959 | char *p; |
960 | const char *termination; | |
4a35548b | 961 | |
4439363a GS |
962 | initial_bom_check(in); |
963 | ||
41d214f6 | 964 | inc = in->priv; |
4a35548b | 965 | |
df0db9fd GS |
966 | termination = get_line_termination(in->buf); |
967 | if (!termination) | |
41d214f6 | 968 | /* Don't have a full line yet. */ |
d0181813 | 969 | return SR_ERR_NA; |
4a35548b | 970 | |
df0db9fd GS |
971 | p = g_strrstr_len(in->buf->str, in->buf->len, termination); |
972 | if (!p) | |
41d214f6 | 973 | /* Don't have a full line yet. */ |
d0181813 | 974 | return SR_ERR_NA; |
41d214f6 BV |
975 | len = p - in->buf->str - 1; |
976 | new_buf = g_string_new_len(in->buf->str, len); | |
977 | g_string_append_c(new_buf, '\0'); | |
4a35548b | 978 | |
41d214f6 BV |
979 | inc->termination = g_strdup(termination); |
980 | ||
981 | if (in->buf->str[0] != '\0') | |
982 | ret = initial_parse(in, new_buf); | |
983 | else | |
984 | ret = SR_OK; | |
985 | ||
986 | g_string_free(new_buf, TRUE); | |
987 | ||
988 | return ret; | |
989 | } | |
990 | ||
7f4c3a62 | 991 | static int process_buffer(struct sr_input *in, gboolean is_eof) |
41d214f6 | 992 | { |
41d214f6 BV |
993 | struct context *inc; |
994 | gsize num_columns; | |
e53f32d2 | 995 | size_t line_idx, col_idx, col_nr; |
836fac9c GS |
996 | const struct column_details *details; |
997 | col_parse_cb parse_func; | |
ad6a2bee | 998 | int ret; |
e53f32d2 | 999 | char *p, **lines, *line, **columns, *column; |
41d214f6 | 1000 | |
41d214f6 | 1001 | inc = in->priv; |
d0181813 | 1002 | if (!inc->started) { |
bee2b016 | 1003 | std_session_send_df_header(in->sdi); |
d0181813 | 1004 | inc->started = TRUE; |
4a35548b MS |
1005 | } |
1006 | ||
4555d3bd GS |
1007 | /* |
1008 | * Consider empty input non-fatal. Keep accumulating input until | |
1009 | * at least one full text line has become available. Grab the | |
1010 | * maximum amount of accumulated data that consists of full text | |
1011 | * lines, and process what has been received so far, leaving not | |
1012 | * yet complete lines for the next invocation. | |
7f4c3a62 GS |
1013 | * |
1014 | * Enforce that all previously buffered data gets processed in | |
1015 | * the "EOF" condition. Do not insist in the presence of the | |
1016 | * termination sequence for the last line (may often be missing | |
1017 | * on Windows). A present termination sequence will just result | |
1018 | * in the "execution of an empty line", and does not harm. | |
4555d3bd GS |
1019 | */ |
1020 | if (!in->buf->len) | |
1021 | return SR_OK; | |
7f4c3a62 GS |
1022 | if (is_eof) { |
1023 | p = in->buf->str + in->buf->len; | |
1024 | } else { | |
1025 | p = g_strrstr_len(in->buf->str, in->buf->len, inc->termination); | |
1026 | if (!p) | |
1027 | return SR_ERR; | |
1028 | *p = '\0'; | |
1029 | p += strlen(inc->termination); | |
1030 | } | |
41d214f6 | 1031 | g_strstrip(in->buf->str); |
4a35548b | 1032 | |
18078d05 | 1033 | ret = SR_OK; |
ef0b9935 | 1034 | lines = g_strsplit(in->buf->str, inc->termination, 0); |
e53f32d2 | 1035 | for (line_idx = 0; (line = lines[line_idx]); line_idx++) { |
41d214f6 | 1036 | inc->line_number++; |
ef0b9935 GS |
1037 | if (inc->line_number < inc->start_line) { |
1038 | sr_spew("Line %zu skipped (before start).", inc->line_number); | |
1039 | continue; | |
1040 | } | |
df0db9fd | 1041 | if (line[0] == '\0') { |
41d214f6 | 1042 | sr_spew("Blank line %zu skipped.", inc->line_number); |
4a35548b MS |
1043 | continue; |
1044 | } | |
1045 | ||
1046 | /* Remove trailing comment. */ | |
df0db9fd GS |
1047 | strip_comment(line, inc->comment); |
1048 | if (line[0] == '\0') { | |
41d214f6 | 1049 | sr_spew("Comment-only line %zu skipped.", inc->line_number); |
4a35548b MS |
1050 | continue; |
1051 | } | |
1052 | ||
160691b9 | 1053 | /* Skip the header line, its content was used as the channel names. */ |
de8fe3b5 | 1054 | if (inc->use_header && !inc->header_seen) { |
160691b9 | 1055 | sr_spew("Header line %zu skipped.", inc->line_number); |
de8fe3b5 | 1056 | inc->header_seen = TRUE; |
160691b9 JS |
1057 | continue; |
1058 | } | |
1059 | ||
e53f32d2 GS |
1060 | /* Split the line into columns, check for minimum length. */ |
1061 | columns = split_line(line, inc); | |
df0db9fd | 1062 | if (!columns) { |
41d214f6 | 1063 | sr_err("Error while parsing line %zu.", inc->line_number); |
2355d229 | 1064 | g_strfreev(lines); |
4a35548b MS |
1065 | return SR_ERR; |
1066 | } | |
4a35548b | 1067 | num_columns = g_strv_length(columns); |
e53f32d2 GS |
1068 | if (num_columns < inc->column_want_count) { |
1069 | sr_err("Insufficient column count %zu in line %zu.", | |
1070 | num_columns, inc->line_number); | |
4a35548b | 1071 | g_strfreev(columns); |
2355d229 | 1072 | g_strfreev(lines); |
4a35548b MS |
1073 | return SR_ERR; |
1074 | } | |
1075 | ||
836fac9c | 1076 | /* Have the columns of the current text line processed. */ |
626c388a | 1077 | clear_logic_samples(inc); |
e53f32d2 GS |
1078 | for (col_idx = 0; col_idx < inc->column_want_count; col_idx++) { |
1079 | column = columns[col_idx]; | |
1080 | col_nr = col_idx + 1; | |
836fac9c GS |
1081 | details = lookup_column_details(inc, col_nr); |
1082 | if (!details || !details->text_format) | |
1083 | continue; | |
1084 | parse_func = col_parse_funcs[details->text_format]; | |
1085 | if (!parse_func) | |
1086 | continue; | |
1087 | ret = parse_func(column, inc, details); | |
e53f32d2 GS |
1088 | if (ret != SR_OK) { |
1089 | g_strfreev(columns); | |
1090 | g_strfreev(lines); | |
1091 | return SR_ERR; | |
1092 | } | |
4a35548b MS |
1093 | } |
1094 | ||
626c388a GS |
1095 | /* Send sample data to the session bus (buffered). */ |
1096 | ret = queue_logic_samples(in); | |
41d214f6 | 1097 | if (ret != SR_OK) { |
4a35548b | 1098 | sr_err("Sending samples failed."); |
cd59e6ec | 1099 | g_strfreev(columns); |
2355d229 | 1100 | g_strfreev(lines); |
4a35548b MS |
1101 | return SR_ERR; |
1102 | } | |
cd59e6ec | 1103 | |
41d214f6 BV |
1104 | g_strfreev(columns); |
1105 | } | |
1106 | g_strfreev(lines); | |
241c386a | 1107 | g_string_erase(in->buf, 0, p - in->buf->str); |
41d214f6 | 1108 | |
7066fd46 | 1109 | return ret; |
41d214f6 BV |
1110 | } |
1111 | ||
7066fd46 | 1112 | static int receive(struct sr_input *in, GString *buf) |
41d214f6 BV |
1113 | { |
1114 | struct context *inc; | |
7066fd46 BV |
1115 | int ret; |
1116 | ||
1117 | g_string_append_len(in->buf, buf->str, buf->len); | |
41d214f6 BV |
1118 | |
1119 | inc = in->priv; | |
1a920e33 | 1120 | if (!inc->column_seen_count) { |
df0db9fd GS |
1121 | ret = initial_receive(in); |
1122 | if (ret == SR_ERR_NA) | |
7066fd46 BV |
1123 | /* Not enough data yet. */ |
1124 | return SR_OK; | |
1125 | else if (ret != SR_OK) | |
1126 | return SR_ERR; | |
1127 | ||
1128 | /* sdi is ready, notify frontend. */ | |
1129 | in->sdi_ready = TRUE; | |
41d214f6 | 1130 | return SR_OK; |
7066fd46 BV |
1131 | } |
1132 | ||
7f4c3a62 | 1133 | ret = process_buffer(in, FALSE); |
7066fd46 BV |
1134 | |
1135 | return ret; | |
1136 | } | |
1137 | ||
1138 | static int end(struct sr_input *in) | |
1139 | { | |
1140 | struct context *inc; | |
7066fd46 | 1141 | int ret; |
41d214f6 | 1142 | |
7066fd46 | 1143 | if (in->sdi_ready) |
7f4c3a62 | 1144 | ret = process_buffer(in, TRUE); |
7066fd46 BV |
1145 | else |
1146 | ret = SR_OK; | |
cd59e6ec GS |
1147 | if (ret != SR_OK) |
1148 | return ret; | |
1149 | ||
626c388a | 1150 | ret = flush_logic_samples(in); |
cd59e6ec GS |
1151 | if (ret != SR_OK) |
1152 | return ret; | |
7066fd46 BV |
1153 | |
1154 | inc = in->priv; | |
3be42bc2 | 1155 | if (inc->started) |
bee2b016 | 1156 | std_session_send_df_end(in->sdi); |
4a35548b | 1157 | |
7066fd46 BV |
1158 | return ret; |
1159 | } | |
1160 | ||
d5cc282f | 1161 | static void cleanup(struct sr_input *in) |
7066fd46 BV |
1162 | { |
1163 | struct context *inc; | |
1164 | ||
affaf540 GS |
1165 | keep_header_for_reread(in); |
1166 | ||
7066fd46 BV |
1167 | inc = in->priv; |
1168 | ||
b1f83103 | 1169 | g_free(inc->termination); |
539188e5 | 1170 | inc->termination = NULL; |
cd59e6ec | 1171 | g_free(inc->datafeed_buffer); |
539188e5 | 1172 | inc->datafeed_buffer = NULL; |
4a35548b MS |
1173 | } |
1174 | ||
ad93bfb0 SA |
1175 | static int reset(struct sr_input *in) |
1176 | { | |
1177 | struct context *inc = in->priv; | |
1178 | ||
1179 | cleanup(in); | |
1180 | inc->started = FALSE; | |
1181 | g_string_truncate(in->buf, 0); | |
1182 | ||
1183 | return SR_OK; | |
1184 | } | |
1185 | ||
c6aa9870 | 1186 | enum option_index { |
2142a79b | 1187 | OPT_COL_FMTS, |
c6aa9870 | 1188 | OPT_SINGLE_COL, |
72903e9d | 1189 | OPT_FIRST_COL, |
c6aa9870 | 1190 | OPT_NUM_LOGIC, |
c6aa9870 | 1191 | OPT_FORMAT, |
c6aa9870 | 1192 | OPT_START, |
72903e9d GS |
1193 | OPT_HEADER, |
1194 | OPT_RATE, | |
1195 | OPT_DELIM, | |
1196 | OPT_COMMENT, | |
c6aa9870 GS |
1197 | OPT_MAX, |
1198 | }; | |
1199 | ||
41d214f6 | 1200 | static struct sr_option options[] = { |
72903e9d GS |
1201 | [OPT_COL_FMTS] = { |
1202 | "column_formats", "Column format specs", | |
1203 | "Specifies text columns data types: comma separated list of [<cols>]<fmt>[<bits>], with -/x/o/b/l format specifiers.", | |
1204 | NULL, NULL, | |
1205 | }, | |
1206 | [OPT_SINGLE_COL] = { | |
1207 | "single_column", "Single column", | |
1208 | "Enable single-column mode, exclusively use text from the specified column (number starting at 1).", | |
1209 | NULL, NULL, | |
1210 | }, | |
1211 | [OPT_FIRST_COL] = { | |
1212 | "first_column", "First column", | |
1213 | "Number of the first column with logic data in simple multi-column mode (number starting at 1, default 1).", | |
1214 | NULL, NULL, | |
1215 | }, | |
1216 | [OPT_NUM_LOGIC] = { | |
1217 | "logic_channels", "Number of logic channels", | |
1218 | "Logic channel count, required in simple single-column mode, defaults to \"all remaining columns\" in simple multi-column mode. Obsoleted by 'column_formats'.", | |
1219 | NULL, NULL, | |
1220 | }, | |
1221 | [OPT_FORMAT] = { | |
1222 | "single_format", "Data format for simple single-column mode.", | |
1223 | "The number format of single-column mode input data: bin, hex, oct.", | |
1224 | NULL, NULL, | |
1225 | }, | |
1226 | [OPT_START] = { | |
1227 | "start_line", "Start line", | |
1228 | "The line number at which to start processing input text (default: 1).", | |
1229 | NULL, NULL, | |
1230 | }, | |
1231 | [OPT_HEADER] = { | |
1232 | "header", "Get channel names from first line.", | |
1233 | "Use the first processed line's column captions (when available) as channel names.", | |
1234 | NULL, NULL, | |
1235 | }, | |
1236 | [OPT_RATE] = { | |
1237 | "samplerate", "Samplerate (Hz)", | |
1238 | "The input data's sample rate in Hz.", | |
1239 | NULL, NULL, | |
1240 | }, | |
1241 | [OPT_DELIM] = { | |
1242 | "column_separator", "Column separator", | |
1243 | "The sequence which separates text columns. Non-empty text, comma by default.", | |
1244 | NULL, NULL, | |
1245 | }, | |
1246 | [OPT_COMMENT] = { | |
1247 | "comment_leader", "Comment leader character", | |
1248 | "The text which starts comments at the end of text lines.", | |
1249 | NULL, NULL, | |
1250 | }, | |
c6aa9870 | 1251 | [OPT_MAX] = ALL_ZERO, |
41d214f6 BV |
1252 | }; |
1253 | ||
2c240774 | 1254 | static const struct sr_option *get_options(void) |
41d214f6 | 1255 | { |
31c41782 UH |
1256 | GSList *l; |
1257 | ||
41d214f6 | 1258 | if (!options[0].def) { |
1a920e33 | 1259 | options[OPT_COL_FMTS].def = g_variant_ref_sink(g_variant_new_string("")); |
e53f32d2 | 1260 | options[OPT_SINGLE_COL].def = g_variant_ref_sink(g_variant_new_uint32(0)); |
72903e9d | 1261 | options[OPT_FIRST_COL].def = g_variant_ref_sink(g_variant_new_uint32(1)); |
e53f32d2 | 1262 | options[OPT_NUM_LOGIC].def = g_variant_ref_sink(g_variant_new_uint32(0)); |
c6aa9870 | 1263 | options[OPT_FORMAT].def = g_variant_ref_sink(g_variant_new_string("bin")); |
31c41782 UH |
1264 | l = NULL; |
1265 | l = g_slist_append(l, g_variant_ref_sink(g_variant_new_string("bin"))); | |
1266 | l = g_slist_append(l, g_variant_ref_sink(g_variant_new_string("hex"))); | |
1267 | l = g_slist_append(l, g_variant_ref_sink(g_variant_new_string("oct"))); | |
c6aa9870 | 1268 | options[OPT_FORMAT].values = l; |
e53f32d2 | 1269 | options[OPT_START].def = g_variant_ref_sink(g_variant_new_uint32(1)); |
72903e9d GS |
1270 | options[OPT_HEADER].def = g_variant_ref_sink(g_variant_new_boolean(FALSE)); |
1271 | options[OPT_RATE].def = g_variant_ref_sink(g_variant_new_uint64(0)); | |
1272 | options[OPT_DELIM].def = g_variant_ref_sink(g_variant_new_string(",")); | |
1273 | options[OPT_COMMENT].def = g_variant_ref_sink(g_variant_new_string(";")); | |
41d214f6 BV |
1274 | } |
1275 | ||
1276 | return options; | |
1277 | } | |
1278 | ||
d4c93774 | 1279 | SR_PRIV struct sr_input_module input_csv = { |
4a35548b | 1280 | .id = "csv", |
41d214f6 BV |
1281 | .name = "CSV", |
1282 | .desc = "Comma-separated values", | |
c7bc82ff | 1283 | .exts = (const char*[]){"csv", NULL}, |
41d214f6 | 1284 | .options = get_options, |
4a35548b | 1285 | .init = init, |
41d214f6 | 1286 | .receive = receive, |
7066fd46 | 1287 | .end = end, |
41d214f6 | 1288 | .cleanup = cleanup, |
ad93bfb0 | 1289 | .reset = reset, |
4a35548b | 1290 | }; |