#include <time.h>
#define MYMAIL_DB_MAGIC_TOKEN "mymail_index_file"
-#define VERSION "0.9.5"
+#define MYMAIL_VERSION "0.9.8"
+
+#define MYMAIL_DB_FORMAT_VERSION 1
#define MAX_NB_SEARCH_CONDITIONS 32
#define BUFFER_SIZE 65536
#define TOKEN_BUFFER_SIZE 1024
-#define LEADING_FROM_LINE_REGEXP_STRING "^From [^ ]* \\(Mon\\|Tue\\|Wed\\|Thu\\|Fri\\|Sat\\|Sun\\) \\(Jan\\|Feb\\|Mar\\|Apr\\|May\\|Jun\\|Jul\\|Aug\\|Sep\\|Oct\\|Nov\\|Dec\\) [ 123][0-9] [0-9][0-9]:[0-9][0-9]:[0-9][0-9] [0-9][0-9][0-9][0-9]\n$"
+#define LEADING_FROM_LINE_REGEXP_STRING "^From .*\\(Mon\\|Tue\\|Wed\\|Thu\\|Fri\\|Sat\\|Sun\\) \\(Jan\\|Feb\\|Mar\\|Apr\\|May\\|Jun\\|Jul\\|Aug\\|Sep\\|Oct\\|Nov\\|Dec\\) [ 0123][0-9] [0-9][0-9]:[0-9][0-9]:[0-9][0-9] [0-9][0-9][0-9][0-9]\n$"
/* Global variables! */
int global_quiet;
+int global_use_leading_time;
regex_t global_leading_from_line_regexp;
ID_DATE,
ID_PARTICIPANT,
ID_BODY,
- ID_INTERVAL,
+ ID_TIME_INTERVAL,
MAX_ID
};
/********************************************************************/
struct search_condition {
- int field_id;
+ int db_key;
+ regex_t db_value_regexp;
int negation;
- regex_t regexp;
- time_t interval_start, interval_stop;
+ time_t time_start, time_stop;
};
/********************************************************************/
const char *parse_token(char *token_buffer, size_t token_buffer_size,
char separator, const char *string) {
char *u = token_buffer;
+
while(*string == separator) { string++; }
- while(u < token_buffer + token_buffer_size - 1 && *string &&
- *string != separator) {
+
+ while(u < token_buffer + token_buffer_size - 1 && *string && *string != separator) {
*(u++) = *(string++);
}
+
while(*string == separator) { string++; }
+
*u = '\0';
return string;
}
/*********************************************************************/
void print_version(FILE *out) {
- fprintf(out, "mymail version %s (%s)\n", VERSION, UNAME);
+ fprintf(out, "mymail version %s (%s)\n", MYMAIL_VERSION, UNAME);
}
void print_usage(FILE *out) {
fprintf(out, " print the version number\n");
fprintf(out, " -q, --quiet\n");
fprintf(out, " do not print information during search\n");
+ fprintf(out, " -t, --use-leading-time\n");
+ fprintf(out, " use the time stamp from the leading line of each mail and not the Date:\n");
+ fprintf(out, " field\n");
fprintf(out, " -p <db filename pattern>, --db-pattern <db filename pattern>\n");
fprintf(out, " set the db filename pattern for recursive search\n");
fprintf(out, " -r <db root path>, --db-root <db root path>\n");
fprintf(out, " set the mbox filename pattern for recursive search\n");
fprintf(out, " -s <search pattern>, --search <search pattern>\n");
fprintf(out, " search for matching mails in the db file\n");
- fprintf(out, " -d <db filename>, --db-file-generate <db filename>\n");
+ fprintf(out, " -d <db filename>, --db-file-output <db filename>\n");
fprintf(out, " set the db filename for indexing\n");
fprintf(out, " -i, --index\n");
fprintf(out, " index mails\n");
regexec(&global_leading_from_line_regexp, mbox_line, 0, 0, 0) == 0;
}
-int mbox_line_match_search(struct search_condition *condition,
- int mbox_id, const char *mbox_value) {
+int db_line_match_search(struct search_condition *condition,
+ int db_key, const char *db_value) {
- if(condition->field_id == ID_INTERVAL) {
- if(mbox_id == ID_LEADING_LINE) {
- const char *c;
- time_t t;
- struct tm tm;
-
- c = mbox_value;
- while(*c && *c != ' ') c++; while(*c && *c == ' ') c++;
- strptime(c, "%a %b %e %k:%M:%S %Y", &tm);
- t = mktime(&tm);
-
- return (t >= condition->interval_start &&
- (condition->interval_stop == 0 ||
- t <= condition->interval_stop));
- } else {
- return 0;
- }
- } else {
- return
- (
-
- (condition->field_id == mbox_id)
+ return
+ (
+ (condition->db_key == db_key)
- ||
+ ||
- (condition->field_id == ID_PARTICIPANT && (mbox_id == ID_LEADING_LINE ||
- mbox_id == ID_FROM ||
- mbox_id == ID_TO))
- ||
+ (condition->db_key == ID_PARTICIPANT && (db_key == ID_LEADING_LINE ||
+ db_key == ID_FROM ||
+ db_key == ID_TO))
+ ||
- (condition->field_id == ID_FROM && mbox_id == ID_LEADING_LINE)
+ (condition->db_key == ID_FROM && db_key == ID_LEADING_LINE)
- )
+ )
- &&
+ &&
- regexec(&condition->regexp, mbox_value, 0, 0, 0) == 0;
- }
+ regexec(&condition->db_value_regexp, db_value, 0, 0, 0) == 0;
}
void update_body_hits(char *mail_filename, int position_in_mail,
if(fgets(raw_mbox_line, BUFFER_SIZE, mail_file)) {
while(nb_body_hits < nb_body_conditions) {
- /* last_mbox_line_was_empty = (raw_mbox_line[0] == '\n'); */
- /* if(last_mbox_line_was_empty) { header = 0; } */
-
if(raw_mbox_line[0] == '\n') { header = 0; }
if(!header) {
for(n = 0; n < nb_search_conditions; n++) {
- if(search_conditions[n].field_id == ID_BODY && !hits[n]) {
+ if(search_conditions[n].db_key == ID_BODY && !hits[n]) {
hits[n] =
- (regexec(&search_conditions[n].regexp, raw_mbox_line, 0, 0, 0) == 0);
+ (regexec(&search_conditions[n].db_value_regexp, raw_mbox_line, 0, 0, 0) == 0);
if(hits[n]) {
nb_body_hits++;
}
}
void extract_mail(const char *mail_filename, unsigned long int position_in_mail,
- FILE *output_file) {
+ FILE *output_file) {
char raw_mbox_line[BUFFER_SIZE];
FILE *mail_file;
+ /* printf("Extract\n"); */
+
mail_file = safe_fopen(mail_filename, "r", "mbox for mail extraction");
fseek(mail_file, position_in_mail, SEEK_SET);
fprintf(output_file, "%s", raw_mbox_line);
while(1) {
if(!fgets(raw_mbox_line, BUFFER_SIZE, mail_file) ||
- (is_a_leading_from_line(raw_mbox_line))
- )
+ is_a_leading_from_line(raw_mbox_line))
break;
fprintf(output_file, "%s", raw_mbox_line);
}
fclose(mail_file);
}
+int check_full_mail_match(char *current_mail_filename,
+ time_t mail_time,
+ int nb_search_conditions,
+ struct search_condition *search_conditions,
+ int nb_body_conditions,
+ int *hits,
+ int current_position_in_mail) {
+ int n, nb_fulfilled_body_conditions;
+
+ for(n = 0; n < nb_search_conditions; n++) {
+ if(search_conditions[n].db_key == ID_TIME_INTERVAL) {
+ hits[n] = (mail_time >= search_conditions[n].time_start &&
+ (search_conditions[n].time_stop == 0 ||
+ mail_time <= search_conditions[n].time_stop));
+ }
+ }
+
+ /* We first check all conditions but the body ones */
+
+ for(n = 0; n < nb_search_conditions &&
+ ((search_conditions[n].db_key == ID_BODY) ||
+ xor(hits[n], search_conditions[n].negation)); n++);
+
+ if(n == nb_search_conditions) {
+
+ /* Now check the body ones */
+
+ nb_fulfilled_body_conditions = 0;
+
+ if(nb_body_conditions > 0) {
+ update_body_hits(current_mail_filename, current_position_in_mail,
+ nb_search_conditions, search_conditions,
+ nb_body_conditions,
+ hits);
+
+ for(n = 0; n < nb_search_conditions; n++) {
+ if(search_conditions[n].db_key == ID_BODY &&
+ xor(hits[n], search_conditions[n].negation)) {
+ nb_fulfilled_body_conditions++;
+ }
+ }
+ }
+ return nb_body_conditions == nb_fulfilled_body_conditions;
+ } else {
+ return 0;
+ }
+}
+
+/* We use the mail leading line time by default, and if we should and
+ can, we update with the Date: field */
+
+void update_time(int db_key, const char *db_value, time_t *t) {
+ const char *c;
+ struct tm tm;
+
+ if(db_key == ID_LEADING_LINE) {
+ c = db_value;
+ while(*c && *c != ' ') c++; while(*c && *c == ' ') c++;
+ /* printf("From %s", db_value); */
+ strptime(c, "%a %b %e %k:%M:%S %Y", &tm);
+ *t = mktime(&tm);
+ } else {
+ if(!global_use_leading_time) {
+ if(db_key == ID_DATE) {
+ if(strptime(db_value, "%a, %d %b %Y %k:%M:%S", &tm) ||
+ strptime(db_value, "%d %b %Y %k:%M:%S", &tm)) {
+ /* printf("Date: %s", db_value); */
+ *t = mktime(&tm);
+ }
+ }
+ }
+ }
+}
+
int search_in_db(const char *db_filename,
int nb_search_conditions,
struct search_condition *search_conditions,
FILE *output_file) {
- int hits[MAX_NB_SEARCH_CONDITIONS];
+ FILE *db_file;
char raw_db_line[BUFFER_SIZE];
char current_mail_filename[PATH_MAX + 1];
+ char db_key_string[TOKEN_BUFFER_SIZE];
+ char position_in_file_string[TOKEN_BUFFER_SIZE];
unsigned long int current_position_in_mail;
- char mbox_key[TOKEN_BUFFER_SIZE];
- const char *mbox_value;
- int mbox_id;
- int already_written, m, n;
- int nb_body_conditions, nb_fulfilled_body_conditions;
- FILE *db_file;
+ const char *db_value;
+ int db_key;
+ int hits[MAX_NB_SEARCH_CONDITIONS];
+ int nb_body_conditions, need_time;
int nb_extracted_mails;
+ time_t mail_time;
+
+ int m, n;
nb_extracted_mails = 0;
/* Then parse the said db file */
current_position_in_mail = 0;
- already_written = 0;
for(n = 0; n < nb_search_conditions; n++) { hits[n] = 0; }
nb_body_conditions = 0;
+ need_time = 0;
+ mail_time = 0;
+
for(n = 0; n < nb_search_conditions; n++) {
- if(search_conditions[n].field_id == ID_BODY) {
+ if(search_conditions[n].db_key == ID_BODY) {
nb_body_conditions++;
}
+ else if(search_conditions[n].db_key == ID_TIME_INTERVAL) {
+ need_time = 1;
+ }
}
strcpy(current_mail_filename, "");
while(fgets(raw_db_line, BUFFER_SIZE, db_file)) {
- mbox_value = parse_token(mbox_key, TOKEN_BUFFER_SIZE, ' ', raw_db_line);
-
- if(strcmp("mail", mbox_key) == 0) {
- char position_in_file_string[TOKEN_BUFFER_SIZE];
+ db_value = parse_token(db_key_string, TOKEN_BUFFER_SIZE, ' ', raw_db_line);
+ if(strcmp("mail", db_key_string) == 0) {
if(current_mail_filename[0]) {
-
- /* We first check all conditions but the body ones */
-
- for(n = 0; n < nb_search_conditions &&
- ((search_conditions[n].field_id == ID_BODY) ||
- xor(hits[n], search_conditions[n].negation)); n++);
-
- if(n == nb_search_conditions) {
-
- /* Now check the body ones */
-
- nb_fulfilled_body_conditions = 0;
-
- if(nb_body_conditions > 0) {
- update_body_hits(current_mail_filename, current_position_in_mail,
- nb_search_conditions, search_conditions,
- nb_body_conditions,
- hits);
-
- for(n = 0; n < nb_search_conditions; n++) {
- if(search_conditions[n].field_id == ID_BODY &&
- xor(hits[n], search_conditions[n].negation)) {
- nb_fulfilled_body_conditions++;
- }
- }
- }
-
- if(nb_body_conditions == nb_fulfilled_body_conditions) {
- nb_extracted_mails++;
- extract_mail(current_mail_filename, current_position_in_mail, output_file);
- }
+ if(check_full_mail_match(current_mail_filename,
+ mail_time,
+ nb_search_conditions, search_conditions,
+ nb_body_conditions, hits, current_position_in_mail)) {
+ extract_mail(current_mail_filename, current_position_in_mail, output_file);
+ nb_extracted_mails++;
}
}
for(n = 0; n < nb_search_conditions; n++) { hits[n] = 0; }
-
- mbox_value = parse_token(position_in_file_string, TOKEN_BUFFER_SIZE, ' ', mbox_value);
- mbox_value = parse_token(current_mail_filename, TOKEN_BUFFER_SIZE, '\n', mbox_value);
+ db_value = parse_token(position_in_file_string, TOKEN_BUFFER_SIZE, ' ', db_value);
+ db_value = parse_token(current_mail_filename, PATH_MAX+1, '\n', db_value);
current_position_in_mail = atol(position_in_file_string);
- already_written = 0;
}
else {
- mbox_id = -1;
- for(m = 0; (m < MAX_ID) && mbox_id == -1; m++) {
- if(strncmp(field_keys[m], mbox_key, strlen(mbox_key)) == 0) {
- mbox_id = m;
+ db_key = -1;
+ for(m = 0; (m < MAX_ID) && db_key == -1; m++) {
+ if(strncmp(field_keys[m], db_key_string, strlen(db_key_string)) == 0) {
+ db_key = m;
}
}
+
for(n = 0; n < nb_search_conditions; n++) {
- hits[n] |= mbox_line_match_search(&search_conditions[n],
- mbox_id, mbox_value);
+ hits[n] |= db_line_match_search(&search_conditions[n],
+ db_key, db_value);
}
+
+ if(need_time) {
+ update_time(db_key, db_value, &mail_time);
+ }
+ }
+ }
+
+ if(current_mail_filename[0]) {
+ if(check_full_mail_match(current_mail_filename,
+ mail_time,
+ nb_search_conditions, search_conditions,
+ nb_body_conditions, hits, current_position_in_mail)) {
+ extract_mail(current_mail_filename, current_position_in_mail, output_file);
+ nb_extracted_mails++;
}
}
while(fgets(raw_mbox_line, BUFFER_SIZE, file)) {
if(is_a_leading_from_line(raw_mbox_line)) {
+ /* This starts a new mail */
if(in_header) {
fprintf(stderr,
"Got a ^\"From \" in the header in %s:%lu.\n",
mbox_filename, position_in_file);
fprintf(stderr, "%s", raw_mbox_line);
}
+
+ /* printf("LEADING_LINE %s", raw_mbox_line); */
+
in_header = 1;
new_header = 1;
} else if(raw_mbox_line[0] == '\n') {
- if(in_header) { in_header = 0; }
+ if(in_header) {
+ in_header = 0;
+ /* We leave the header, index the current line */
+ if(full_line[0]) {
+ /* printf("INDEX %s\n", full_line); */
+ index_one_mbox_line(nb_fields_to_parse, fields_to_parse, full_line, db_file);
+ }
+ end_of_full_line = full_line;
+ *end_of_full_line = '\0';
+ }
}
if(in_header) {
}
if(raw_mbox_line[0] == ' ' || raw_mbox_line[0] == '\t') {
+ /* Continuation of a line */
char *start = raw_mbox_line;
while(*start == ' ' || *start == '\t') start++;
*(end_of_full_line++) = ' ';
}
else {
- /*
- if(!((raw_mbox_line[0] >= 'a' && raw_mbox_line[0] <= 'z') ||
- (raw_mbox_line[0] >= 'A' && raw_mbox_line[0] <= 'Z'))) {
- fprintf(stderr,
- "Header line syntax error %s:%lu.\n",
- mbox_filename, position_in_file);
- fprintf(stderr, "%s", raw_mbox_line);
- }
- */
+ /* Start a new header line, not a continuation */
if(full_line[0]) {
+ /* printf("INDEX %s\n", full_line); */
index_one_mbox_line(nb_fields_to_parse, fields_to_parse, full_line, db_file);
}
{ "help", no_argument, 0, 'h' },
{ "version", no_argument, 0, 'v' },
{ "quiet", no_argument, 0, 'q' },
- { "db-file-generate", 1, 0, 'd' },
+ { "use-leading-time", no_argument, 0, 't' },
+ { "db-file-output", 1, 0, 'd' },
{ "db-pattern", 1, 0, 'p' },
{ "db-root", 1, 0, 'r' },
{ "db-list", 1, 0, 'l' },
struct time_criterion {
char *label;
+ int day_criterion;
int start_hour, end_hour;
int past_week_day;
};
static struct time_criterion time_criteria[] = {
- { "8h", 8, -1, -1 },
- { "today", 24, -1, -1 },
- { "24h", 24, -1, -1 },
- { "48h", 48, -1, -1 },
- { "week", 24 * 7, -1, -1 },
- { "month", 24 * 31, -1, -1 },
- { "year", 24 * 365, -1, -1 },
-
- { "yesterday", 48, 24, -1 },
-
- { "monday", -1, -1, 1 },
- { "tuesday", -1, -1, 2 },
- { "wednesday", -1, -1, 3 },
- { "thursday", -1, -1, 4 },
- { "friday", -1, -1, 5 },
- { "saturday", -1, -1, 6 },
- { "sunday", -1, -1, 7 },
+ { "8h", 0, 8, -1, -1 },
+ { "24h", 0, 24, -1, -1 },
+ { "48h", 0, 48, -1, -1 },
+ { "week", 0, 24 * 7, -1, -1 },
+ { "month", 0, 24 * 31, -1, -1 },
+ { "trimester", 0, 24 * 92, -1, -1 },
+ { "year", 0, 24 * 365, -1, -1 },
+
+ { "yesterday", 1, -1, -1, -1 },
+ { "today", 1, -1, -1, 0 },
+
+ { "monday", 1, -1, -1, 1 },
+ { "tuesday", 1, -1, -1, 2 },
+ { "wednesday", 1, -1, -1, 3 },
+ { "thursday", 1, -1, -1, 4 },
+ { "friday", 1, -1, -1, 5 },
+ { "saturday", 1, -1, -1, 6 },
+ { "sunday", 1, -1, -1, 7 },
};
int delta_day;
t = time(0);
tm = localtime(&t);
- delta_day = (7 + tm->tm_wday - day) % 7;
- if(delta_day == 0) { delta_day = 7; }
+ if(day > 0) {
+ delta_day = (7 + tm->tm_wday - day) % 7;
+ } else {
+ delta_day = - day;
+ }
return t - (delta_day * 3600 * 24 + tm->tm_sec + 60 * tm->tm_min + 3600 * tm->tm_hour);
}
condition->negation = 0;
}
- condition->field_id = -1;
+ condition->db_key = -1;
- /* Recently */
+ /* Time condition */
for(k = 0; k < sizeof(time_criteria) / sizeof(struct time_criterion); k++) {
if(strcmp(time_criteria[k].label, search_field) == 0) {
- condition->field_id = ID_INTERVAL;
- if(time_criteria[k].past_week_day < 0) {
- condition->interval_start = time(0) - 3600 * time_criteria[k].start_hour;
+ condition->db_key = ID_TIME_INTERVAL;
+ if(time_criteria[k].day_criterion) {
+ condition->time_start = time_for_past_day(time_criteria[k].past_week_day);
+ condition->time_stop = condition->time_start + 3600 * 24;
+ } else {
+ condition->time_start = time(0) - 3600 * time_criteria[k].start_hour;
if(time_criteria[k].end_hour >= 0) {
- condition->interval_stop = time(0) - 3600 * time_criteria[k].end_hour;
+ condition->time_stop = time(0) - 3600 * time_criteria[k].end_hour;
} else {
- condition->interval_stop = 0;
+ condition->time_stop = 0;
}
- } else {
- condition->interval_start = time_for_past_day(time_criteria[k].past_week_day);
- condition->interval_stop = condition->interval_start + 3600 * 24;
}
+
break;
}
}
- if(condition->field_id == -1) {
+ if(condition->db_key == -1) {
/* No time condition matched, look for the search fields */
- for(m = 0; (m < MAX_ID) && condition->field_id == -1; m++) {
+ for(m = 0; (m < MAX_ID) && condition->db_key == -1; m++) {
if(strncmp(field_keys[m], search_field, strlen(search_field)) == 0) {
- condition->field_id = m;
+ condition->db_key = m;
}
}
/* None match, if there is a default search field, re-run the search with it */
- if(condition->field_id == -1) {
+ if(condition->db_key == -1) {
if(default_search_field) {
- for(m = 0; (m < MAX_ID) && condition->field_id == -1; m++) {
+ for(m = 0; (m < MAX_ID) && condition->db_key == -1; m++) {
if(strncmp(field_keys[m],
default_search_field, strlen(default_search_field)) == 0) {
- condition->field_id = m;
+ condition->db_key = m;
}
}
string = full_string;
+ if(string[0] == '!') { string++; }
}
}
- if(condition->field_id == -1) {
+ if(condition->db_key == -1) {
fprintf(stderr,
"mymail: Syntax error in field key \"%s\".\n",
search_field);
exit(EXIT_FAILURE);
}
- if(regcomp(&condition->regexp,
+ if(regcomp(&condition->db_value_regexp,
string,
REG_ICASE)) {
fprintf(stderr,
"mymail: Syntax error in regexp \"%s\" for field \"%s\".\n",
string,
- field_keys[condition->field_id]);
+ field_keys[condition->db_key]);
exit(EXIT_FAILURE);
}
}
}
void free_condition(struct search_condition *condition) {
- if(condition->field_id != ID_INTERVAL) {
- regfree(&condition->regexp);
+ if(condition->db_key != ID_TIME_INTERVAL) {
+ regfree(&condition->db_value_regexp);
}
}
}
global_quiet = 0;
+ global_use_leading_time = 0;
default_search_field = 0;
strncpy(output_filename, "", PATH_MAX);
global_quiet = 1;
break;
+ case 't':
+ global_use_leading_time = 1;
+ break;
+
case 'i':
action_index = 1;
break;
}
}
- fprintf(db_file, "%s version_%s raw\n", MYMAIL_DB_MAGIC_TOKEN, VERSION);
+ fprintf(db_file, "%s version_%s format_%d raw\n", MYMAIL_DB_MAGIC_TOKEN, MYMAIL_VERSION, MYMAIL_DB_FORMAT_VERSION);
while(optind < argc) {
recursive_index_mbox(db_file,
projects / mymail.git / blobdiff