X-Git-Url: https://fleuret.org/cgi-bin/gitweb/gitweb.cgi?p=mymail.git;a=blobdiff_plain;f=mymail.c;h=a04fde8eab17dba46399cac7ee8a1fd266628866;hp=6388365093510e913221db4b9e8fd5f8f7f9cef8;hb=HEAD;hpb=319a695493cac88fc6b14541d30ae0107f38caf9 diff --git a/mymail.c b/mymail.c index 6388365..a04fde8 100644 --- a/mymail.c +++ b/mymail.c @@ -19,10 +19,25 @@ * */ +/* + + mymail is a simple mail indexer. It can: + + (1) scan mbox files, and create a db file containing for each mail a + list of fields computed from its header. + + (2) read such a db file, gets all the mails matching regexp-defined + conditions on the fields, and generates a resulting mbox file. + + It is low-tech, simple, light and fast. + +*/ + #define _GNU_SOURCE #include #include +#include #include #include #include @@ -31,229 +46,754 @@ #include #include #include +#include + +#define MYMAIL_DB_MAGIC_TOKEN "mymail_index_file" +#define MYMAIL_VERSION "0.9.10" + +#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\\) [ 0123][0-9] [0-9][0-9]:[0-9][0-9]:[0-9][0-9] [0-9][0-9][0-9][0-9]\n$" + +/********************************************************************/ + +struct alias_node { + char *alias, *value; + struct alias_node *next; +}; + +/* Global variables! */ + +int global_quiet; +int global_use_leading_time; +int global_nb_mails_max; +int global_discard_mail_from_the_future; +regex_t global_leading_from_line_regexp; +struct alias_node *global_alias_list; +time_t global_current_time; + +/********************************************************************/ + +enum { + ID_MAIL = 0, + ID_LEADING_LINE, + ID_FROM, + ID_TO, + ID_SUBJECT, + ID_DATE, + ID_PARTICIPANT, + ID_BODY, + ID_TIME_INTERVAL, + ID_MAIL_ID, + ID_REFERENCE_ID, + ID_THREAD_ID, + MAX_ID +}; + +static char *field_keys[] = { + "mail", + "lead", + "from", + "to", + "subject", + "date", + "part", + "body", + "interval", + "mailid", + "reference", + "thread" +}; + +/********************************************************************/ -#define VERSION "0.1" +struct search_condition { + int db_key; + regex_t db_value_regexp; + int negation; + time_t time_start, time_stop; +}; -#define BUFFER_SIZE 16384 +/********************************************************************/ struct parsable_field { - char *name; + int id; + int cflags; char *regexp_string; regex_t regexp; }; -char *db_filename; -char *search_pattern; +static struct parsable_field fields_to_parse[] = { + { + ID_LEADING_LINE, + 0, + "^From ", + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } + }, + + { + ID_FROM, + REG_ICASE, + "^\\(from\\|reply-to\\|sender\\|return-path\\): ", + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } + }, + + { + ID_TO, + REG_ICASE, + "^\\(to\\|cc\\|bcc\\): ", + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } + }, + + { + ID_SUBJECT, + REG_ICASE, + "^subject: ", + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } + }, -int action_index; + { + ID_DATE, + REG_ICASE, + "^date: ", + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } + }, + + { + ID_MAIL_ID, + REG_ICASE, + "^message-id: ", + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } + }, -char *segment_next_field(char *current) { - while(*current && *current != ' ') current++; - *current = '\0'; current++; - while(*current && *current == ' ') current++; - return current; + { + ID_REFERENCE_ID, + REG_ICASE, + "^\\(in-reply-to\\|references\\): ", + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } + }, + +}; + +/********************************************************************/ + +int xor(int a, int b) { + return (a && !b) || (!a && b); } -void remove_eof(char *c) { - while(*c && *c != '\n' && *c != '\r') c++; - *c = '\0'; +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) { + *(u++) = *(string++); + } + + while(*string == separator) { string++; } + + *u = '\0'; + return string; } -/********************************************************************/ +char *default_value(char *current_value, + const char *env_variable, + const char *hard_default_value) { + if(current_value) { + return current_value; + } else { + char *env_value = getenv(env_variable); + if(env_value) { + return strdup(env_value); + } else if(hard_default_value) { + return strdup(hard_default_value); + } else { + return 0; + } + } +} -/* malloc with error checking. */ +/********************************************************************/ void *safe_malloc(size_t n) { void *p = malloc(n); if(!p && n != 0) { fprintf(stderr, - "mymail: can not allocate memory: %s\n", strerror(errno)); + "mymail: cannot allocate memory: %s\n", strerror(errno)); exit(EXIT_FAILURE); } return p; } +FILE *safe_fopen(const char *path, const char *mode, const char *comment) { + FILE *result = fopen(path, mode); + if(result) { + return result; + } else { + fprintf(stderr, + "mymail: Cannot open file '%s' (%s) with mode \"%s\": %s\n", + path, comment, mode, + strerror(errno)); + exit(EXIT_FAILURE); + } +} + /*********************************************************************/ -void usage(FILE *out) { - fprintf(out, "mymail version %s (%s)\n", VERSION, UNAME); +void print_version(FILE *out) { + fprintf(out, "mymail version %s (%s)\n", MYMAIL_VERSION, UNAME); +} + +void print_usage(FILE *out) { + print_version(out); fprintf(out, "Written by Francois Fleuret .\n"); fprintf(out, "\n"); - fprintf(out, "Usage: mymail [options] [ [ ...]]\n"); + fprintf(out, "Usage: mymail [options] [ [ ...]| [ ...]]\n"); fprintf(out, "\n"); + fprintf(out, " -h, --help\n"); + fprintf(out, " show this help\n"); + fprintf(out, " -v, --version\n"); + 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, " -f, --do-not-discard-mails-from-the-future\n"); + fprintf(out, " do not ignore mails with a date more than 24h in the future\n"); + fprintf(out, " -p , --db-pattern \n"); + fprintf(out, " set the db filename pattern for recursive search\n"); + fprintf(out, " -r , --db-root \n"); + fprintf(out, " set the db root path for recursive search\n"); + fprintf(out, " -l , --db-list \n"); + fprintf(out, " set the semicolon-separated list of db files for search\n"); + fprintf(out, " -m , --mbox-pattern \n"); + fprintf(out, " set the mbox filename pattern for recursive search\n"); + fprintf(out, " -s , --search \n"); + fprintf(out, " search for matching mails in the db file\n"); + fprintf(out, " -d , --db-file-output \n"); + fprintf(out, " set the db filename for indexing\n"); + fprintf(out, " -i, --index\n"); + fprintf(out, " index mails\n"); + fprintf(out, " -o , --output \n"); + fprintf(out, " set the result file, use stdout if unset\n"); + fprintf(out, " -n , --nb-mails-max \n"); + fprintf(out, " set the maximum number of mails to extract\n"); + fprintf(out, " -a , --default-search \n"); + fprintf(out, " set the default search field\n"); + } /*********************************************************************/ -void search_in_db(const char *search_name, const char *search_regexp_string, - FILE *db_file) { - char raw_line[BUFFER_SIZE]; - char current_mail_filename[BUFFER_SIZE]; - unsigned int current_position_in_mail; - char *name, *value; - regex_t regexp; - int already_written; +int ignore_entry(const char *name) { + return + strcmp(name, ".") == 0 || + strcmp(name, "..") == 0 || + (name[0] == '.' && name[1] != '/'); +} + +int is_a_leading_from_line(char *mbox_line) { + return + strncmp(mbox_line, "From ", 5) == 0 && + regexec(&global_leading_from_line_regexp, mbox_line, 0, 0, 0) == 0; +} + +int db_line_match_search(struct search_condition *condition, + int db_key, const char *db_value) { + + return + ( + (condition->db_key == db_key) + + || + + (condition->db_key == ID_PARTICIPANT && (db_key == ID_LEADING_LINE || + db_key == ID_FROM || + db_key == ID_TO)) + || + + (condition->db_key == ID_FROM && db_key == ID_LEADING_LINE) + + || + + (condition->db_key == ID_THREAD_ID && (db_key == ID_MAIL_ID || + db_key == ID_REFERENCE_ID)) + ) + + && + + regexec(&condition->db_value_regexp, db_value, 0, 0, 0) == 0; +} + +void update_body_hits(char *mail_filename, int position_in_mail, + int nb_search_conditions, struct search_condition *search_conditions, + int nb_body_conditions, + int *hits) { + FILE *mail_file; + int header, n; + char raw_mbox_line[BUFFER_SIZE]; + int nb_body_hits; + + nb_body_hits = 0; + + header = 1; + mail_file = safe_fopen(mail_filename, "r", "mbox for body scan"); + + fseek(mail_file, position_in_mail, SEEK_SET); + + if(fgets(raw_mbox_line, BUFFER_SIZE, mail_file)) { + while(nb_body_hits < nb_body_conditions) { + if(raw_mbox_line[0] == '\n') { header = 0; } + + if(!header) { + for(n = 0; n < nb_search_conditions; n++) { + if(search_conditions[n].db_key == ID_BODY && !hits[n]) { + hits[n] = + (regexec(&search_conditions[n].db_value_regexp, raw_mbox_line, 0, 0, 0) == 0); + if(hits[n]) { + nb_body_hits++; + } + } + } + } + + if(!fgets(raw_mbox_line, BUFFER_SIZE, mail_file) || + (is_a_leading_from_line(raw_mbox_line))) + break; + } + } + + fclose(mail_file); +} + +void extract_mail(const char *mail_filename, unsigned long int position_in_mail, + 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"); + /* fchmod(fileno(mail_file), 0x660); */ + fseek(mail_file, position_in_mail, SEEK_SET); + + if(fgets(raw_mbox_line, BUFFER_SIZE, mail_file)) { + 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)) + 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; + + memset(&tm, 0, sizeof(struct 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_extracted_mails, + int nb_search_conditions, + struct search_condition *search_conditions, + FILE *output_file) { + + 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; + const char *db_value; + int db_key; + int hits[MAX_NB_SEARCH_CONDITIONS]; + int nb_body_conditions, need_time; + time_t mail_time; + + int m, n; + + if(!global_quiet) { + printf("Searching in '%s' ... ", db_filename); + fflush(stdout); + } + + db_file = safe_fopen(db_filename, "r", "index file for search"); + + /* First, check the db file leading line integrity */ - if(regcomp(®exp, - search_regexp_string, - REG_ICASE)) { + if(fgets(raw_db_line, BUFFER_SIZE, db_file)) { + if(strncmp(raw_db_line, MYMAIL_DB_MAGIC_TOKEN, strlen(MYMAIL_DB_MAGIC_TOKEN))) { + fprintf(stderr, + "mymail: Header line in '%s' does not match the mymail db format.\n", + db_filename); + exit(EXIT_FAILURE); + } + } else { fprintf(stderr, - "mymail: Syntax error in regexp \"%s\" for field \"%s\".\n", - search_regexp_string, - search_name); + "mymail: Cannot read the header line in '%s'.\n", + db_filename); exit(EXIT_FAILURE); } + /* Then parse the said db file */ + current_position_in_mail = 0; - already_written = 0; - - while(fgets(raw_line, BUFFER_SIZE, db_file)) { - name = raw_line; - value = segment_next_field(raw_line); - - /* printf("LINE [%s] %s", name, value); */ - - if(strcmp("mail", name) == 0) { - char *position_in_file_string = value; - char *mail_filename = segment_next_field(value); - current_position_in_mail = atoi(position_in_file_string); - strcpy(current_mail_filename, mail_filename); - remove_eof(current_mail_filename); - /* printf("READING [%s]\n", current_mail_filename); */ - already_written = 0; - } else if(!already_written) { - if(strcmp(search_name, name) == 0 && regexec(®exp, value, 0, 0, 0) == 0) { - FILE *mail_file; - /* printf("%s:%u\n", current_mail_filename, current_position_in_mail); */ - mail_file = fopen(current_mail_filename, "r"); - if(!mail_file) { - fprintf(stderr, "mymail: Can not open `%s'.\n", current_mail_filename); - exit(EXIT_FAILURE); - } - fseek(mail_file, current_position_in_mail, SEEK_SET); - if(fgets(raw_line, BUFFER_SIZE, mail_file)) { - printf("%s", raw_line); - while(fgets(raw_line, BUFFER_SIZE, mail_file) && - strncmp(raw_line, "From ", 5)) { - printf("%s", raw_line); - } + + for(n = 0; n < nb_search_conditions; n++) { hits[n] = 0; } + + nb_body_conditions = 0; + need_time = global_discard_mail_from_the_future; + mail_time = 0; + + for(n = 0; n < nb_search_conditions; n++) { + 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(nb_extracted_mails < global_nb_mails_max && + fgets(raw_db_line, BUFFER_SIZE, db_file)) { + + /* Removes the CR */ + char *s = raw_db_line; + while(*s && *s != '\n') { s++; } + *s = '\0'; + + 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] && + (!global_discard_mail_from_the_future || mail_time < global_current_time + 3600 * 24) && + 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; } + db_value = parse_token(position_in_file_string, TOKEN_BUFFER_SIZE, ' ', db_value); + strncpy(current_mail_filename, db_value, PATH_MAX + 1); + current_position_in_mail = atol(position_in_file_string); + } + + else { + 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; } - fclose(mail_file); - already_written = 1; + } + + for(n = 0; n < nb_search_conditions; n++) { + hits[n] |= db_line_match_search(&search_conditions[n], + db_key, db_value); + } + + if(need_time) { + update_time(db_key, db_value, &mail_time); } } } + + /* printf("global_discard_mail_from_the_future = %d\n", + global_discard_mail_from_the_future); */ + + if(nb_extracted_mails < global_nb_mails_max && + current_mail_filename[0] && + (!global_discard_mail_from_the_future || mail_time < global_current_time + 3600 * 24) && + 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++; + } + + fclose(db_file); + + if(!global_quiet) { + printf("done.\n"); + fflush(stdout); + } + + return nb_extracted_mails; } +int recursive_search_in_db(const char *entry_name, regex_t *db_filename_regexp, + int nb_extracted_mails, + int nb_search_conditions, + struct search_condition *search_conditions, + FILE *output_file) { + DIR *dir; + struct dirent *dir_e; + struct stat sb; + char subname[PATH_MAX + 1]; + + if(lstat(entry_name, &sb) != 0) { + fprintf(stderr, + "mymail: Cannot stat \"%s\": %s\n", + entry_name, + strerror(errno)); + exit(EXIT_FAILURE); + } + + /* printf("recursive_search_in_db %s\n", entry_name); */ + + dir = opendir(entry_name); + + if(dir) { + while((dir_e = readdir(dir)) && + nb_extracted_mails < global_nb_mails_max) { + if(!ignore_entry(dir_e->d_name)) { + snprintf(subname, PATH_MAX, "%s/%s", entry_name, dir_e->d_name); + nb_extracted_mails = recursive_search_in_db(subname, db_filename_regexp, + nb_extracted_mails, + nb_search_conditions, search_conditions, + output_file); + } + } + closedir(dir); + } + + else { + const char *s = entry_name, *filename = entry_name; + while(*s) { if(*s == '/') { filename = s+1; } s++; } + + if(regexec(db_filename_regexp, filename, 0, 0, 0) == 0) { + nb_extracted_mails = + search_in_db(entry_name, + nb_extracted_mails, + nb_search_conditions, search_conditions, output_file); + } + } + + return nb_extracted_mails; +} /*********************************************************************/ -void read_file(const char *input_filename, - int nb_fields_to_parse, struct parsable_field *fields_to_parse, - FILE *db_file) { - char raw_line[BUFFER_SIZE]; +void index_one_mbox_line(unsigned int nb_fields_to_parse, + struct parsable_field *fields_to_parse, + char *raw_mbox_line, FILE *db_file) { + regmatch_t matches; + unsigned int f; + for(f = 0; f < nb_fields_to_parse; f++) { + if(regexec(&fields_to_parse[f].regexp, raw_mbox_line, 1, &matches, 0) == 0) { + fprintf(db_file, "%s %s\n", + field_keys[fields_to_parse[f].id], + raw_mbox_line + matches.rm_eo); + } + } +} + +void index_mbox(const char *mbox_filename, + int nb_fields_to_parse, struct parsable_field *fields_to_parse, + FILE *db_file) { + char raw_mbox_line[BUFFER_SIZE], full_line[BUFFER_SIZE]; + char *end_of_full_line; FILE *file; int in_header, new_header; - unsigned int position_in_file; + unsigned long int position_in_file; - file = fopen(input_filename, "r"); - - if(!file) { - fprintf(stderr, "mymail: Can not open `%s'.\n", input_filename); - exit(EXIT_FAILURE); - } + file = safe_fopen(mbox_filename, "r", "mbox for indexing"); in_header = 0; new_header = 0; position_in_file = 0; + end_of_full_line = 0; + full_line[0] = '\0'; - while(fgets(raw_line, BUFFER_SIZE, file)) { - if(strncmp(raw_line, "From ", 5) == 0) { + 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:%u.\n", - input_filename, position_in_file); - fprintf(stderr, "%s", raw_line); - exit(EXIT_FAILURE); + "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(strncmp(raw_line, "\n", 1) == 0) { - if(in_header) { in_header = 0; } + } else if(raw_mbox_line[0] == '\n') { + 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) { */ - /* printf("LINE.H %s", raw_line); */ - /* } else { */ - /* printf("LINE.B %s", raw_line); */ - /* } */ - if(in_header) { - int f; - regmatch_t matches; if(new_header) { - fprintf(db_file, "mail %u %s\n", position_in_file, input_filename); + fprintf(db_file, "mail %lu %s\n", position_in_file, mbox_filename); new_header = 0; } - for(f = 0; f < nb_fields_to_parse; f++) { - if(regexec(&fields_to_parse[f].regexp, raw_line, 1, &matches, 0) == 0) { - fprintf(db_file, "%s %s", - fields_to_parse[f].name, - raw_line + matches.rm_eo); + + 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++) = ' '; + strcpy(end_of_full_line, start); + while(*end_of_full_line && *end_of_full_line != '\n') { + end_of_full_line++; } + *end_of_full_line = '\0'; } + + else { + /* 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); + } + + end_of_full_line = full_line; + strcpy(end_of_full_line, raw_mbox_line); + while(*end_of_full_line && *end_of_full_line != '\n') { + end_of_full_line++; + } + *end_of_full_line = '\0'; + } + } - position_in_file += strlen(raw_line); + position_in_file += strlen(raw_mbox_line); } fclose(file); } -int ignore_entry(const char *name) { - return - strcmp(name, ".") == 0 || - strcmp(name, "..") == 0 || - (name[0] == '.' && name[1] != '/'); -} - -void process_entry(const char *dir_name, - int nb_fields_to_parse, struct parsable_field *fields_to_parse, - FILE *db_file) { +void recursive_index_mbox(FILE *db_file, + const char *entry_name, regex_t *mbox_filename_regexp, + int nb_fields_to_parse, struct parsable_field *fields_to_parse) { DIR *dir; struct dirent *dir_e; struct stat sb; char subname[PATH_MAX + 1]; - if(lstat(dir_name, &sb) != 0) { + if(lstat(entry_name, &sb) != 0) { fprintf(stderr, - "mymail: Can not stat \"%s\": %s\n", - dir_name, + "mymail: Cannot stat \"%s\": %s\n", + entry_name, strerror(errno)); exit(EXIT_FAILURE); - } else { - } - - if(S_ISLNK(sb.st_mode)) { - return; } - dir = opendir(dir_name); + dir = opendir(entry_name); if(dir) { - printf("Processing directory '%s'.\n", dir_name); while((dir_e = readdir(dir))) { if(!ignore_entry(dir_e->d_name)) { - snprintf(subname, PATH_MAX, "%s/%s", dir_name, dir_e->d_name); - process_entry(subname, nb_fields_to_parse, fields_to_parse, db_file); + snprintf(subname, PATH_MAX, "%s/%s", entry_name, dir_e->d_name); + recursive_index_mbox(db_file, subname, mbox_filename_regexp, + nb_fields_to_parse, fields_to_parse); } } closedir(dir); } else { - if(S_ISREG(sb.st_mode)) { - /* printf("Processing regular file '%s'.\n", dir_name); */ - read_file(dir_name, nb_fields_to_parse, fields_to_parse, db_file); + const char *s = entry_name, *filename = s; + while(*s) { if(*s == '/') { filename = s+1; }; s++; } + if(!mbox_filename_regexp || regexec(mbox_filename_regexp, filename, 0, 0, 0) == 0) { + index_mbox(entry_name, nb_fields_to_parse, fields_to_parse, db_file); } } } @@ -262,45 +802,312 @@ void process_entry(const char *dir_name, /* For long options that have no equivalent short option, use a non-character as a pseudo short option, starting with CHAR_MAX + 1. */ -enum -{ +enum { OPT_BASH_MODE = CHAR_MAX + 1 }; static struct option long_options[] = { { "help", no_argument, 0, 'h' }, - { "db-prefix", 1, 0, 'p' }, - { "search-pattern", 1, 0, 's' }, + { "version", no_argument, 0, 'v' }, + { "quiet", no_argument, 0, 'q' }, + { "use-leading-time", no_argument, 0, 't' }, + { "do-not-discard-mails-from-the-future", no_argument, 0, 'f' }, + { "db-file-output", 1, 0, 'd' }, + { "db-pattern", 1, 0, 'p' }, + { "db-root", 1, 0, 'r' }, + { "db-list", 1, 0, 'l' }, + { "mbox-pattern", 1, 0, 'm' }, + { "search", 1, 0, 's' }, { "index", 0, 0, 'i' }, + { "output", 1, 0, 'o' }, + { "default-search", 1, 0, 'a' }, + { "nb-mails-max", 1, 0, 'n' }, { 0, 0, 0, 0 } }; -static struct parsable_field fields_to_parse[] = { - { - "from", - "^\\([Ff][Rr][Oo][Mm]:\\|From\\) *", - { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } - }, +struct time_criterion { + char *label; + int day_criterion; + int start_hour, end_hour; + int past_week_day; +}; + +/*********************************************************************/ + +static struct time_criterion time_criteria[] = { + + { "1h", 0, 1, -1, -1 }, + { "2h", 0, 2, -1, -1 }, + { "4h", 0, 4, -1, -1 }, + { "8h", 0, 8, -1, -1 }, + { "24h", 0, 24, -1, -1 }, + { "48h", 0, 48, -1, -1 }, + { "week", 0, 24 * 7, -1, -1 }, + { "2weeks", 0, 24 * 14, -1, -1 }, + { "month", 0, 24 * 31, -1, -1 }, + { "semester", 0, 24 * 185, -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 }, - { - "dest", - "^\\([Tt][Oo]\\|[Cc][Cc]\\|[Bb][Cc][Cc]\\): *", - { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } - }, }; +/*********************************************************************/ + +time_t time_for_past_day(int day) { + struct tm *tm; + int delta_day; + tm = localtime(&global_current_time); + if(day > 0) { + delta_day = (7 + tm->tm_wday - day) % 7; + } else { + delta_day = - day; + } + return global_current_time - (delta_day * 3600 * 24 + tm->tm_sec + 60 * tm->tm_min + 3600 * tm->tm_hour); +} + +void init_condition(struct search_condition *condition, const char *full_string, + const char *default_search_field) { + char full_search_field[TOKEN_BUFFER_SIZE], *search_field; + unsigned int k, m; + const char *string; + struct alias_node *a; + + for(a = global_alias_list; a; a = a->next) { + if(strcmp(full_string, a->alias) == 0) { + full_string = a->value; + break; + } + } + + string = parse_token(full_search_field, TOKEN_BUFFER_SIZE, ' ', full_string); + search_field = full_search_field; + + if(search_field[0] == '!') { + search_field++; + condition->negation = 1; + } else { + condition->negation = 0; + } + + condition->db_key = -1; + + /* Time condition */ + + for(k = 0; k < sizeof(time_criteria) / sizeof(struct time_criterion); k++) { + if(strcmp(time_criteria[k].label, search_field) == 0) { + 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 = global_current_time - 3600 * time_criteria[k].start_hour; + if(time_criteria[k].end_hour >= 0) { + condition->time_stop = global_current_time - 3600 * time_criteria[k].end_hour; + } else { + condition->time_stop = 0; + } + } + + break; + } + } + + if(condition->db_key == -1) { + + /* No time condition matched, look for the search fields */ + + for(m = 0; (m < MAX_ID) && condition->db_key == -1; m++) { + if(strncmp(field_keys[m], search_field, strlen(search_field)) == 0) { + condition->db_key = m; + } + } + + /* None match, if there is a default search field, re-run the search with it */ + + if(condition->db_key == -1) { + if(default_search_field) { + 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->db_key = m; + } + } + string = full_string; + if(string[0] == '!') { string++; } + } + } + + if(condition->db_key == -1) { + fprintf(stderr, + "mymail: Syntax error in field key \"%s\".\n", + search_field); + exit(EXIT_FAILURE); + } + + 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->db_key]); + exit(EXIT_FAILURE); + } + } +} + +void free_condition(struct search_condition *condition) { + if(condition->db_key != ID_TIME_INTERVAL) { + regfree(&condition->db_value_regexp); + } +} + +const char *eat_space(const char *s) { + while(*s == ' ' || *s == '\t') { s++; } + return s; +} + +void read_rc_file(const char *rc_filename) { + char raw_line[BUFFER_SIZE]; + char command[TOKEN_BUFFER_SIZE], tmp_token[TOKEN_BUFFER_SIZE]; + + FILE *rc_file; + int line_number; + const char *s; + char *t; + + rc_file = fopen(rc_filename, "r"); + + if(rc_file) { + line_number = 1; + while(fgets(raw_line, BUFFER_SIZE, rc_file)) { + t = raw_line; + while(*t) { if(*t == '\n') { *t = '\0'; }; t++; } + + s = raw_line; + s = eat_space(s); + + if(*s && *s != '#') { + s = parse_token(command, TOKEN_BUFFER_SIZE, ' ', s); + + if(strcmp(command, "alias") == 0) { + struct alias_node *a = safe_malloc(sizeof(struct alias_node)); + a->next = global_alias_list; + global_alias_list = a; + if(s) { + s = eat_space(s); + s = parse_token(tmp_token, TOKEN_BUFFER_SIZE, '=', s); + a->alias = strdup(tmp_token); + if(s) { + s = eat_space(s); + a->value = strdup(s); + } else { + fprintf(stderr, "%s:%d syntax error, missing alias value.\n", + rc_filename, + line_number); + exit(EXIT_FAILURE); + } + } else { + fprintf(stderr, "%s:%d syntax error, missing alias key.\n", + rc_filename, + line_number); + exit(EXIT_FAILURE); + } + } else { + fprintf(stderr, "%s:%d syntax error, unknown command '%s'.\n", + rc_filename, + line_number, + command); + exit(EXIT_FAILURE); + } + } + + line_number++; + } + fclose(rc_file); + } +} + +/*********************************************************************/ +/*********************************************************************/ +/*********************************************************************/ + int main(int argc, char **argv) { + char *db_filename = 0; + char *db_filename_regexp_string = 0; + char *db_root_path = 0; + char *db_filename_list = 0; + char *mbox_filename_regexp_string = 0; + char *default_search_field; + char output_filename[PATH_MAX + 1]; + char rc_filename[PATH_MAX + 1]; + int action_index = 0; int error = 0, show_help = 0; - const int nb_fields_to_parse = sizeof(fields_to_parse) / sizeof(struct parsable_field); + const unsigned int nb_fields_to_parse = + sizeof(fields_to_parse) / sizeof(struct parsable_field); char c; - int f; + unsigned int f, n; + unsigned int nb_search_conditions; + struct search_condition search_conditions[MAX_NB_SEARCH_CONDITIONS]; + struct alias_node *a, *b; + + /* Group and others have no access */ + umask(S_IRWXG | S_IRWXO); + + if(regcomp(&global_leading_from_line_regexp, LEADING_FROM_LINE_REGEXP_STRING, 0)) { + fprintf(stderr, + "mymail: Cannot compile leading \"from\" line regexp. That is strange.\n"); + exit(EXIT_FAILURE); + } + + if(getenv("MYMAILRC")) { + sprintf(rc_filename, "%s", getenv("MYMAILRC")); + } else if(getenv("HOME")) { + sprintf(rc_filename, "%s/.mymailrc", getenv("HOME")); + } else { + rc_filename[0] = '\0'; + } + + global_alias_list = 0; + global_quiet = 0; + global_use_leading_time = 0; + global_nb_mails_max = 250; + global_discard_mail_from_the_future = 1; + global_current_time = time(0); + + default_search_field = 0; + strncpy(output_filename, "", PATH_MAX); - action_index = 0; - search_pattern = 0; + if(rc_filename[0]) { + read_rc_file(rc_filename); + } + + /* + { + #warning Test code added on 2013 May 02 11:17:01 + struct alias_node *a; + for(a = global_alias_list; a; a = a->next) { + printf ("ALIAS [%s] [%s]\n", a->alias, a->value); + } + } + */ setlocale(LC_ALL, ""); - while ((c = getopt_long(argc, argv, "hip:s:", + nb_search_conditions = 0; + + while ((c = getopt_long(argc, argv, "hvqtfip:s:d:r:l:o:a:m:", long_options, NULL)) != -1) { switch(c) { @@ -309,21 +1116,85 @@ int main(int argc, char **argv) { show_help = 1; break; + case 'v': + print_version(stdout); + break; + + case 'q': + global_quiet = 1; + break; + + case 't': + global_use_leading_time = 1; + break; + + case 'f': + global_discard_mail_from_the_future = 0; + break; + case 'i': action_index = 1; break; - case 'p': + case 'd': + if(db_filename) { + fprintf(stderr, "mymail: Can not set the db filename twice.\n"); + exit(EXIT_FAILURE); + } db_filename = strdup(optarg); break; + case 'p': + if(db_filename_regexp_string) { + fprintf(stderr, "mymail: Can not set the db filename pattern twice.\n"); + exit(EXIT_FAILURE); + } + db_filename_regexp_string = strdup(optarg); + break; + + case 'm': + if(mbox_filename_regexp_string) { + fprintf(stderr, "mymail: Can not set the mbox filename pattern twice.\n"); + exit(EXIT_FAILURE); + } + mbox_filename_regexp_string = strdup(optarg); + break; + + case 'o': + strncpy(output_filename, optarg, PATH_MAX); + break; + + case 'r': + if(db_root_path) { + fprintf(stderr, "mymail: Can not set the db root path twice.\n"); + exit(EXIT_FAILURE); + } + db_root_path = strdup(optarg); + break; + + case 'l': + if(db_filename_list) { + fprintf(stderr, "mymail: Can not set the db filename list twice.\n"); + exit(EXIT_FAILURE); + } + db_filename_list = strdup(optarg); + break; + case 's': - if(search_pattern) { - fprintf(stderr, - "mymail: Search pattern already defined.\n"); + if(nb_search_conditions == MAX_NB_SEARCH_CONDITIONS) { + fprintf(stderr, "mymail: Too many search patterns.\n"); exit(EXIT_FAILURE); } - search_pattern = strdup(optarg); + init_condition(&search_conditions[nb_search_conditions], optarg, default_search_field); + nb_search_conditions++; + break; + + case 'a': + default_search_field = optarg; + break; + + case 'n': + global_nb_mails_max = atoi(optarg); break; default: @@ -332,89 +1203,202 @@ int main(int argc, char **argv) { } } - if(!db_filename) { - db_filename = strdup("/tmp/mymail"); - } - if(error) { - usage(stderr); + print_usage(stderr); exit(EXIT_FAILURE); } if(show_help) { - usage(stdout); + print_usage(stdout); exit(EXIT_SUCCESS); } + /* Set all the values that may defined in the arguments, through + environment variables, or hard-coded */ + + db_filename = default_value(db_filename, + "MYMAIL_DB_FILE", + "mymail.db"); + + db_filename_regexp_string = default_value(db_filename_regexp_string, + "MYMAIL_DB_FILE", + "\\.db$"); + + db_root_path = default_value(db_root_path, + "MYMAIL_DB_ROOT", + 0); + + db_filename_list = default_value(db_filename_list, + "MYMAIL_DB_LIST", + 0); + + mbox_filename_regexp_string = default_value(mbox_filename_regexp_string, + "MYMAIL_MBOX_PATTERN", + 0); + + /* mbox indexing */ + if(action_index) { - FILE *db_file = fopen(db_filename, "w"); - if(!db_file) { - fprintf(stderr, - "mymail: Can not open \"%s\" for writing: %s\n", - db_filename, - strerror(errno)); - exit(EXIT_FAILURE); + FILE *db_file; + regex_t mbox_filename_regexp_static; + regex_t *mbox_filename_regexp; + + if(mbox_filename_regexp_string) { + if(regcomp(&mbox_filename_regexp_static, + mbox_filename_regexp_string, + 0)) { + fprintf(stderr, + "mymail: Syntax error in regexp \"%s\".\n", + mbox_filename_regexp_string); + exit(EXIT_FAILURE); + } + mbox_filename_regexp = &mbox_filename_regexp_static; + } else { + mbox_filename_regexp = 0; } + db_file = safe_fopen(db_filename, "w", "index file for indexing"); + for(f = 0; f < nb_fields_to_parse; f++) { if(regcomp(&fields_to_parse[f].regexp, fields_to_parse[f].regexp_string, - REG_ICASE)) { + fields_to_parse[f].cflags)) { fprintf(stderr, "mymail: Syntax error in regexp \"%s\" for field \"%s\".\n", fields_to_parse[f].regexp_string, - fields_to_parse[f].name); + field_keys[fields_to_parse[f].id]); exit(EXIT_FAILURE); } } + fprintf(db_file, + "%s version_%s format_%d raw\n", + MYMAIL_DB_MAGIC_TOKEN, + MYMAIL_VERSION, + MYMAIL_DB_FORMAT_VERSION); + while(optind < argc) { - process_entry(argv[optind], - nb_fields_to_parse, fields_to_parse, db_file); + recursive_index_mbox(db_file, + argv[optind], mbox_filename_regexp, + nb_fields_to_parse, fields_to_parse); optind++; } + fflush(db_file); fclose(db_file); + if(mbox_filename_regexp) { + regfree(mbox_filename_regexp); + } + for(f = 0; f < nb_fields_to_parse; f++) { regfree(&fields_to_parse[f].regexp); } } + /* Mail search */ + else { - if(search_pattern) { - FILE *db_file; - char *search_name; - char *search_regexp_string; - search_name = search_pattern; - search_regexp_string = segment_next_field(search_pattern); - if(!*search_regexp_string) { - fprintf(stderr, - "Syntax error in the search pattern.\n"); - exit(EXIT_FAILURE); + + FILE *output_file; + int nb_extracted_mails = 0; + + if(output_filename[0]) { + output_file = safe_fopen(output_filename, "w", "result mbox"); + } else { + output_file = stdout; + global_quiet = 1; + } + + if(nb_search_conditions > 0) { + + /* Recursive search if db_root_path is set */ + + if(db_root_path) { + regex_t db_filename_regexp; + if(regcomp(&db_filename_regexp, + db_filename_regexp_string, + 0)) { + fprintf(stderr, + "mymail: Syntax error in regexp \"%s\".\n", + db_filename_regexp_string); + exit(EXIT_FAILURE); + } + + nb_extracted_mails = recursive_search_in_db(db_root_path, &db_filename_regexp, + nb_extracted_mails, + nb_search_conditions, search_conditions, + output_file); + + regfree(&db_filename_regexp); } - /* printf("Starting search in %s for field \"%s\" matching \"%s\".\n", */ - /* db_filename, */ - /* search_name, */ - /* search_regexp_string); */ + /* Search in all db files listed in db_filename_list */ - db_file = fopen(db_filename, "r"); + if(db_filename_list) { + char db_filename[PATH_MAX + 1]; + const char *s; - if(!db_file) { - fprintf(stderr, - "mymail: Can not open \"%s\" for reading: %s\n", - db_filename, - strerror(errno)); - exit(EXIT_FAILURE); + s = db_filename_list; + + while(*s) { + s = parse_token(db_filename, PATH_MAX + 1, ';', s); + + if(db_filename[0]) { + nb_extracted_mails = + search_in_db(db_filename, + nb_extracted_mails, + nb_search_conditions, search_conditions, output_file); + } + } } - search_in_db(search_name, search_regexp_string, db_file); + /* Search in all db files listed in the command arguments */ - fclose(db_file); - free(search_pattern); + while(optind < argc) { + nb_extracted_mails = + search_in_db(argv[optind], + nb_extracted_mails, + nb_search_conditions, search_conditions, output_file); + optind++; + } + } + + if(!global_quiet) { + if(nb_extracted_mails > 0) { + printf("Found %d matching mails.\n", nb_extracted_mails); + } else { + printf("No matching mail found.\n"); + } + } + + fflush(output_file); + + if(output_file != stdout) { + fclose(output_file); } } + for(n = 0; n < nb_search_conditions; n++) { + free_condition(&search_conditions[n]); + } + + a = global_alias_list; + while(a) { + b = a->next; + free(a->alias); + free(a->value); + free(a); + a = b; + } + + free(db_filename); + free(db_filename_regexp_string); + free(db_root_path); + free(db_filename_list); + free(mbox_filename_regexp_string); + + regfree(&global_leading_from_line_regexp); + exit(EXIT_SUCCESS); }