X-Git-Url: https://fleuret.org/cgi-bin/gitweb/gitweb.cgi?a=blobdiff_plain;f=mymail.c;h=dcb44bff6944428852ca26ce7f189c6f87a198f6;hb=e40b536f28d3d6af2e49a48dc830f15a2fd88ab0;hp=25121a4f61432ca71c049ed1bf5221f95a4e0719;hpb=5166266bf624c0a4d94d7751bb819c26cc1cc22d;p=mymail.git diff --git a/mymail.c b/mymail.c index 25121a4..dcb44bf 100644 --- a/mymail.c +++ b/mymail.c @@ -19,6 +19,18 @@ * */ +/* + + This command is a dumb mail indexer. It can either (1) scan + directories containing mbox files, and create a db file containing + for each mail a list of fields computed from the header, or (2) + read such a db file and get all the mails matching regexp-defined + conditions on the fields. + + It is low-tech, simple, light and fast. + +*/ + #define _GNU_SOURCE #include @@ -32,21 +44,54 @@ #include #include +#define MYMAIL_DB_MAGIC_TOKEN "mymail_index_file" #define VERSION "0.1" -#define BUFFER_SIZE 16384 +#define MAX_NB_SEARCH_PATTERNS 10 + +#define BUFFER_SIZE 65536 + +enum { + ID_MAIL, + ID_FROM, + ID_DEST, + ID_SUBJECT, + ID_FROMDEST, + MAX_ID +}; + +static char *field_names[] = { + "mail", + "from", + "dest", + "subj", + "fromdest" +}; struct parsable_field { - char *name; + int id; char *regexp_string; regex_t regexp; }; char *db_filename; -char *search_pattern; +char *db_root_path; +int paranoid; int action_index; +char *segment_next_field(char *current) { + while(*current && *current != ' ') current++; + *current = '\0'; current++; + while(*current && *current == ' ') current++; + return current; +} + +void remove_eof(char *c) { + while(*c && *c != '\n' && *c != '\r') c++; + *c = '\0'; +} + /********************************************************************/ /* malloc with error checking. */ @@ -55,7 +100,7 @@ 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; @@ -63,121 +108,333 @@ void *safe_malloc(size_t n) { /*********************************************************************/ -void usage(FILE *out) { +void print_version(FILE *out) { fprintf(out, "mymail version %s (%s)\n", 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, " -i, --index\n"); + fprintf(out, " index mails\n"); + fprintf(out, " -s , --search \n"); + fprintf(out, " search for matching mails in the data-base file\n"); + fprintf(out, " -d , --db-file \n"); + fprintf(out, " set the data-base file\n"); + fprintf(out, " -r , --db-root \n"); + fprintf(out, " set the data-base root path for recursive search\n"); +} + +/*********************************************************************/ + +int ignore_entry(const char *name) { + return + /* strcmp(name, ".") == 0 || */ + /* strcmp(name, "..") == 0 || */ + (name[0] == '.' && name[1] != '/'); +} + +int mbox_line_match_search(int search_id, regex_t *search_regexp, + int mbox_id, char *mbox_value) { + return search_id == mbox_id && regexec(search_regexp, mbox_value, 0, 0, 0) == 0; +} + +void search_in_db(int nb_search_patterns, + int *search_ids, char **search_regexp_strings, + FILE *db_file) { + int hits[MAX_NB_SEARCH_PATTERNS]; + char raw_db_line[BUFFER_SIZE]; + char raw_mbox_line[BUFFER_SIZE]; + char current_mail_filename[PATH_MAX + 1]; + unsigned long int current_position_in_mail; + char *mbox_name, *mbox_value; + int mbox_id; + regex_t search_regexps[MAX_NB_SEARCH_PATTERNS]; + int already_written, m, n; + + for(n = 0; n < nb_search_patterns; n++) { + if(regcomp(&search_regexps[n], + search_regexp_strings[n], + REG_ICASE)) { + fprintf(stderr, + "mymail: Syntax error in regexp \"%s\" for field \"%s\".\n", + search_regexp_strings[n], + field_names[search_ids[n]]); + exit(EXIT_FAILURE); + } + } + + current_position_in_mail = 0; + already_written = 0; + + for(n = 0; n < nb_search_patterns; n++) { hits[n] = 0; } + + while(fgets(raw_db_line, BUFFER_SIZE, db_file)) { + mbox_name = raw_db_line; + mbox_value = segment_next_field(raw_db_line); + + if(strcmp("mail", mbox_name) == 0) { + char *position_in_file_string; + char *mail_filename; + + for(n = 0; n < nb_search_patterns && hits[n]; n++); + + if(n == nb_search_patterns) { + FILE *mail_file; + + mail_file = fopen(current_mail_filename, "r"); + + if(!mail_file) { + fprintf(stderr, "mymail: Cannot open mbox '%s'.\n", current_mail_filename); + exit(EXIT_FAILURE); + } + + fseek(mail_file, current_position_in_mail, SEEK_SET); + + if(fgets(raw_mbox_line, BUFFER_SIZE, mail_file)) { + printf("%s", raw_mbox_line); + while(fgets(raw_mbox_line, BUFFER_SIZE, mail_file) && + strncmp(raw_mbox_line, "From ", 5)) { + printf("%s", raw_mbox_line); + } + } + + fclose(mail_file); + } + + for(n = 0; n < nb_search_patterns; n++) { hits[n] = 0; } + + position_in_file_string = mbox_value; + mail_filename = segment_next_field(mbox_value); + current_position_in_mail = atol(position_in_file_string); + strcpy(current_mail_filename, mail_filename); + + remove_eof(current_mail_filename); + already_written = 0; + } + + else { + mbox_id = -1; + for(m = 0; (m < MAX_ID) && mbox_id == -1; m++) { + if(strncmp(field_names[m], mbox_name, strlen(mbox_name)) == 0) { + mbox_id = m; + } + } + for(n = 0; n < nb_search_patterns; n++) { + hits[n] |= mbox_line_match_search(search_ids[n], &search_regexps[n], + mbox_id, mbox_value); + } + } + } + + for(n = 0; n < nb_search_patterns; n++) { + regfree(&search_regexps[n]); + } +} + +void recursive_search_in_db(const char *entry_name, + int nb_search_patterns, + int *search_ids, char **search_regexp_strings) { + DIR *dir; + struct dirent *dir_e; + struct stat sb; + char raw_db_line[BUFFER_SIZE]; + 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); + } + + dir = opendir(entry_name); + + if(dir) { + while((dir_e = readdir(dir))) { + if(!ignore_entry(dir_e->d_name)) { + snprintf(subname, PATH_MAX, "%s/%s", entry_name, dir_e->d_name); + recursive_search_in_db(subname, + nb_search_patterns, + search_ids, search_regexp_strings); + } + } + closedir(dir); + } else { + const char *s = entry_name, *filename = entry_name; + while(*s) { if(*s == '/') { filename = s+1; } s++; } + + if(strcmp(filename, db_filename) == 0) { + FILE *db_file = fopen(entry_name, "r"); + + if(!db_file) { + fprintf(stderr, + "mymail: Cannot open \"%s\" for reading: %s\n", + db_filename, + strerror(errno)); + exit(EXIT_FAILURE); + } + + 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", + entry_name); + exit(EXIT_FAILURE); + } + } else { + fprintf(stderr, + "mymail: Cannot read the header line in '%s'.\n", + entry_name); + exit(EXIT_FAILURE); + } + + search_in_db(nb_search_patterns, search_ids, search_regexp_strings, + db_file); + + fclose(db_file); + } + } +} + +/*********************************************************************/ + +void index_one_mbox_line(int nb_fields_to_parse, struct parsable_field *fields_to_parse, + char *raw_mbox_line, FILE *db_file) { + regmatch_t matches; + 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_names[fields_to_parse[f].id], + raw_mbox_line + matches.rm_eo); + } + } } -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_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"); + file = fopen(mbox_filename, "r"); if(!file) { - fprintf(stderr, "mymail: Can not open `%s'.\n", input_filename); - exit(EXIT_FAILURE); + fprintf(stderr, "mymail: Cannot open '%s'.\n", mbox_filename); + if(paranoid) { exit(EXIT_FAILURE); } + return; } 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(strncmp(raw_mbox_line, "From ", 5) == 0) { 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); + if(paranoid) { exit(EXIT_FAILURE); } } in_header = 1; new_header = 1; - } else if(strncmp(raw_line, "\n", 1) == 0) { + } else if(strncmp(raw_mbox_line, "\n", 1) == 0) { if(in_header) { in_header = 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 %s:%d\n", input_filename, position_in_file); + 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') { + 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 { + /* + 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); + } + */ + + if(full_line[0]) { + 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, + 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, 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); - } + index_mbox(entry_name, nb_fields_to_parse, fields_to_parse, db_file); } } @@ -185,45 +442,65 @@ 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' }, + { "db-file", 1, 0, 'd' }, + { "db-root", 1, 0, 'r' }, + { "search", 1, 0, 's' }, { "index", 0, 0, 'i' }, { 0, 0, 0, 0 } }; static struct parsable_field fields_to_parse[] = { { - "from", + ID_FROM, "^\\([Ff][Rr][Oo][Mm]:\\|From\\) *", { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } }, { - "dest", + ID_DEST, "^\\([Tt][Oo]\\|[Cc][Cc]\\|[Bb][Cc][Cc]\\): *", { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } }, + + { + ID_SUBJECT, + "^[Ss][Uu][Bb][Jj][Ee][Cc][Tt]: *", + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } + }, + }; +/*********************************************************************/ + int main(int argc, char **argv) { int error = 0, show_help = 0; const int nb_fields_to_parse = sizeof(fields_to_parse) / sizeof(struct parsable_field); char c; int f; + int nb_search_patterns; + char *search_pattern[MAX_NB_SEARCH_PATTERNS]; + + /* for(f = 0; f < argc; f++) { */ + /* printf("arg %d \"%s\"\n", f, argv[f]); */ + /* } */ + paranoid = 0; action_index = 0; - search_pattern = 0; + db_filename = 0; + db_root_path = 0; setlocale(LC_ALL, ""); - while ((c = getopt_long(argc, argv, "hip:s:", + nb_search_patterns = 0; + + while ((c = getopt_long(argc, argv, "hvip:s:d:r:", long_options, NULL)) != -1) { switch(c) { @@ -232,21 +509,28 @@ int main(int argc, char **argv) { show_help = 1; break; + case 'v': + print_version(stdout); + break; + case 'i': action_index = 1; break; - case 'p': + case 'd': db_filename = strdup(optarg); break; + case 'r': + db_root_path = strdup(optarg); + break; + case 's': - if(search_pattern) { - fprintf(stderr, - "mymail: Search pattern already defined.\n"); + if(nb_search_patterns == MAX_NB_SEARCH_PATTERNS) { + fprintf(stderr, "mymail: Too many search patterns.\n"); exit(EXIT_FAILURE); } - search_pattern = strdup(optarg); + search_pattern[nb_search_patterns++] = strdup(optarg); break; default: @@ -256,24 +540,48 @@ int main(int argc, char **argv) { } if(!db_filename) { - db_filename = strdup("/tmp/mymail"); + char *default_db_filename = getenv("MYMAIL_DB_FILE"); + + if(!default_db_filename) { + default_db_filename = "mymail.db"; + } + + db_filename = strdup(default_db_filename); } + if(!db_root_path) { + char *default_db_root_path = getenv("MYMAIL_DB_ROOT"); + + if(default_db_root_path) { + db_root_path = strdup(default_db_root_path); + } + } + + if(!db_root_path) { + fprintf(stderr, + "mymail: db root path is not set\n"); + exit(EXIT_FAILURE); + } + + if(error) { - usage(stderr); + print_usage(stderr); exit(EXIT_FAILURE); } if(show_help) { - usage(stdout); + print_usage(stdout); exit(EXIT_SUCCESS); } if(action_index) { - FILE *db_file = fopen(db_filename, "w"); + FILE *db_file; + + db_file = fopen(db_filename, "w"); + if(!db_file) { fprintf(stderr, - "mymail: Can not open \"%s\" for writing: %s\n", + "mymail: Cannot open \"%s\" for writing: %s\n", db_filename, strerror(errno)); exit(EXIT_FAILURE); @@ -286,14 +594,17 @@ int main(int argc, char **argv) { fprintf(stderr, "mymail: Syntax error in regexp \"%s\" for field \"%s\".\n", fields_to_parse[f].regexp_string, - fields_to_parse[f].name); + field_names[fields_to_parse[f].id]); exit(EXIT_FAILURE); } } + fprintf(db_file, "%s version_%s raw version\n", MYMAIL_DB_MAGIC_TOKEN, VERSION); + while(optind < argc) { - process_entry(argv[optind], - nb_fields_to_parse, fields_to_parse, db_file); + recursive_index_mbox(db_file, + argv[optind], + nb_fields_to_parse, fields_to_parse); optind++; } @@ -305,26 +616,39 @@ int main(int argc, char **argv) { } else { - if(search_pattern) { - char *search_name; - char *search_regexp; - search_name = search_pattern; - search_regexp = search_pattern; - while(*search_regexp && *search_regexp != ' ') search_regexp++; - *search_regexp = '\0'; search_regexp++; - while(*search_regexp && *search_regexp == ' ') search_regexp++; - if(!*search_regexp) { + + if(nb_search_patterns > 0) { + int search_ids[MAX_NB_SEARCH_PATTERNS]; + char *search_regexp_strings[MAX_NB_SEARCH_PATTERNS]; + int m, n; + + for(n = 0; n < nb_search_patterns; n++) { + search_regexp_strings[n] = segment_next_field(search_pattern[n]); + search_ids[n] = -1; + for(m = 0; (m < MAX_ID) && search_ids[n] == -1; m++) { + if(strncmp(field_names[m], search_pattern[n], strlen(search_pattern[n])) == 0) { + search_ids[n] = m; + } + } + } + + if(!*search_regexp_strings) { fprintf(stderr, "Syntax error in the search pattern.\n"); exit(EXIT_FAILURE); } - printf("Starting search in %s for field \"%s\" matching \"%s\".\n", - db_filename, - search_name, - search_regexp); - free(search_pattern); + + recursive_search_in_db(db_root_path, + nb_search_patterns, search_ids, search_regexp_strings); + + for(n = 0; n < nb_search_patterns; n++) { + free(search_pattern[n]); + } } } + free(db_filename); + free(db_root_path); + exit(EXIT_SUCCESS); }