#include <sys/wait.h>
#include <dirent.h>
#include <time.h>
+#include <regex.h>
#include "intl.h"
#include "utils.h"
return NULL;
}
-gint get_wcs_len(const gchar *s)
+gint get_mbs_len(const gchar *s)
{
const gchar *p = s;
gint mb_len;
eliminate_parenthesis(str, '(', ')');
g_strstrip(str);
- while (!strncasecmp(str, "Re:", 3)) {
- srcp = str + 3;
- while (isspace(*srcp)) srcp++;
+ srcp = str + subject_get_prefix_length(str);
+ if (srcp != str)
memmove(str, srcp, strlen(srcp) + 1);
- }
}
void trim_subject_for_sort(gchar *str)
g_strstrip(str);
- while (!strncasecmp(str, "Re:", 3)) {
- srcp = str + 3;
- while (isspace(*srcp)) srcp++;
+ srcp = str + subject_get_prefix_length(str);
+ if (srcp != str)
memmove(str, srcp, strlen(srcp) + 1);
- }
}
void trim_subject(gchar *str)
gchar op, cl;
gint in_brace;
- destp = str;
- while (!strncasecmp(destp, "Re:", 3)) {
- destp += 3;
- while (isspace(*destp)) destp++;
- }
+ destp = str + subject_get_prefix_length(str);
if (*destp == '[') {
op = '[';
gboolean is_ascii_str(const guchar *str)
{
+ g_return_val_if_fail(str, FALSE);
+
while (*str != '\0') {
if (*str != '\t' && *str != ' ' &&
*str != '\r' && *str != '\n' &&
} else if (!strncmp(zone, "UT" , 2) ||
!strncmp(zone, "GMT", 2)) {
remoteoffset = 0;
- } else if (strlen(zone3) == 3 &&
- (p = strstr(ustzstr, zone3)) != NULL &&
- (p - ustzstr) % 3 == 0) {
- iustz = ((gint)(p - ustzstr) / 3 + 1) / 2 - 8;
- remoteoffset = iustz * 3600;
+ } else if (strlen(zone3) == 3) {
+ for (p = ustzstr; *p != '\0'; p += 3) {
+ if (!strncasecmp(p, zone3, 3)) {
+ iustz = ((gint)(p - ustzstr) / 3 + 1) / 2 - 8;
+ remoteoffset = iustz * 3600;
+ break;
+ }
+ }
+ if (*p == '\0')
+ return -1;
} else if (strlen(zone3) == 1) {
switch (zone[0]) {
case 'Z': remoteoffset = 0; break;
default: remoteoffset = 0; break;
}
remoteoffset = remoteoffset * 3600;
- }
+ } else
+ return -1;
return remoteoffset;
}
debug_mode = mode;
}
-gboolean debug_get_mode()
+gboolean debug_get_mode(void)
{
return debug_mode;
}
{
if (subject == NULL)
subject = "";
-
- if (g_strncasecmp(subject, "Re: ", 4) == 0)
- return g_hash_table_lookup(subject_table, subject + 4);
else
- return g_hash_table_lookup(subject_table, subject);
+ subject += subject_get_prefix_length(subject);
+
+ return g_hash_table_lookup(subject_table, subject);
}
void subject_table_insert(GHashTable *subject_table, gchar * subject,
void * data)
{
- if (subject == NULL)
- return;
- if (* subject == 0)
+ if (subject == NULL || *subject == 0)
return;
- if (g_strcasecmp(subject, "Re:") == 0)
- return;
- if (g_strcasecmp(subject, "Re: ") == 0)
- return;
-
- if (g_strncasecmp(subject, "Re: ", 4) == 0)
- g_hash_table_insert(subject_table, subject + 4, data);
- else
- g_hash_table_insert(subject_table, subject, data);
+ subject += subject_get_prefix_length(subject);
+ g_hash_table_insert(subject_table, subject, data);
}
void subject_table_remove(GHashTable *subject_table, gchar * subject)
if (subject == NULL)
return;
- if (g_strncasecmp(subject, "Re: ", 4) == 0)
- g_hash_table_remove(subject_table, subject + 4);
- else
- g_hash_table_remove(subject_table, subject);
+ subject += subject_get_prefix_length(subject);
+ g_hash_table_remove(subject_table, subject);
}
-gboolean subject_is_reply(const gchar *subject)
-{
- /* XXX: just simply here so someone can handle really
- * advanced Re: detection like "Re[4]", "ANTW:" or
- * Re: Re: Re: Re: Re: Re: Re: Re:" stuff. */
- if (subject == NULL) return FALSE;
- else return 0 == g_strncasecmp(subject, "Re: ", 4);
+/*!
+ *\brief Check if a string is prefixed with known (combinations)
+ * of prefixes. The function assumes that each prefix
+ * is terminated by zero or exactly _one_ space.
+ *
+ *\param str String to check for a prefixes
+ *
+ *\return int Number of chars in the prefix that should be skipped
+ * for a "clean" subject line. If no prefix was found, 0
+ * is returned.
+ */
+int subject_get_prefix_length(const gchar *subject)
+{
+ /*!< Array with allowable reply prefixes regexps. */
+ static const gchar * const prefixes[] = {
+ "Re\\:", /* "Re:" */
+ "Re\\[[1-9][0-9]*\\]\\:", /* "Re[XXX]:" (non-conforming news mail clients) */
+ "Antw\\:", /* "Antw:" (Dutch / German Outlook) */
+ "Aw\\:", /* "Aw:" (German) */
+ "Antwort\\:", /* "Antwort:" (German Lotus Notes) */
+ "Res\\:", /* "Res:" (Brazilian Outlook) */
+ "Fw\\:", /* "Fw:" Forward */
+ "Enc\\:" /* "Enc:" Forward (Brazilian Outlook) */
+ /* add more */
+ };
+ const int PREFIXES = sizeof prefixes / sizeof prefixes[0];
+ int n;
+ regmatch_t pos;
+ static regex_t regex;
+ static gboolean init_;
+
+ if (!subject) return 0;
+ if (!*subject) return 0;
+
+ if (!init_) {
+ GString *s = g_string_new("");
+
+ for (n = 0; n < PREFIXES; n++)
+ /* Terminate each prefix regexpression by a
+ * "\ ?" (zero or ONE space), and OR them */
+ g_string_sprintfa(s, "(%s\\ ?)%s",
+ prefixes[n],
+ n < PREFIXES - 1 ?
+ "|" : "");
+
+ g_string_prepend(s, "(");
+ g_string_append(s, ")+"); /* match at least once */
+ g_string_prepend(s, "^\\ *"); /* from beginning of line */
+
+
+ /* We now have something like "^\ *((PREFIX1\ ?)|(PREFIX2\ ?))+"
+ * TODO: Should this be "^\ *(((PREFIX1)|(PREFIX2))\ ?)+" ??? */
+ if (regcomp(®ex, s->str, REG_EXTENDED | REG_ICASE)) {
+ debug_print("Error compiling regexp %s\n", s->str);
+ g_string_free(s, TRUE);
+ return 0;
+ } else {
+ init_ = TRUE;
+ g_string_free(s, TRUE);
+ }
+ }
+
+ if (!regexec(®ex, subject, 1, &pos, 0) && pos.rm_so != -1)
+ return pos.rm_eo;
+ else
+ return 0;
}
FILE *get_tmpfile_in_dir(const gchar *dir, gchar **filename)
{
return GPOINTER_TO_INT(a) - GPOINTER_TO_INT(b);
}
+
+gchar *generate_msgid(const gchar *address, gchar *buf, gint len)
+{
+ /* steal from compose.c::compose_generate_msgid() */
+ struct tm *lt;
+ time_t t;
+ gchar *addr;
+
+ t = time(NULL);
+ lt = localtime(&t);
+
+ if (address && *address) {
+ if (strchr(address, '@'))
+ addr = g_strdup(address);
+ else
+ addr = g_strconcat(address, "@", get_domain_name(), NULL);
+ } else
+ addr = g_strconcat(g_get_user_name(), "@", get_domain_name(),
+ NULL);
+
+ g_snprintf(buf, len, "%04d%02d%02d%02d%02d%02d.%08x.%s",
+ lt->tm_year + 1900, lt->tm_mon + 1,
+ lt->tm_mday, lt->tm_hour,
+ lt->tm_min, lt->tm_sec,
+ (guint)random(), addr);
+
+ g_free(addr);
+ return buf;
+}
+
+/**
+ * Create a new boundary in a way that it is very unlikely that this
+ * will occur in the following text. It would be easy to ensure
+ * uniqueness if everything is either quoted-printable or base64
+ * encoded (note that conversion is allowed), but because MIME bodies
+ * may be nested, it may happen that the same boundary has already
+ * been used. We avoid scanning the message for conflicts and hope the
+ * best.
+ *
+ * boundary := 0*69<bchars> bcharsnospace
+ * bchars := bcharsnospace / " "
+ * bcharsnospace := DIGIT / ALPHA / "'" / "(" / ")" /
+ * "+" / "_" / "," / "-" / "." /
+ * "/" / ":" / "=" / "?"
+ *
+ * ":" and "," removed because of buggy MTAs
+ */
+
+gchar *generate_mime_boundary(void)
+{
+ static gchar tbl[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
+ "abcdefghijklmnopqrstuvwxyz"
+ "1234567890'()+_./=?";
+ gchar bufuniq[17];
+ gchar bufdate[BUFFSIZE];
+ int i, equal;
+ int pid;
+
+ pid = getpid();
+
+ /* We make the boundary depend on the pid, so that all running
+ * processed generate different values even when they have been
+ * started within the same second and srand48(time(NULL)) has been
+ * used. I can't see whether this is really an advantage but it
+ * doesn't do any harm.
+ */
+ equal = -1;
+ for (i = 0; i < sizeof(bufuniq) - 1; i++) {
+ bufuniq[i] = tbl[(lrand48() ^ pid) % (sizeof(tbl) - 1)]; /* fill with random */
+ if (bufuniq[i] == '=' && equal == -1)
+ equal = i;
+ }
+ bufuniq[i] = 0;
+
+ /* now make sure that we do have the sequence "=." in it which cannot
+ * be matched by quoted-printable or base64 encoding */
+ if (equal != -1 && (equal + 1) < i)
+ bufuniq[equal + 1] = '.';
+ else {
+ bufuniq[0] = '=';
+ bufuniq[1] = '.';
+ }
+
+ get_rfc822_date(bufdate, sizeof(bufdate));
+ subst_char(bufdate, ' ', '_');
+ subst_char(bufdate, ',', '_');
+ subst_char(bufdate, ':', '_');
+
+ return g_strdup_printf("Multipart_%s_%s",
+ bufdate, bufuniq);
+}