add patch RFE 4244, 'Handle additional Date header formats'
[claws.git] / src / procheader.c
index 9e67aeda13eb2d617f687019820f922e263ae460..f5755b6a395694be369d68aa363a4a918629e301 100644 (file)
@@ -41,6 +41,7 @@
 #include "hooks.h"
 #include "utils.h"
 #include "defs.h"
+#include "file-utils.h"
 
 #define BUFFSIZE       8192
 
@@ -49,15 +50,15 @@ static gchar monthstr[] = "JanFebMarAprMayJunJulAugSepOctNovDec";
 typedef char *(*getlinefunc) (char *, size_t, void *);
 typedef int (*peekcharfunc) (void *);
 typedef int (*getcharfunc) (void *);
-typedef gint (*get_one_field_func) (gchar *, size_t, void *, HeaderEntry[]);
+typedef gint (*get_one_field_func) (gchar **, void *, HeaderEntry[]);
 
-static gint string_get_one_field(gchar *buf, size_t len, char **str,
+static gint string_get_one_field(gchar **buf, char **str,
                                 HeaderEntry hentry[]);
 
 static char *string_getline(char *buf, size_t len, char **str);
 static int string_peekchar(char **str);
 static int file_peekchar(FILE *fp);
-static gint generic_get_one_field(gchar *buf, size_t len, void *data,
+static gint generic_get_one_field(gchar **bufptr, void *data,
                                  HeaderEntry hentry[],
                                  getlinefunc getline, 
                                  peekcharfunc peekchar,
@@ -66,23 +67,41 @@ static MsgInfo *parse_stream(void *data, gboolean isstring, MsgFlags flags,
                             gboolean full, gboolean decrypted);
 
 
-gint procheader_get_one_field(gchar *buf, size_t len, FILE *fp,
+gint procheader_get_one_field(gchar **buf, FILE *fp,
                              HeaderEntry hentry[])
 {
-       return generic_get_one_field(buf, len, fp, hentry,
+       return generic_get_one_field(buf, fp, hentry,
                                     (getlinefunc)fgets_crlf, (peekcharfunc)file_peekchar,
                                     TRUE);
 }
 
-static gint string_get_one_field(gchar *buf, size_t len, char **str,
+static gint string_get_one_field(gchar **buf, char **str,
                                 HeaderEntry hentry[])
 {
-       return generic_get_one_field(buf, len, str, hentry,
+       return generic_get_one_field(buf, str, hentry,
                                     (getlinefunc)string_getline,
                                     (peekcharfunc)string_peekchar,
                                     TRUE);
 }
 
+gboolean procheader_skip_headers(FILE *fp)
+{
+       gchar *buf = g_malloc(BUFFSIZE);
+       do {
+               if (fgets_crlf(buf, BUFFSIZE - 1, fp) == NULL) {
+                       g_free(buf);
+                       return FALSE;
+               }
+               if (buf[0] == '\r' || buf[0] == '\n') {
+                       break;
+               }
+       } while (TRUE);
+       g_free(buf);
+
+       return TRUE;
+}
+
+
 static char *string_getline(char *buf, size_t len, char **str)
 {
        gboolean is_cr = FALSE;
@@ -119,23 +138,44 @@ static int file_peekchar(FILE *fp)
        return ungetc(getc(fp), fp);
 }
 
-static gint generic_get_one_field(gchar *buf, size_t len, void *data,
+static gint generic_get_one_field(gchar **bufptr, void *data,
                          HeaderEntry *hentry,
                          getlinefunc getline, peekcharfunc peekchar,
                          gboolean unfold)
 {
+       /* returns -1 in case of failure of any kind, whatever it's a parsing error
+          or an allocation error. if returns -1, *bufptr is always NULL, and vice-versa,
+          and if returning 0 (OK), *bufptr is always non-NULL, so callers just have to
+          test the return value
+       */
        gint nexthead;
        gint hnum = 0;
        HeaderEntry *hp = NULL;
+       size_t len;
+       gchar *buf;
+
+       cm_return_val_if_fail(bufptr != NULL, -1);
+
+       len = BUFFSIZE;
+       buf = g_malloc(len);
 
        if (hentry != NULL) {
                /* skip non-required headers */
+               /* and get hentry header line */
                do {
                        do {
-                               if (getline(buf, len, data) == NULL)
+                               if (getline(buf, len, data) == NULL) {
+                                       debug_print("generic_get_one_field: getline\n");
+                                       g_free(buf);
+                                       *bufptr = NULL;
                                        return -1;
-                               if (buf[0] == '\r' || buf[0] == '\n')
+                               }
+                               if (buf[0] == '\r' || buf[0] == '\n') {
+                                       debug_print("generic_get_one_field: empty line\n");
+                                       g_free(buf);
+                                       *bufptr = NULL;
                                        return -1;
+                               }
                        } while (buf[0] == ' ' || buf[0] == '\t');
 
                        for (hp = hentry, hnum = 0; hp->name != NULL;
@@ -146,8 +186,27 @@ static gint generic_get_one_field(gchar *buf, size_t len, void *data,
                        }
                } while (hp->name == NULL);
        } else {
-               if (getline(buf, len, data) == NULL) return -1;
-               if (buf[0] == '\r' || buf[0] == '\n') return -1;
+               /* read first line */
+               if (getline(buf, len, data) == NULL) {
+                       debug_print("generic_get_one_field: getline\n");
+                       g_free(buf);
+                       *bufptr = NULL;
+                       return -1;
+               }
+               if (buf[0] == '\r' || buf[0] == '\n') {
+                       debug_print("generic_get_one_field: empty line\n");
+                       g_free(buf);
+                       *bufptr = NULL;
+                       return -1;
+               }
+       }
+       /* reduce initial buffer to its useful part */
+       len = strlen(buf)+1;
+       buf = g_realloc(buf, len);
+       if (buf == NULL) {
+               debug_print("generic_get_one_field: reallocation error\n");
+               *bufptr = NULL;
+               return -1;
        }
 
        /* unfold line */
@@ -156,6 +215,9 @@ static gint generic_get_one_field(gchar *buf, size_t len, void *data,
                /* ([*WSP CRLF] 1*WSP) */
                if (nexthead == ' ' || nexthead == '\t') {
                        size_t buflen;
+                       gchar *tmpbuf;
+                       size_t tmplen;
+
                        gboolean skiptab = (nexthead == '\t');
                        /* trim previous trailing \n if requesting one header or
                         * unfolding was requested */
@@ -164,15 +226,29 @@ static gint generic_get_one_field(gchar *buf, size_t len, void *data,
 
                        buflen = strlen(buf);
                        
-                       /* concatenate next line */
-                       if ((len - buflen) > 2) {
-                               if (getline(buf + buflen, len - buflen, data) == NULL)
-                                       break;
-                               if (skiptab) { /* replace tab with space */
-                                       *(buf + buflen) = ' ';
-                               }
-                       } else
+                       /* read next line */
+                       tmpbuf = g_malloc(BUFFSIZE);
+
+                       if (getline(tmpbuf, BUFFSIZE, data) == NULL) {
+                               g_free(tmpbuf);
                                break;
+                       }
+                       tmplen = strlen(tmpbuf)+1;
+
+                       /* extend initial buffer and concatenate next line */
+                       len += tmplen;
+                       buf = g_realloc(buf, len);
+                       if (buf == NULL) {
+                               debug_print("generic_get_one_field: reallocation error\n");
+                               g_free(buf);
+                               *bufptr = NULL;
+                               return -1;
+                       }
+                       memcpy(buf+buflen, tmpbuf, tmplen);
+                       g_free(tmpbuf);
+                       if (skiptab) { /* replace tab with space */
+                               *(buf + buflen) = ' ';
+                       }
                } else {
                        /* remove trailing new line */
                        strretchomp(buf);
@@ -180,12 +256,14 @@ static gint generic_get_one_field(gchar *buf, size_t len, void *data,
                }
        }
 
+       *bufptr = buf;
+
        return hnum;
 }
 
-gint procheader_get_one_field_asis(gchar *buf, size_t len, FILE *fp)
+gint procheader_get_one_field_asis(gchar **buf, FILE *fp)
 {
-       return generic_get_one_field(buf, len, fp, NULL,
+       return generic_get_one_field(buf, fp, NULL,
                                     (getlinefunc)fgets_crlf, 
                                     (peekcharfunc)file_peekchar,
                                     FALSE);
@@ -193,7 +271,7 @@ gint procheader_get_one_field_asis(gchar *buf, size_t len, FILE *fp)
 
 GPtrArray *procheader_get_header_array_asis(FILE *fp)
 {
-       gchar buf[BUFFSIZE];
+       gchar *buf = NULL;
        GPtrArray *headers;
        Header *header;
 
@@ -201,9 +279,11 @@ GPtrArray *procheader_get_header_array_asis(FILE *fp)
 
        headers = g_ptr_array_new();
 
-       while (procheader_get_one_field_asis(buf, sizeof(buf), fp) != -1) {
+       while (procheader_get_one_field_asis(&buf, fp) != -1) {
                if ((header = procheader_parse_header(buf)) != NULL)
                        g_ptr_array_add(headers, header);
+               g_free(buf);
+               buf = NULL;
        }
 
        return headers;
@@ -214,6 +294,8 @@ void procheader_header_array_destroy(GPtrArray *harray)
        gint i;
        Header *header;
 
+       cm_return_if_fail(harray != NULL);
+
        for (i = 0; i < harray->len; i++) {
                header = g_ptr_array_index(harray, i);
                procheader_header_free(header);
@@ -290,6 +372,8 @@ Header * procheader_parse_header(gchar * buf)
        Header * header;
        gboolean addr_field = FALSE;
 
+       cm_return_val_if_fail(buf != NULL, NULL);
+
        if ((*buf == ':') || (*buf == ' '))
                return NULL;
 
@@ -309,15 +393,14 @@ Header * procheader_parse_header(gchar * buf)
 
 void procheader_get_header_fields(FILE *fp, HeaderEntry hentry[])
 {
-       gchar buf[BUFFSIZE];
+       gchar *buf = NULL;
        HeaderEntry *hp;
        gint hnum;
        gchar *p;
 
        if (hentry == NULL) return;
 
-       while ((hnum = procheader_get_one_field(buf, sizeof(buf), fp, hentry))
-              != -1) {
+       while ((hnum = procheader_get_one_field(&buf, fp, hentry)) != -1) {
                hp = hentry + hnum;
 
                p = buf + strlen(hp->name);
@@ -331,36 +414,76 @@ void procheader_get_header_fields(FILE *fp, HeaderEntry hentry[])
                        hp->body = g_strconcat(tp, ", ", p, NULL);
                        g_free(tp);
                }
+               g_free(buf);
+               buf = NULL;
        }
 }
 
 MsgInfo *procheader_parse_file(const gchar *file, MsgFlags flags,
                               gboolean full, gboolean decrypted)
 {
+#ifdef G_OS_WIN32
+       GFile *f;
+       GFileInfo *fi;
+       GTimeVal tv;
+       GError *error = NULL;
+#else
        GStatBuf s;
+#endif
        FILE *fp;
        MsgInfo *msginfo;
 
+#ifdef G_OS_WIN32
+       f = g_file_new_for_path(file);
+       fi = g_file_query_info(f, "standard::size,standard::type,time::modified",
+                       G_FILE_QUERY_INFO_NONE, NULL, &error);
+       if (error != NULL) {
+               g_warning(error->message);
+               g_error_free(error);
+               g_object_unref(f);
+       }
+#else
        if (g_stat(file, &s) < 0) {
                FILE_OP_ERROR(file, "stat");
                return NULL;
        }
+#endif
+
+#ifdef G_OS_WIN32
+       if (g_file_info_get_file_type(fi) != G_FILE_TYPE_REGULAR) {
+               g_object_unref(fi);
+               g_object_unref(f);
+               return NULL;
+       }
+#else
        if (!S_ISREG(s.st_mode))
                return NULL;
+#endif
 
-       if ((fp = g_fopen(file, "rb")) == NULL) {
-               FILE_OP_ERROR(file, "fopen");
+       if ((fp = claws_fopen(file, "rb")) == NULL) {
+               FILE_OP_ERROR(file, "claws_fopen");
                return NULL;
        }
 
        msginfo = procheader_parse_stream(fp, flags, full, decrypted);
-       fclose(fp);
+       claws_fclose(fp);
 
        if (msginfo) {
+#ifdef G_OS_WIN32
+               msginfo->size = g_file_info_get_size(fi);
+               g_file_info_get_modification_time(fi, &tv);
+               msginfo->mtime = tv.tv_sec;
+#else
                msginfo->size = s.st_size;
                msginfo->mtime = s.st_mtime;
+#endif
        }
 
+#ifdef G_OS_WIN32
+       g_object_unref(fi);
+       g_object_unref(f);
+#endif
+
        return msginfo;
 }
 
@@ -484,13 +607,13 @@ static gboolean avatar_from_some_face(gpointer source, gpointer userdata)
        return FALSE;
 }
 
-static guint avatar_hook_id = 0;
+static gulong avatar_hook_id = HOOK_NONE;
 
 static MsgInfo *parse_stream(void *data, gboolean isstring, MsgFlags flags,
                             gboolean full, gboolean decrypted)
 {
        MsgInfo *msginfo;
-       gchar buf[BUFFSIZE];
+       gchar *buf = NULL;
        gchar *p, *tmp;
        gchar *hp;
        HeaderEntry *hentry;
@@ -504,12 +627,15 @@ static MsgInfo *parse_stream(void *data, gboolean isstring, MsgFlags flags,
        hentry = procheader_get_headernames(full);
 
        if (MSG_IS_QUEUED(flags) || MSG_IS_DRAFT(flags)) {
-               while (get_one_field(buf, sizeof(buf), data, NULL) != -1) {
+               while (get_one_field(&buf, data, NULL) != -1) {
                        if ((!strncmp(buf, "X-Claws-End-Special-Headers: 1",
                                strlen("X-Claws-End-Special-Headers:"))) ||
                            (!strncmp(buf, "X-Sylpheed-End-Special-Headers: 1",
-                               strlen("X-Sylpheed-End-Special-Headers:"))))
+                               strlen("X-Sylpheed-End-Special-Headers:")))) {
+                               g_free(buf);
+                               buf = NULL;
                                break;
+                       }
                        /* from other mailers */
                        if (!strncmp(buf, "Date: ", 6)
                        ||  !strncmp(buf, "To: ", 4)
@@ -519,8 +645,12 @@ static MsgInfo *parse_stream(void *data, gboolean isstring, MsgFlags flags,
                                        data = orig_data;
                                else 
                                        rewind((FILE *)data);
+                               g_free(buf);
+                               buf = NULL;
                                break;
                        }
+                       g_free(buf);
+                       buf = NULL;
                }
        }
 
@@ -533,15 +663,14 @@ static MsgInfo *parse_stream(void *data, gboolean isstring, MsgFlags flags,
        
        msginfo->inreplyto = NULL;
 
-       if (avatar_hook_id == 0 && (prefs_common.enable_avatars & AVATARS_ENABLE_CAPTURE)) {
+       if (avatar_hook_id == HOOK_NONE && (prefs_common.enable_avatars & AVATARS_ENABLE_CAPTURE)) {
                avatar_hook_id = hooks_register_hook(AVATAR_HEADER_UPDATE_HOOKLIST, avatar_from_some_face, NULL);
-       } else if (avatar_hook_id != 0 && !(prefs_common.enable_avatars & AVATARS_ENABLE_CAPTURE)) {
+       } else if (avatar_hook_id != HOOK_NONE && !(prefs_common.enable_avatars & AVATARS_ENABLE_CAPTURE)) {
                hooks_unregister_hook(AVATAR_HEADER_UPDATE_HOOKLIST, avatar_hook_id);
-               avatar_hook_id = 0;
+               avatar_hook_id = HOOK_NONE;
        }
 
-       while ((hnum = get_one_field(buf, sizeof(buf), data, hentry))
-              != -1) {
+       while ((hnum = get_one_field(&buf, data, hentry)) != -1) {
                hp = buf + strlen(hentry[hnum].name);
                while (*hp == ' ' || *hp == '\t') hp++;
 
@@ -743,6 +872,8 @@ static MsgInfo *parse_stream(void *data, gboolean isstring, MsgFlags flags,
                        hooks_invoke(AVATAR_HEADER_UPDATE_HOOKLIST, (gpointer)acd);
                        g_free(acd);
                }
+               g_free(buf);
+               buf = NULL;
        }
 
        if (!msginfo->inreplyto && msginfo->references)
@@ -807,6 +938,10 @@ static gint procheader_scan_date_string(const gchar *str,
                        weekday, day, month, year, hh, mm, ss, zone);
        if (result == 8) return 0;
 
+       result = sscanf(str, "%3s %3s %d %2d:%2d:%2d %d %6s",
+                       weekday, month, day, hh, mm, ss, year, zone);
+       if (result == 8) return 0;
+
        result = sscanf(str, "%d %9s %d %2d:%2d:%2d %6s",
                        day, month, year, hh, mm, ss, zone);
        if (result == 7) return 0;
@@ -816,6 +951,10 @@ static gint procheader_scan_date_string(const gchar *str,
                        weekday, day, month, year, hh, mm, ss);
        if (result == 7) return 0;
 
+       result = sscanf(str, "%3s %3s %d %2d:%2d:%2d %d",
+                       weekday, month, day, hh, mm, ss, year);
+       if (result == 7) return 0;
+
        result = sscanf(str, "%d %9s %d %2d:%2d:%2d",
                        day, month, year, hh, mm, ss);
        if (result == 6) return 0;
@@ -843,7 +982,6 @@ static gint procheader_scan_date_string(const gchar *str,
        /* RFC3339 subset, with fraction of second */
        result = sscanf(str, "%4d-%2d-%2d%c%2d:%2d:%2d.%d%6s",
                        year, &month_n, day, &sep1, hh, mm, ss, &secfract, zonestr);
-       debug_print("str |%s|, result %d\n", str, result);
        if (result == 9
                        && (sep1 == 'T' || sep1 == 't' || sep1 == ' ')) {
                if (month_n >= 1 && month_n <= 12) {
@@ -861,7 +999,6 @@ static gint procheader_scan_date_string(const gchar *str,
        /* RFC3339 subset, no fraction of second */
        result = sscanf(str, "%4d-%2d-%2d%c%2d:%2d:%2d%6s",
                        year, &month_n, day, &sep1, hh, mm, ss, zonestr);
-       debug_print("str |%s|, result %d\n", str, result);
        if (result == 8
                        && (sep1 == 'T' || sep1 == 't' || sep1 == ' ')) {
                if (month_n >= 1 && month_n <= 12) {
@@ -878,8 +1015,8 @@ static gint procheader_scan_date_string(const gchar *str,
 
        *zone = '\0';
 
-       /* RFC3339 subset */
-       /* This particular "subset" is invalid, RFC requires the time offset */
+       /* RFC3339 subset, no fraction of second, and no timezone offset */
+       /* This particular "subset" is invalid, RFC requires the offset */
        result = sscanf(str, "%4d-%2d-%2d %2d:%2d:%2d",
                        year, &month_n, day, hh, mm, ss);
        if (result == 6) {
@@ -889,6 +1026,17 @@ static gint procheader_scan_date_string(const gchar *str,
                }
        }
 
+       /* ISO8601 format with just date (YYYY-MM-DD) */
+       result = sscanf(str, "%4d-%2d-%2d",
+                       year, &month_n, day);
+       if (result == 3) {
+               *hh = *mm = *ss = 0;
+               if (1 <= month_n && month_n <= 12) {
+                       strncpy2(month, monthstr+((month_n-1)*3), 4);
+                       return 0;
+               }
+       }
+
        return -1;
 }
 
@@ -957,10 +1105,8 @@ time_t procheader_date_parse(gchar *dest, const gchar *src, gint len)
        gint hh, mm, ss;
        gchar zone[7];
        GDateMonth dmonth = G_DATE_BAD_MONTH;
-       struct tm t;
        gchar *p;
        time_t timer;
-       time_t tz_offset;
 
        if (procheader_scan_date_string(src, weekday, &day, month, &year,
                                        &hh, &mm, &ss, zone) < 0) {
@@ -969,14 +1115,6 @@ time_t procheader_date_parse(gchar *dest, const gchar *src, gint len)
                return 0;
        }
 
-       /* Y2K compliant :) */
-       if (year < 1000) {
-               if (year < 50)
-                       year += 2000;
-               else
-                       year += 1900;
-       }
-
        month[3] = '\0';
        for (p = monthstr; *p != '\0'; p += 3) {
                if (!g_ascii_strncasecmp(p, month, 3)) {
@@ -985,6 +1123,31 @@ time_t procheader_date_parse(gchar *dest, const gchar *src, gint len)
                }
        }
 
+#ifdef G_OS_WIN32
+       GTimeZone *tz;
+       GDateTime *dt, *dt2;
+
+       tz = g_time_zone_new(zone); // can't return NULL no need to check for it
+       dt = g_date_time_new(tz, 1, 1, 1, 0, 0, 0);
+       g_time_zone_unref(tz);
+       dt2 = g_date_time_add_full(dt, year-1, dmonth-1, day-1, hh, mm, ss);
+       g_date_time_unref(dt);
+
+       timer = g_date_time_to_unix(dt2);
+       g_date_time_unref(dt2);
+
+#else
+       struct tm t;
+       time_t tz_offset;
+
+       /* Y2K compliant :) */
+       if (year < 1000) {
+               if (year < 50)
+                       year += 2000;
+               else
+                       year += 1900;
+       }
+
        t.tm_sec = ss;
        t.tm_min = mm;
        t.tm_hour = hh;
@@ -1002,6 +1165,7 @@ time_t procheader_date_parse(gchar *dest, const gchar *src, gint len)
 
        if (dest)
                procheader_date_get_localtime(dest, len, timer);
+#endif
 
        return timer;
 }
@@ -1039,34 +1203,44 @@ void procheader_date_get_localtime(gchar *dest, gint len, const time_t timer)
 
 /* Added by Mel Hadasht on 27 Aug 2001 */
 /* Get a header from msginfo */
-gint procheader_get_header_from_msginfo(MsgInfo *msginfo, gchar *buf, gint len, gchar *header)
+gint procheader_get_header_from_msginfo(MsgInfo *msginfo, gchar **buf, gchar *header)
 {
        gchar *file;
        FILE *fp;
        HeaderEntry hentry[]={ { NULL, NULL, TRUE  },
-                               { NULL, NULL, FALSE } };
+                                                  { NULL, NULL, FALSE } };
        gint val;
 
-        hentry[0].name = header;
-       
        cm_return_val_if_fail(msginfo != NULL, -1);
+       cm_return_val_if_fail(buf != NULL, -1);
+       cm_return_val_if_fail(header != NULL, -1);
+
+       hentry[0].name = header;
+
        file = procmsg_get_message_file_path(msginfo);
-       if ((fp = g_fopen(file, "rb")) == NULL) {
-               FILE_OP_ERROR(file, "fopen");
-               g_free(file);
-               return -1;
+       if ((fp = claws_fopen(file, "rb")) == NULL) {
+               FILE_OP_ERROR(file, "claws_fopen");
+               g_free(file);
+               g_free(*buf);
+               *buf = NULL;
+               return -1;
        }
-       val = procheader_get_one_field(buf,len, fp, hentry);
-       if (fclose(fp) == EOF) {
-               FILE_OP_ERROR(file, "fclose");
+       val = procheader_get_one_field(buf, fp, hentry);
+
+       if (claws_fclose(fp) == EOF) {
+               FILE_OP_ERROR(file, "claws_fclose");
                claws_unlink(file);
                g_free(file);
+               g_free(*buf);
+               *buf = NULL;
                return -1;
        }
 
        g_free(file);
-        if (val == -1)
+       if (val == -1) {
+               /* *buf is already NULL in that case, see procheader_get_one_field() */
                return -1;
+       }
 
        return 0;
 }