X-Git-Url: https://diplodocus.org/git/nmh/blobdiff_plain/40cbba6fa1e5926cebd5e2a26b92a8f0b647474e..ccafa1944:/sbr/m_getfld.c?ds=sidebyside diff --git a/sbr/m_getfld.c b/sbr/m_getfld.c index 05b9e953..af44f09a 100644 --- a/sbr/m_getfld.c +++ b/sbr/m_getfld.c @@ -11,6 +11,75 @@ #include #include +/* + Purpose + ======= + Reads an Internet message (RFC 5322), or one or more messages + stored in a maildrop in mbox (RFC 4155) or MMDF format, from a file + stream. Each call to m_getfld() reads one header field, or a + portion of the body, in sequence. + + Inputs + ====== + gstate: opaque parse state + bufsz: maximum number of characters to load into buf + iob: input file stream + + Outputs + ======= + name: header field name (array of size NAMESZ=999) + buf: either a header field body or message body + bufsz: number of characters loaded into buf + (return value): message parse state on return from function + + Functions + ========= + void m_getfld_state_destroy (m_getfld_state_t *gstate): destroys + the parse state pointed to by the gstate argument. + + m_getfld_state_reset (m_getfld_state_t *gstate): resets the parse + state to FLD. + + void m_unknown(FILE *iob): Determines the message delimiter string + for the maildrop. Called by inc, scan, and msh when reading from a + maildrop file. + + void m_eomsbr (int (*action)(int)): Sets the hook to check for end + of message in a maildrop. Called only by msh. + + State variables + =============== + m_getfld() retains state internally between calls in the + m_getfld_state_t variable. These are used for detecting the end of + each message when reading maildrops: + + unsigned char **pat_map + unsigned char *fdelim + unsigned char *delimend + int fdelimlen + unsigned char *edelim + int edelimlen + char *msg_delim + int msg_style + int (*eom_action)(int) + + Usage + ===== + m_getfld_state_t gstate = 0; + ... + int state = m_getfld (&gstate, ...); + ... + m_getfld_state_destroy (&gstate); + + The state is retained internally by gstate. To reset its state to FLD: + m_getfld_state_reset (&gstate); +*/ + +/* The following described the old implementation. The high-level + structure hasn't changed, but some of the details have. I'm + leaving this as-is, though, for posterity. + */ + /* This module has a long and checkered history. First, it didn't burst maildrops correctly because it considered two CTRL-A:s in a row to be an inter-message delimiter. It really is four CTRL-A:s followed by a @@ -134,112 +203,31 @@ there is data in "name" or "buf"). */ -/* -Purpose -======= -Reads an Internet message (RFC 5322), or one or more messages stored in a -maildrop in mbox (RFC 4155) or MMDF format, from a file stream. Each call -to m_getfld() reads one header field, or a portion of the body, in sequence. - -Inputs -====== -state: message parse state -bufsz: maximum number of characters to load into buf -iob: input file stream - -Outputs -======= -name: header field name (array of size NAMESZ=999) -buf: either a header field body or message body -bufsz: number of characters loaded into buf -(return value): message parse state on return from function - -Functions -========= -void m_unknown(FILE *iob): Determines the message delimiter string for the - maildrop. Called by inc, scan, and msh when reading from a maildrop file. - -void m_eomsbr (int (*action)(int)): Sets the hook to check for end of - message in a maildrop. Called only by msh. - -Those functions save state in the State variables listed below. - -State variables -=============== -m_getfld() retains state internally between calls in some state variables. -These are used for detecting the end of each message when reading maildrops: -static unsigned char **pat_map -static unsigned char *fdelim -static unsigned char *delimend -static int fdelimlen -static unsigned char *edelim -static int edelimlen -char * msg_delim -int msg_style -int (*eom_action)(int) - -Restrictions -============ -m_getfld() is restricted to operate on one file stream at a time -because of the retained state (see "State variables" above). And the -first call to m_getfld() on that file stream requires that the read -pointer be at the beginning of the file (ftell() of 0). - -Current usage -============= -The first call to m_getfld() on a file stream is with a state of FLD. -Subsequent calls provide the state returned by the previous call. -Therefore, given the Restrictions above, the state variable could be -removed from the signature and just retained internally. -*/ - /* * static prototypes */ -static int m_Eom (int, FILE *); +struct m_getfld_state; +static int m_Eom (m_getfld_state_t, int); static unsigned char *matchc(int, char *, int, char *); -#define eom(c,iob) (msg_style != MS_DEFAULT && \ - (((c) == *msg_delim && m_Eom(c,iob)) ||\ - (eom_action && (*eom_action)(c)))) - -static unsigned char **pat_map; - -static int msg_style = MS_DEFAULT; - -/* - * The "full" delimiter string for a packed maildrop consists - * of a newline followed by the actual delimiter. E.g., the - * full string for a Unix maildrop would be: "\n\nFrom ". - * "Fdelim" points to the start of the full string and is used - * in the BODY case of the main routine to search the buffer for - * a possible eom. Msg_delim points to the first character of - * the actual delim. string (i.e., fdelim+1). Edelim - * points to the 2nd character of actual delimiter string. It - * is used in m_Eom because the first character of the string - * has been read and matched before m_Eom is called. - */ -static char *msg_delim = ""; -static unsigned char *fdelim; -static unsigned char *delimend; -static int fdelimlen; -static unsigned char *edelim; -static int edelimlen; - -static int (*eom_action)(int) = NULL; +#define eom(c,s) (s->msg_style != MS_DEFAULT && \ + (((c) == *s->msg_delim && m_Eom(s,c)) || \ + (s->eom_action && (*s->eom_action)(c)))) /* This replaces the old approach, with its direct access to stdio * internals. It uses one fread() to load a buffer that we manage. * * MSG_INPUT_SIZE is the size of the buffer. * MAX_DELIMITER_SIZE is the maximum size of the delimiter used to - * separate messages in a maildrop, such as mbox "\nFrom ". + * separate messages in a maildrop, such as mbox "From ". * - * test/inc/lots_of_headers.txt depends on the particular value of - * MSG_INPUT_SIZE. */ -#define MSG_INPUT_SIZE 8192 -#define MAX_DELIMITER_SIZE 32 -static struct m_getfld_buffer { + * Some of the tests in the test suite assume a MSG_INPUT_SIZE + * of 4096. + */ +#define MSG_INPUT_SIZE 4096 +#define MAX_DELIMITER_SIZE 5 + +struct m_getfld_state { unsigned char msg_buf[2 * MSG_INPUT_SIZE + MAX_DELIMITER_SIZE]; unsigned char *readpos; unsigned char *end; /* One past the last character read in. */ @@ -247,14 +235,89 @@ static struct m_getfld_buffer { input file stream so that callers can interleave m_getfld() calls with ftell() and fseek(). ytes_read replaces the old m_getfld() msg_count global. last_caller_pos is stored when - leaving m_getfld()/m_unkown(), then checked on the next entry. + leaving m_getfld()/m_unknown(), then checked on the next entry. last_internal_pos is used to remember the position used internally by m_getfld() (read_more(), actually). */ off_t bytes_read; off_t total_bytes_read; /* by caller, not necessarily from input file */ off_t last_caller_pos; off_t last_internal_pos; -} m; + FILE *iob; + + unsigned char **pat_map; + int msg_style; + /* + * The "full" delimiter string for a packed maildrop consists + * of a newline followed by the actual delimiter. E.g., the + * full string for a Unix maildrop would be: "\n\nFrom ". + * "Fdelim" points to the start of the full string and is used + * in the BODY case of the main routine to search the buffer for + * a possible eom. Msg_delim points to the first character of + * the actual delim. string (i.e., fdelim+1). Edelim + * points to the 2nd character of actual delimiter string. It + * is used in m_Eom because the first character of the string + * has been read and matched before m_Eom is called. + */ + char *msg_delim; + unsigned char *fdelim; + unsigned char *delimend; + int fdelimlen; + unsigned char *edelim; + int edelimlen; + int (*eom_action)(int); + int state; + int track_filepos; +}; + +static +void +m_getfld_state_init (m_getfld_state_t *gstate, FILE *iob) { + m_getfld_state_t s; + + s = *gstate = (m_getfld_state_t) mh_xmalloc(sizeof (struct m_getfld_state)); + s->readpos = s->end = s->msg_buf; + s->bytes_read = s->total_bytes_read = 0; + s->last_caller_pos = s->last_internal_pos = 0; + s->iob = iob; + s->pat_map = NULL; + s->msg_style = MS_DEFAULT; + s->msg_delim = ""; + s->fdelim = s->delimend = s->edelim = NULL; + s->fdelimlen = s->edelimlen = 0; + s->eom_action = NULL; + s->state = FLD; + s->track_filepos = 0; +} + +/* scan() needs to force a state an initial state of FLD for each message. */ +void +m_getfld_state_reset (m_getfld_state_t *gstate) { + if (*gstate) { + (*gstate)->state = FLD; + } +} + +/* If the caller interleaves ftell*()/fseek*() calls with m_getfld() + calls, m_getfld() must keep track of the file position. The caller + must use this function to inform m_getfld(). */ +void +m_getfld_track_filepos (m_getfld_state_t *gstate, FILE *iob) { + if (! *gstate) { + m_getfld_state_init (gstate, iob); + } + + (*gstate)->track_filepos = 1; +} + +void m_getfld_state_destroy (m_getfld_state_t *gstate) { + m_getfld_state_t s = *gstate; + + if (s) { + if (s->fdelim) free (s->fdelim-1); + free (s); + *gstate = 0; + } +} /* Summary of file and message input buffer positions: @@ -286,166 +349,187 @@ static struct m_getfld_buffer { static void -enter_getfld (FILE *iob, struct m_getfld_buffer *m) { +enter_getfld (m_getfld_state_t *gstate, FILE *iob) { + m_getfld_state_t s; off_t pos = ftello (iob); - /* Rely on Restriction that the first call to m_getfld (), etc., - is with the read position for the file stream set to 0. */ - if (pos == 0) { - /* A new file stream, so reset the buffer state. */ - m->readpos = m->end = m->msg_buf; - m->total_bytes_read = 0; - m->last_caller_pos = m->last_internal_pos = ftello (iob); - } else { - off_t pos_movement = pos - m->last_caller_pos; /* Can be < 0. */ - - if (pos_movement == 0) { - pos = m->last_internal_pos; + if (! *gstate) { + m_getfld_state_init (gstate, iob); + } + s = *gstate; + s->bytes_read = 0; + + /* This is ugly and no longer necessary, but is retained just in + case it's needed again. The parser used to open the input file + multiple times, so we had to always use the FILE * that's + passed to m_getfld(). Now the parser inits a new + m_getfld_state for each file. See comment below about the + readpos shift code being currently unused. */ + s->iob = iob; + + if (s->track_filepos && (pos != 0 || s->last_internal_pos != 0)) { + if (s->last_internal_pos == 0) { + s->total_bytes_read = pos; } else { - /* The current file stream position differs from the last one, so - caller must have called ftell/o(). Adjust accordingly. */ - if (m->readpos + pos_movement >= m->msg_buf && - m->readpos + pos_movement < m->end) { - /* We can shift readpos and remain within the bounds of - msg_buf. */ - m->readpos += pos_movement; - m->total_bytes_read += pos_movement; - pos = m->last_internal_pos; + off_t pos_movement = pos - s->last_caller_pos; /* Can be < 0. */ + + if (pos_movement == 0) { + pos = s->last_internal_pos; } else { - size_t num_read; - - /* This seek skips past an integral number of chunks of - size MSG_INPUT_SIZE. */ - fseeko (iob, pos / MSG_INPUT_SIZE * MSG_INPUT_SIZE, SEEK_SET); - num_read = fread (m->msg_buf, 1, MSG_INPUT_SIZE, iob); - m->readpos = m->msg_buf + pos % MSG_INPUT_SIZE; - m->end = m->msg_buf + num_read; - m->total_bytes_read = pos; + /* The current file stream position differs from the + last one, so caller must have called ftell/o(). + Or, this is the first call and the file position + was not at 0. */ + + if (s->readpos + pos_movement >= s->msg_buf && + s->readpos + pos_movement < s->end) { + /* This is currently unused. It could be used by + parse_mime() if it was changed to use a global + m_getfld_state. */ + /* We can shift readpos and remain within the + bounds of msg_buf. */ + s->readpos += pos_movement; + s->total_bytes_read += pos_movement; + pos = s->last_internal_pos; + } else { + size_t num_read; + + /* This seek skips past an integral number of + chunks of size MSG_INPUT_SIZE. */ + fseeko (iob, pos/MSG_INPUT_SIZE * MSG_INPUT_SIZE, SEEK_SET); + num_read = fread (s->msg_buf, 1, MSG_INPUT_SIZE, iob); + s->readpos = s->msg_buf + pos % MSG_INPUT_SIZE; + s->end = s->msg_buf + num_read; + s->total_bytes_read = pos; + } } - } - fseeko (iob, pos, SEEK_SET); + fseeko (iob, pos, SEEK_SET); + } } - - m->bytes_read = 0; } static void -leave_getfld (struct m_getfld_buffer *m, FILE *iob) { - /* Save the internal file position that we use for the input buffer. */ - m->last_internal_pos = ftello (iob); - - /* Set file stream position so that callers can use ftell(). */ - m->total_bytes_read += m->bytes_read; - fseeko (iob, m->total_bytes_read, SEEK_SET); - m->last_caller_pos = ftello (iob); +leave_getfld (m_getfld_state_t s) { + s->total_bytes_read += s->bytes_read; + + if (s->track_filepos) { + /* Save the internal file position that we use for the input buffer. */ + s->last_internal_pos = ftello (s->iob); + + /* Set file stream position so that callers can use ftell(). */ + fseeko (s->iob, s->total_bytes_read, SEEK_SET); + s->last_caller_pos = ftello (s->iob); + } } static size_t -read_more (struct m_getfld_buffer *m, FILE *iob) { +read_more (m_getfld_state_t s) { /* Retain at least edelimlen characters that have already been read so that we can back up to them in m_Eom(). */ - ssize_t retain = edelimlen; + ssize_t retain = s->edelimlen; size_t num_read; - if (retain < m->end - m->readpos) retain = m->end - m->readpos; - /* assert (retain <= m->readpos - m->msg_buf <= sizeof msg_buf); */ + if (retain < s->end - s->readpos) retain = s->end - s->readpos; + /* assert (retain <= s->readpos - s->msg_buf <= sizeof msg_buf); */ /* Move what we want to retain at end of the buffer to the beginning. */ - memmove (m->msg_buf, m->readpos - retain, retain); + memmove (s->msg_buf, s->readpos - retain, retain); - m->readpos = m->msg_buf + retain; - num_read = fread (m->readpos, 1, MSG_INPUT_SIZE, iob); - m->end = m->readpos + num_read; + s->readpos = s->msg_buf + retain; + num_read = fread (s->readpos, 1, MSG_INPUT_SIZE, s->iob); + s->end = s->readpos + num_read; return num_read; } static int -Getc (FILE *iob) { - if (m.end - m.readpos < 1) { - if (read_more (&m, iob) == 0) { +Getc (m_getfld_state_t s) { + if (s->end - s->readpos < 1) { + if (read_more (s) == 0) { /* Pretend that we read a character. That's what stdio does. */ - ++m.readpos; + ++s->readpos; return EOF; } } - ++m.bytes_read; - return m.readpos < m.end ? *m.readpos++ : EOF; + ++s->bytes_read; + return s->readpos < s->end ? *s->readpos++ : EOF; } static int -Peek (FILE *iob) { - int next_char = Getc (iob); - --m.readpos; - --m.bytes_read; +Peek (m_getfld_state_t s) { + if (s->end - s->readpos < 1) { + if (read_more (s) == 0) { + /* Pretend that we read a character. That's what stdio does. */ + ++s->readpos; + return EOF; + } + } - return next_char; + return s->readpos < s->end ? *s->readpos : EOF; } static int -Ungetc (int c, FILE *iob) { - NMH_UNUSED (iob); - - if (m.readpos == m.msg_buf) { +Ungetc (int c, m_getfld_state_t s) { + if (s->readpos == s->msg_buf) { return EOF; } else { - --m.bytes_read; - return *--m.readpos = c; + --s->bytes_read; + return *--s->readpos = c; } } int -m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, - int *bufsz, FILE *iob) +m_getfld (m_getfld_state_t *gstate, unsigned char name[NAMESZ], + unsigned char *buf, int *bufsz, FILE *iob) { + m_getfld_state_t s; register unsigned char *cp; register int max, n, c; - enter_getfld (iob, &m); + enter_getfld (gstate, iob); + s = *gstate; - if ((c = Getc(iob)) < 0) { + if ((c = Getc(s)) < 0) { *bufsz = *buf = 0; - leave_getfld (&m, iob); - return FILEEOF; + leave_getfld (s); + return s->state = FILEEOF; } - if (eom (c, iob)) { - if (! eom_action) { + if (eom (c, s)) { + if (! s->eom_action) { /* flush null messages */ - while ((c = Getc(iob)) >= 0 && eom (c, iob)) + while ((c = Getc(s)) >= 0 && eom (c, s)) ; if (c >= 0) - Ungetc(c, iob); + Ungetc(c, s); } *bufsz = *buf = 0; - leave_getfld (&m, iob); - return FILEEOF; + leave_getfld (s); + return s->state = FILEEOF; } - switch (state) { - case FLDEOF: - case BODYEOF: + switch (s->state) { case FLD: if (c == '\n' || c == '-') { /* we hit the header/body separator */ - while (c != '\n' && (c = Getc(iob)) >= 0) continue; + while (c != '\n' && (c = Getc(s)) >= 0) continue; - if (c < 0 || (c = Getc(iob)) < 0 || eom (c, iob)) { - if (! eom_action) { + if (c < 0 || (c = Getc(s)) < 0 || eom (c, s)) { + if (! s->eom_action) { /* flush null messages */ - while ((c = Getc(iob)) >= 0 && eom (c, iob)) + while ((c = Getc(s)) >= 0 && eom (c, s)) ; if (c >= 0) - Ungetc(c, iob); + Ungetc(c, s); } *bufsz = *buf = 0; - leave_getfld (&m, iob); - return FILEEOF; + leave_getfld (s); + return s->state = FILEEOF; } - state = BODY; + s->state = BODY; goto body; } /* @@ -461,7 +545,7 @@ m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, account for that. */ for (n = 1; c != ':' && c != '\n' && c != EOF && n < max; - ++n, c = Getc (iob)) { + ++n, c = Getc (s)) { *cp++ = c; } @@ -469,11 +553,11 @@ m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, the ':' or the first folded whitespace. */ { int next_char; - if (c == EOF || (next_char = Peek (iob)) == EOF) { + if (c == EOF || (next_char = Peek (s)) == EOF) { *bufsz = *cp = *buf = 0; advise (NULL, "eof encountered in field \"%s\"", name); - leave_getfld (&m, iob); - return FMTERR; + leave_getfld (s); + return s->state = FMTERR; } } @@ -498,18 +582,18 @@ m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, /* No, it can't. Oh well, guess we'll blow up. */ *bufsz = *cp = *buf = 0; advise (NULL, "eol encountered in field \"%s\"", name); - state = FMTERR; + s->state = FMTERR; break; } memcpy (buf, name, n - 1); buf[n - 1] = '\n'; buf[n] = '\0'; - /* The last character read was '\n'. m.bytes_read + /* The last character read was '\n'. s->bytes_read (and n) include that, but it was not put into the name array in the for loop above. So subtract 1. */ - *bufsz = --m.bytes_read; /* == n - 1 */ - leave_getfld (&m, iob); - return BODY; + *bufsz = --s->bytes_read; /* == n - 1 */ + leave_getfld (s); + return s->state = BODY; } else if (max <= n) { /* By design, the loop above discards the last character it had read. It's in c, use it. */ @@ -517,7 +601,7 @@ m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, *bufsz = *cp = *buf = 0; advise (NULL, "field name \"%s\" exceeds %d bytes", name, NAMESZ - 2); - state = LENERR; + s->state = LENERR; break; } @@ -540,26 +624,26 @@ m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, n = 0; for (finished = 0; ! finished; ) { while (c != '\n' && c != EOF && n++ < max) { - *cp++ = c = Getc (iob); + *cp++ = c = Getc (s); } - if (c != EOF) c = Peek (iob); + if (c != EOF) c = Peek (s); if (max < n) { /* the dest buffer is full */ - state = FLDPLUS; + s->state = FLDPLUS; finished = 1; } else if (c != ' ' && c != '\t') { /* The next character is not folded whitespace, so prepare to move on to the next field. It's OK if c is EOF, it will be handled on the next call to m_getfld (). */ - state = FLD; + s->state = FLD; finished = 1; } else { /* Folded header field, continues on the next line. */ } } - *bufsz = m.bytes_read; + *bufsz = s->bytes_read; break; } @@ -572,10 +656,10 @@ m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, unsigned char *bp; max = *bufsz-1; - /* Back up and store the current position and update cnt. */ - bp = --m.readpos; - c = m.end - m.readpos < max ? m.end - m.readpos : max; - if (msg_style != MS_DEFAULT && c > 1) { + /* Back up and store the current position. */ + bp = --s->readpos; + c = s->end - s->readpos < max ? s->end - s->readpos : max; + if (s->msg_style != MS_DEFAULT && c > 1) { /* * packed maildrop - only take up to the (possible) * start of the next message. This "matchc" should @@ -590,7 +674,7 @@ m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, */ unsigned char *ep; - if ((ep = matchc( fdelimlen, fdelim, c, bp ))) + if ((ep = matchc( s->fdelimlen, s->fdelim, c, bp ))) c = ep - bp + 1; else { /* @@ -609,7 +693,7 @@ m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, unsigned char *sp; ep = bp + c - 1; - if ((sp = pat_map[*ep])) { + if ((sp = s->pat_map[*ep])) { do { /* This if() is true unless (a) the buffer is too * small to contain this delimiter prefix, or @@ -621,7 +705,7 @@ m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, * should have found it. Thus it's not a delim * and we know we won't get a match. */ - if (((sp - fdelim) + 2) <= c) { + if (((sp - s->fdelim) + 2) <= c) { cp = sp; /* Unfortunately although fdelim has a preceding NUL * we can't use this as a sentinel in case the buffer @@ -629,9 +713,9 @@ m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, * would cause us to run off the front of fdelim). */ while (*--ep == *--cp) - if (cp < fdelim) + if (cp < s->fdelim) break; - if (cp < fdelim) { + if (cp < s->fdelim) { /* we matched the entire delim prefix, * so only take the buffer up to there. * we know ep >= bp -- check above prevents underrun @@ -642,7 +726,7 @@ m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, } /* try matching one less char of delim string */ ep = bp + c - 1; - } while (--sp > fdelim); + } while (--sp > s->fdelim); } } } @@ -650,35 +734,39 @@ m_getfld (int state, unsigned char name[NAMESZ], unsigned char *buf, /* Advance the current position to reflect the copy out. c is less than or equal to the number of bytes remaining in the read buffer, so will not overrun it. */ - m.readpos += c; + s->readpos += c; cp = buf + c; /* Subtract 1 from c because the first character was read by - Getc(), and therefore already accounted for in m.bytes_read. */ - m.bytes_read += c - 1; - *bufsz = m.bytes_read; + Getc(), and therefore already accounted for in s->bytes_read. */ + s->bytes_read += c - 1; + *bufsz = s->bytes_read; break; } default: - adios (NULL, "m_getfld() called with bogus state of %d", state); + adios (NULL, "m_getfld() called with bogus state of %d", s->state); } *cp = 0; - leave_getfld (&m, iob); + leave_getfld (s); - return state; + return s->state; } void -m_unknown(FILE *iob) +m_unknown(m_getfld_state_t *gstate, FILE *iob) { + m_getfld_state_t s; register int c; char text[MAX_DELIMITER_SIZE]; + char from[] = "From "; register char *cp; register char *delimstr; + unsigned int i; - enter_getfld (iob, &m); + enter_getfld (gstate, iob); + s = *gstate; /* * Figure out what the message delimitter string is for this @@ -693,40 +781,35 @@ m_unknown(FILE *iob) * specified when nmh was built (or from the mts.conf file). */ - msg_style = MS_UNKNOWN; + s->msg_style = MS_UNKNOWN; - for (c = 0, cp = text; c < 5; ++c, ++cp) { - if ((*cp = Getc (iob)) == EOF) { + for (i = 0, cp = text; i < sizeof text; ++i, ++cp) { + if ((*cp = Getc (s)) == EOF) { break; } } - if (c == 5 && strncmp (text, "From ", 5) == 0) { - msg_style = MS_MBOX; + if (i == sizeof from-1 && strncmp (text, "From ", sizeof from-1) == 0) { + s->msg_style = MS_MBOX; delimstr = "\nFrom "; - while ((c = Getc (iob)) != '\n' && c >= 0) continue; - /* m_unknown is only called on maildrop files, and they are only - read using m_getfld (). The caller musn't try to read from - the stream directly because the file position indicator was - not advanced based on bytes_read, but instead on whatever - was read into the message buffer. */ + while ((c = Getc (s)) != '\n' && c >= 0) continue; } else { /* not a Unix style maildrop */ - m.readpos -= m.bytes_read; + s->readpos -= s->bytes_read; delimstr = mmdlm2; - msg_style = MS_MMDF; + s->msg_style = MS_MMDF; } c = strlen (delimstr); - fdelim = (unsigned char *) mh_xmalloc((size_t) (c + 3)); - *fdelim++ = '\0'; - *fdelim = '\n'; - msg_delim = (char *)fdelim+1; - edelim = (unsigned char *)msg_delim+1; - fdelimlen = c + 1; - edelimlen = c - 1; /* == strlen (delimstr) */ - strcpy (msg_delim, delimstr); - delimend = (unsigned char *)msg_delim + edelimlen; - if (edelimlen <= 1) + s->fdelim = (unsigned char *) mh_xmalloc((size_t) (c + 3)); + *s->fdelim++ = '\0'; + *s->fdelim = '\n'; + s->msg_delim = (char *)s->fdelim+1; + s->edelim = (unsigned char *)s->msg_delim+1; + s->fdelimlen = c + 1; + s->edelimlen = c - 1; /* == strlen (delimstr) */ + strcpy (s->msg_delim, delimstr); + s->delimend = (unsigned char *)s->msg_delim + s->edelimlen; + if (s->edelimlen <= 1) adios (NULL, "maildrop delimiter must be at least 2 bytes"); /* * build a Boyer-Moore end-position map for the matcher in m_getfld. @@ -734,36 +817,36 @@ m_unknown(FILE *iob) * separator) or the last char (since the matchc would have found it * if it was a real delim). */ - pat_map = (unsigned char **) calloc (256, sizeof(unsigned char *)); + s->pat_map = (unsigned char **) calloc (256, sizeof(unsigned char *)); - for (cp = (char *) fdelim + 1; cp < (char *) delimend; cp++ ) - pat_map[(unsigned char)*cp] = (unsigned char *) cp; + for (cp = (char *) s->fdelim + 1; cp < (char *) s->delimend; cp++ ) + s->pat_map[(unsigned char)*cp] = (unsigned char *) cp; - if (msg_style == MS_MMDF) { + if (s->msg_style == MS_MMDF) { /* flush extra msg hdrs */ - while ((c = Getc(iob)) >= 0 && eom (c, iob)) + while ((c = Getc(s)) >= 0 && eom (c, s)) ; if (c >= 0) - Ungetc(c, iob); + Ungetc(c, s); } - leave_getfld (&m, iob); + leave_getfld (s); } void -m_eomsbr (int (*action)(int)) +m_eomsbr (m_getfld_state_t s, int (*action)(int)) { - if ((eom_action = action)) { - msg_style = MS_MSH; - *msg_delim = 0; - fdelimlen = 1; - delimend = fdelim; + if ((s->eom_action = action)) { + s->msg_style = MS_MSH; + *s->msg_delim = 0; + s->fdelimlen = 1; + s->delimend = s->fdelim; } else { - msg_style = MS_MMDF; - msg_delim = (char *)fdelim + 1; - fdelimlen = strlen((char *)fdelim); - delimend = (unsigned char *)(msg_delim + edelimlen); + s->msg_style = MS_MMDF; + s->msg_delim = (char *)s->fdelim + 1; + s->fdelimlen = strlen((char *)s->fdelim); + s->delimend = (unsigned char *)(s->msg_delim + s->edelimlen); } } @@ -773,20 +856,21 @@ m_eomsbr (int (*action)(int)) */ static int -m_Eom (int c, FILE *iob) +m_Eom (m_getfld_state_t s, int c) { register int i; char text[MAX_DELIMITER_SIZE]; char *cp; - for (i = 0, cp = text; i < edelimlen; ++i, ++cp) { - if ((*cp = Getc (iob)) == EOF) { + for (i = 0, cp = text; i < s->edelimlen; ++i, ++cp) { + if ((*cp = Getc (s)) == EOF) { break; } } - if (i != edelimlen || strncmp (text, (char *)edelim, edelimlen)) { - if (i == 0 && msg_style == MS_MBOX) + if (i != s->edelimlen || + strncmp (text, (char *)s->edelim, s->edelimlen)) { + if (i == 0 && s->msg_style == MS_MBOX) /* the final newline in the (brain damaged) unix-format * maildrop is part of the delimitter - delete it. */ @@ -795,12 +879,12 @@ m_Eom (int c, FILE *iob) /* Did not find delimiter, so restore the read position. Note that on input, a character had already been read with Getc(). It will be unget by m_getfld () on return. */ - m.readpos -= m.bytes_read - 1; + s->readpos -= s->bytes_read - 1; return 0; } - if (msg_style == MS_MBOX) { - while ((c = Getc (iob)) != '\n') + if (s->msg_style == MS_MBOX) { + while ((c = Getc (s)) != '\n') if (c < 0) break; }