*************************************************/
/* Copyright (c) University of Cambridge 1995 - 2018 */
-/* Copyright (c) The Exim Maintainers 2020 */
+/* Copyright (c) The Exim Maintainers 2020 - 2021 */
/* See the file NOTICE for conditions of use and distribution. */
/* Functions for parsing addresses */
#include "exim.h"
-static uschar *last_comment_position;
+static const uschar *last_comment_position;
*/
uschar *
-parse_find_address_end(uschar *s, BOOL nl_ends)
+parse_find_address_end(const uschar *s, BOOL nl_ends)
{
BOOL source_routing = *s == '@';
int no_term = source_routing? 1 : 0;
}
}
-return s;
+return US s;
}
Returns: pointer to the last @ in an address, or NULL if none
*/
-uschar *
-parse_find_at(uschar *s)
+const uschar *
+parse_find_at(const uschar *s)
{
-uschar *t = s + Ustrlen(s);
+const uschar * t = s + Ustrlen(s);
while (--t >= s)
- {
if (*t == '@')
{
int backslash_count = 0;
- uschar *tt = t - 1;
+ const uschar *tt = t - 1;
while (tt > s && *tt-- == '\\') backslash_count++;
if ((backslash_count & 1) == 0) return t;
}
- else if (*t == '\"') return NULL;
- }
+ else if (*t == '\"')
+ return NULL;
+
return NULL;
}
Returns: new character pointer
*/
-static uschar *
-skip_comment(uschar *s)
+static const uschar *
+skip_comment(const uschar *s)
{
last_comment_position = s;
while (*s)
in []. Make sure the output is set to the null string if there is a syntax
error as well as if there is no domain at all.
+Optionally, msg_id domain literals ( printable-ascii enclosed in [] )
+are permitted.
+
Arguments:
s current character pointer
t where to put the domain
+ msg_id_literals flag for relaxed domain-literal processing
errorptr put error message here on failure (*t will be 0 on exit)
Returns: new character pointer
*/
-static uschar *
-read_domain(uschar *s, uschar *t, uschar **errorptr)
+static const uschar *
+read_domain(const uschar *s, uschar *t, BOOL msg_id_literals, uschar **errorptr)
{
uschar *tt = t;
s = skip_comment(s);
t += 5;
s += 5;
}
- while (*s == '.' || *s == ':' || isxdigit(*s)) *t++ = *s++;
+
+ if (msg_id_literals)
+ while (*s >= 33 && *s <= 90 || *s >= 94 && *s <= 126) *t++ = *s++;
+ else
+ while (*s == '.' || *s == ':' || isxdigit(*s)) *t++ = *s++;
if (*s == ']') *t++ = *s++; else
{
*tt = 0;
}
- if (!allow_domain_literals)
+ if (!allow_domain_literals && !msg_id_literals)
{
*errorptr = US"domain literals not allowed";
*tt = 0;
Returns: new character pointer
*/
-static uschar *
-read_local_part(uschar *s, uschar *t, uschar **error, BOOL allow_null)
+static const uschar *
+read_local_part(const uschar *s, uschar *t, uschar **error, BOOL allow_null)
{
uschar *tt = t;
*error = NULL;
Returns: new character pointer
*/
-static uschar *
-read_route(uschar *s, uschar *t, uschar **errorptr)
+static const uschar *
+read_route(const uschar *s, uschar *t, uschar **errorptr)
{
BOOL commas = FALSE;
*errorptr = NULL;
while (*s == '@')
{
*t++ = '@';
- s = read_domain(s+1, t, errorptr);
+ s = read_domain(s+1, t, FALSE, errorptr);
if (*t == 0) return s;
t += Ustrlen((const uschar *)t);
if (*s != ',') break;
Returns: new character pointer
*/
-static uschar *
-read_addr_spec(uschar *s, uschar *t, int term, uschar **errorptr,
+static const uschar *
+read_addr_spec(const uschar *s, uschar *t, int term, uschar **errorptr,
uschar **domainptr)
{
s = read_local_part(s, t, errorptr, FALSE);
t += Ustrlen((const uschar *)t);
*t++ = *s++;
*domainptr = t;
- s = read_domain(s, t, errorptr);
+ s = read_domain(s, t, FALSE, errorptr);
}
return s;
}
#define FAILED(s) { *errorptr = s; goto PARSE_FAILED; }
uschar *
-parse_extract_address(uschar *mailbox, uschar **errorptr, int *start, int *end,
+parse_extract_address(const uschar *mailbox, uschar **errorptr, int *start, int *end,
int *domain, BOOL allow_null)
{
uschar *yield = store_get(Ustrlen(mailbox) + 1, is_tainted(mailbox));
-uschar *startptr, *endptr;
-uschar *s = US mailbox;
+const uschar *startptr, *endptr;
+const uschar *s = US mailbox;
uschar *t = US yield;
*domain = 0;
{
if (*s == 0 || *s == ';')
{
- if (*t == 0) FAILED(US"empty address");
+ if (!*t) FAILED(US"empty address");
endptr = last_comment_position;
goto PARSE_SUCCEEDED; /* Bare local part */
}
}
endptr = s;
- if (*errorptr != NULL) goto PARSE_FAILED;
+ if (*errorptr) goto PARSE_FAILED;
while (bracket_count-- > 0) if (*s++ != '>')
{
*errorptr = s[-1] == 0
not enclosed in <> as well, which is indicated by an empty first local
part preceding '@'. The source routing is, however, ignored. */
-else if (*t == 0)
+else if (!*t)
{
uschar *domainptr = yield;
s = read_route(s, t, errorptr);
- if (*errorptr != NULL) goto PARSE_FAILED;
+ if (*errorptr) goto PARSE_FAILED;
*t = 0; /* Ensure route is ignored - probably overkill */
s = read_addr_spec(s, t, 0, errorptr, &domainptr);
- if (*errorptr != NULL) goto PARSE_FAILED;
+ if (*errorptr) goto PARSE_FAILED;
*domain = domainptr - yield;
endptr = last_comment_position;
if (*domain == 0) FAILED(US"domain missing in source-routed address");
t += Ustrlen((const uschar *)t);
*t++ = *s++;
*domain = t - yield;
- s = read_domain(s, t, errorptr);
- if (*t == 0) goto PARSE_FAILED;
+ s = read_domain(s, t, TRUE, errorptr);
+ if (!*t) goto PARSE_FAILED;
endptr = last_comment_position;
}
move it back past white space if necessary. */
PARSE_SUCCEEDED:
-if (*s != 0)
+if (*s)
{
if (f.parse_found_group && *s == ';')
{
*end = endptr - US mailbox;
/* Although this code has no limitation on the length of address extracted,
-other parts of Exim may have limits, and in any case, RFC 2821 limits local
-parts to 64 and domains to 255, so we do a check here, giving an error if the
-address is ridiculously long. */
+other parts of Exim may have limits, and in any case, RFC 5321 limits email
+addresses to 256, so we do a check here, giving an error if the address is
+ridiculously long. */
-if (*end - *start > ADDRESS_MAXLENGTH)
+if (*end - *start > EXIM_EMAILADDR_MAX)
{
*errorptr = string_sprintf("address is ridiculously long: %.64s...", yield);
return NULL;
*/
const uschar *
-parse_quote_2047(const uschar *string, int len, uschar *charset, BOOL fold)
+parse_quote_2047(const uschar *string, int len, const uschar *charset,
+ BOOL fold)
{
const uschar * s = string;
int hlen, l;
{ g = string_catn(g, s, 1); first_byte = FALSE; }
}
-g = string_catn(g, US"?=", 2);
-return coded ? string_from_gstring(g) : string;
+if (coded)
+ string = string_from_gstring(g = string_catn(g, US"?=", 2));
+else
+ g->ptr = -1;
+
+gstring_release_unused(g);
+return string;
}
/* No non-printers; use the RFC 822 quoting rules */
-buffer = store_get(len*4, is_tainted(phrase));
+if (len <= 0 || len >= INT_MAX/4)
+ {
+ return string_copy_taint(CUS"", is_tainted(phrase));
+ }
+
+buffer = store_get((len+1)*4, is_tainted(phrase));
s = phrase;
end = s + len;
{
if (ss >= end) ss--;
*t++ = '(';
- Ustrncpy(t, s, ss-s);
- t += ss-s;
- s = ss;
+ if (ss > s)
+ {
+ Ustrncpy(t, s, ss-s);
+ t += ss-s;
+ s = ss;
+ }
}
}
*/
int
-parse_forward_list(uschar *s, int options, address_item **anchor,
+parse_forward_list(const uschar *s, int options, address_item **anchor,
uschar **error, const uschar *incoming_domain, uschar *directory,
error_block **syntax_errors)
{
int special = 0;
int specopt = 0;
int specbit = 0;
- uschar *ss, *nexts;
+ const uschar *ss, *nexts;
address_item *addr;
BOOL inquote = FALSE;
syntax error has been skipped. I now think it is the wrong approach, but
have left this here just in case, and for the record. */
- #ifdef NEVER
+#ifdef NEVER
if (count > 0) return FF_DELIVERED; /* Something was generated */
if (syntax_errors == NULL || /* Not skipping syntax errors, or */
*error = string_sprintf("no addresses generated: syntax error in %s: %s",
(*syntax_errors)->text2, (*syntax_errors)->text1);
return FF_ERROR;
- #endif
-
+#endif
}
/* Find the end of the next address. Quoted strings in addresses may contain
len = ss - s;
- DEBUG(D_route)
- {
- int save = s[len];
- s[len] = 0;
- debug_printf("extract item: %s\n", s);
- s[len] = save;
- }
+ DEBUG(D_route) debug_printf("extract item: %.*s\n", len, s);
/* Handle special addresses if permitted. If the address is :unknown:
ignore it - this is for backward compatibility with old alias files. You
else if (Ustrncmp(s, ":fail:", 6) == 0)
{ special = FF_FAIL; specopt = RDO_FAIL; } /* specbit is 0 */
- if (special != 0)
+ if (special)
{
uschar *ss = Ustrchr(s+1, ':') + 1;
if ((options & specopt) == specbit)
*error = string_sprintf("\"%.*s\" is not permitted", len, s);
return FF_ERROR;
}
- while (*ss != 0 && isspace(*ss)) ss++;
- while (s[len] != 0 && s[len] != '\n') len++;
- s[len] = 0;
- *error = string_copy(ss);
+ while (*ss && isspace(*ss)) ss++;
+ while (s[len] && s[len] != '\n') len++;
+ *error = string_copyn(ss, s + len - ss);
return special;
}
{
uschar *filebuf;
uschar filename[256];
- uschar *t = s+9;
+ const uschar * t = s+9;
int flen = len - 9;
int frc;
struct stat statbuf;
return FF_ERROR;
}
- if (is_tainted(filename))
- {
- *error = string_sprintf("Tainted name '%s' for included file not permitted\n",
- filename);
+ if ((*error = is_tainted2(filename, 0, "Tainted name '%s' for included file not permitted\n", filename)))
return FF_ERROR;
- }
/* Check file name if required */
else
{
int start, end, domain;
- uschar *recipient = NULL;
- int save = s[len];
- s[len] = 0;
+ const uschar *recipient = NULL;
+ uschar * s_ltd = string_copyn(s, len);
/* If it starts with \ and the rest of it parses as a valid mail address
without a domain, carry on with that address, but qualify it with the
incoming domain. Otherwise arrange for the address to fall through,
causing an error message on the re-parse. */
- if (*s == '\\')
+ if (*s_ltd == '\\')
{
recipient =
- parse_extract_address(s+1, error, &start, &end, &domain, FALSE);
+ parse_extract_address(s_ltd+1, error, &start, &end, &domain, FALSE);
if (recipient)
recipient = domain != 0 ? NULL :
string_sprintf("%s@%s", recipient, incoming_domain);
/* Try parsing the item as an address. */
if (!recipient) recipient =
- parse_extract_address(s, error, &start, &end, &domain, FALSE);
+ parse_extract_address(s_ltd, error, &start, &end, &domain, FALSE);
/* If item starts with / or | and is not a valid address, or there
is no domain, treat it as a file or pipe. If it was a quoted item,
remove the quoting occurrences of \ within it. */
- if ((*s == '|' || *s == '/') && (recipient == NULL || domain == 0))
+ if ((*s_ltd == '|' || *s_ltd == '/') && (recipient == NULL || domain == 0))
{
- uschar *t = store_get(Ustrlen(s) + 1, is_tainted(s));
+ uschar *t = store_get(Ustrlen(s_ltd) + 1, is_tainted(s_ltd));
uschar *p = t;
- uschar *q = s;
+ uschar *q = s_ltd;
while (*q != 0)
{
if (inquote)
*p = 0;
addr = deliver_make_addr(t, TRUE);
setflag(addr, af_pfr); /* indicates pipe/file/reply */
- if (*s != '|') setflag(addr, af_file); /* indicates file */
+ if (*s_ltd != '|') setflag(addr, af_file); /* indicates file */
}
/* Item must be an address. Complain if not, else qualify, rewrite and set
else
{
- if (recipient == NULL)
+ if (!recipient)
{
if (Ustrcmp(*error, "empty address") == 0)
{
*error = NULL;
- s[len] = save;
s = nexts;
continue;
}
- if (syntax_errors != NULL)
+ if (syntax_errors)
{
error_block *e = store_get(sizeof(error_block), FALSE);
error_block *last = *syntax_errors;
- if (last == NULL) *syntax_errors = e; else
+ if (!last) *syntax_errors = e; else
{
- while (last->next != NULL) last = last->next;
+ while (last->next) last = last->next;
last->next = e;
}
e->next = NULL;
e->text1 = *error;
- e->text2 = string_copy(s);
- s[len] = save;
+ e->text2 = s_ltd;
s = nexts;
continue;
}
else
{
- *error = string_sprintf("%s in \"%s\"", *error, s);
- s[len] = save; /* _after_ using it for *error */
+ *error = string_sprintf("%s in \"%s\"", *error, s_ltd);
return FF_ERROR;
}
}
recipient = ((options & RDO_REWRITE) != 0)?
rewrite_address(recipient, TRUE, FALSE, global_rewrite_rules,
rewrite_existflags) :
- rewrite_address_qualify(recipient, TRUE);
- addr = deliver_make_addr(recipient, TRUE); /* TRUE => copy recipient */
+ rewrite_address_qualify(recipient, TRUE); /*XXX loses track of const */
+ addr = deliver_make_addr(US recipient, TRUE); /* TRUE => copy recipient, so deconst ok */
}
- /* Restore the final character in the original data, and add to the
- output chain. */
+ /* Add the original data to the output chain. */
- s[len] = save;
addr->next = *anchor;
*anchor = addr;
count++;
Returns: points after the processed message-id or NULL on error
*/
-uschar *
-parse_message_id(uschar *str, uschar **yield, uschar **error)
+const uschar *
+parse_message_id(const uschar *str, uschar **yield, uschar **error)
{
uschar *domain = NULL;
uschar *id;
*id++ = 0;
store_release_above(id);
-str = skip_comment(str);
-return str;
+return skip_comment(str);
}
Returns: points after the processed date or NULL on error
*/
-static uschar *
-parse_number(uschar *str, int *n, int digits)
+static const uschar *
+parse_number(const uschar *str, int *n, int digits)
{
- *n=0;
- while (digits--)
+*n=0;
+while (digits--)
{
- if (*str<'0' || *str>'9') return NULL;
- *n=10*(*n)+(*str++-'0');
+ if (*str<'0' || *str>'9') return NULL;
+ *n=10*(*n)+(*str++-'0');
}
- return str;
+return str;
}
Returns: points after the parsed day or NULL on error
*/
-static uschar *
-parse_day_of_week(uschar *str)
+static const uschar *
+parse_day_of_week(const uschar * str)
{
/*
day-of-week = ([FWS] day-name) / obs-day-of-week
int i;
uschar day[4];
-str=skip_comment(str);
-for (i=0; i<3; ++i)
+str = skip_comment(str);
+for (i = 0; i < 3; ++i)
{
- if ((day[i]=tolower(*str))=='\0') return NULL;
+ if ((day[i] = tolower(*str)) == '\0') return NULL;
++str;
}
-day[3]='\0';
-for (i=0; i<7; ++i) if (Ustrcmp(day,day_name[i])==0) break;
-if (i==7) return NULL;
-str=skip_comment(str);
-return str;
+day[3] = '\0';
+for (i = 0; i<7; ++i) if (Ustrcmp(day,day_name[i]) == 0) break;
+if (i == 7) return NULL;
+return skip_comment(str);
}
Returns: points after the processed date or NULL on error
*/
-static uschar *
-parse_date(uschar *str, int *d, int *m, int *y)
+static const uschar *
+parse_date(const uschar *str, int *d, int *m, int *y)
{
/*
date = day month year
obs-day = [CFWS] 1*2DIGIT [CFWS]
*/
-uschar *c,*n;
+const uschar * s, * n;
static const uschar *month_name[]={ US"jan", US"feb", US"mar", US"apr", US"may", US"jun", US"jul", US"aug", US"sep", US"oct", US"nov", US"dec" };
int i;
uschar month[4];
-str=skip_comment(str);
-if ((str=parse_number(str,d,1))==NULL) return NULL;
-if (*str>='0' && *str<='9') *d=10*(*d)+(*str++-'0');
-c=skip_comment(str);
-if (c==str) return NULL;
-else str=c;
-for (i=0; i<3; ++i) if ((month[i]=tolower(*(str+i)))=='\0') return NULL;
-month[3]='\0';
-for (i=0; i<12; ++i) if (Ustrcmp(month,month_name[i])==0) break;
-if (i==12) return NULL;
+str = skip_comment(str);
+if ((str = parse_number(str,d,1)) == NULL) return NULL;
+
+if (*str>='0' && *str<='9') *d = 10*(*d)+(*str++-'0');
+s = skip_comment(str);
+if (s == str) return NULL;
+str = s;
+
+for (i = 0; i<3; ++i) if ((month[i]=tolower(*(str+i))) == '\0') return NULL;
+month[3] = '\0';
+for (i = 0; i<12; ++i) if (Ustrcmp(month,month_name[i]) == 0) break;
+if (i == 12) return NULL;
str+=3;
-*m=i;
-c=skip_comment(str);
-if (c==str) return NULL;
-else str=c;
-if ((n=parse_number(str,y,4)))
+*m = i;
+s = skip_comment(str);
+if (s == str) return NULL;
+str=s;
+
+if ((n = parse_number(str,y,4)))
{
- str=n;
+ str = n;
if (*y<1900) return NULL;
- *y=*y-1900;
+ *y = *y-1900;
}
-else if ((n=parse_number(str,y,2)))
+else if ((n = parse_number(str,y,2)))
{
- str=skip_comment(n);
- while (*(str-1)==' ' || *(str-1)=='\t') --str; /* match last FWS later */
+ str = skip_comment(n);
+ while (*(str-1) == ' ' || *(str-1) == '\t') --str; /* match last FWS later */
if (*y<50) *y+=100;
}
else return NULL;
Returns: points after the processed time or NULL on error
*/
-static uschar *
-parse_time(uschar *str, int *h, int *m, int *s, int *z)
+static const uschar *
+parse_time(const uschar *str, int *h, int *m, int *s, int *z)
{
/*
time = time-of-day FWS zone
%d107-122 ; upper and lower case
*/
-uschar *c;
+const uschar * c;
-str=skip_comment(str);
-if ((str=parse_number(str,h,2))==NULL) return NULL;
-str=skip_comment(str);
+str = skip_comment(str);
+if ((str = parse_number(str,h,2)) == NULL) return NULL;
+str = skip_comment(str);
if (*str!=':') return NULL;
++str;
-str=skip_comment(str);
-if ((str=parse_number(str,m,2))==NULL) return NULL;
-c=skip_comment(str);
-if (*str==':')
+str = skip_comment(str);
+if ((str = parse_number(str,m,2)) == NULL) return NULL;
+c = skip_comment(str);
+if (*str == ':')
{
++str;
- str=skip_comment(str);
- if ((str=parse_number(str,s,2))==NULL) return NULL;
- c=skip_comment(str);
+ str = skip_comment(str);
+ if ((str = parse_number(str,s,2)) == NULL) return NULL;
+ c = skip_comment(str);
}
-if (c==str) return NULL;
+if (c == str) return NULL;
else str=c;
-if (*str=='+' || *str=='-')
+if (*str == '+' || *str == '-')
{
int neg;
- neg=(*str=='-');
+ neg = (*str == '-');
++str;
- if ((str=parse_number(str,z,4))==NULL) return NULL;
- *z=(*z/100)*3600+(*z%100)*60;
- if (neg) *z=-*z;
+ if ((str = parse_number(str,z,4)) == NULL) return NULL;
+ *z = (*z/100)*3600+(*z%100)*60;
+ if (neg) *z = -*z;
}
else
{
char zone[5];
- struct { const char *name; int off; } zone_name[10]=
+ struct { const char *name; int off; } zone_name[10] =
{ {"gmt",0}, {"ut",0}, {"est",-5}, {"edt",-4}, {"cst",-6}, {"cdt",-5}, {"mst",-7}, {"mdt",-6}, {"pst",-8}, {"pdt",-7}};
int i,j;
- for (i=0; i<4; ++i)
+ for (i = 0; i<4; ++i)
{
- zone[i]=tolower(*(str+i));
+ zone[i] = tolower(*(str+i));
if (zone[i]<'a' || zone[i]>'z') break;
}
- zone[i]='\0';
- for (j=0; j<10 && strcmp(zone,zone_name[j].name); ++j);
+ zone[i] = '\0';
+ for (j = 0; j<10 && strcmp(zone,zone_name[j].name); ++j);
/* Besides zones named in the grammar, RFC 2822 says other alphabetic */
/* time zones should be treated as unknown offsets. */
if (j<10)
{
- *z=zone_name[j].off*3600;
+ *z = zone_name[j].off*3600;
str+=i;
}
else if (zone[0]<'a' || zone[1]>'z') return 0;
else
{
while ((*str>='a' && *str<='z') || (*str>='A' && *str<='Z')) ++str;
- *z=0;
+ *z = 0;
}
}
return str;
Returns: points after the processed date-time or NULL on error
*/
-uschar *
-parse_date_time(uschar *str, time_t *t)
+const uschar *
+parse_date_time(const uschar *str, time_t *t)
{
/*
date-time = [ day-of-week "," ] date FWS time [CFWS]
char **old_environ;
static char gmt0[]="TZ=GMT0";
static char *gmt_env[]={ gmt0, (char*)0 };
-uschar *try;
+const uschar * try;
-if ((try=parse_day_of_week(str)))
+if ((try = parse_day_of_week(str)))
{
- str=try;
+ str = try;
if (*str!=',') return 0;
++str;
}
-if ((str=parse_date(str,&tm.tm_mday,&tm.tm_mon,&tm.tm_year))==NULL) return NULL;
+if ((str = parse_date(str,&tm.tm_mday,&tm.tm_mon,&tm.tm_year)) == NULL) return NULL;
if (*str!=' ' && *str!='\t') return NULL;
-while (*str==' ' || *str=='\t') ++str;
-if ((str=parse_time(str,&tm.tm_hour,&tm.tm_min,&tm.tm_sec,&zone))==NULL) return NULL;
-tm.tm_isdst=0;
-old_environ=environ;
-environ=gmt_env;
-*t=mktime(&tm);
-environ=old_environ;
-if (*t==-1) return NULL;
+while (*str == ' ' || *str == '\t') ++str;
+if ((str = parse_time(str,&tm.tm_hour,&tm.tm_min,&tm.tm_sec,&zone)) == NULL) return NULL;
+tm.tm_isdst = 0;
+old_environ = environ;
+environ = gmt_env;
+*t = mktime(&tm);
+environ = old_environ;
+if (*t == -1) return NULL;
*t-=zone;
-str=skip_comment(str);
-return str;
+return skip_comment(str);
}
int start, end, domain;
uschar buffer[1024];
+store_init();
big_buffer = store_malloc(big_buffer_size);
/* strip_trailing_dot = TRUE; */
git://git.exim.org / exim.git / blobdiff