Fixed qregex-20060423 (pristine is against netqmail)

diff -r 1510847ae5bf Makefile

--- a/Makefile	Thu Nov 01 16:23:16 2007 +0100

+++ b/Makefile	Thu Nov 01 16:24:02 2007 +0100

@@ -1534,16 +1534,16 @@ auto_split.h

 	./compile qmail-showctl.c

 qmail-smtpd: \

-load qmail-smtpd.o rcpthosts.o commands.o timeoutread.o \

+load qmail-smtpd.o rcpthosts.o qregex.o commands.o timeoutread.o \

 timeoutwrite.o ip.o ipme.o ipalloc.o control.o constmap.o received.o \

 date822fmt.o now.o qmail.o cdb.a fd.a wait.a datetime.a getln.a \

-open.a sig.a case.a env.a stralloc.a alloc.a substdio.a error.a str.a \

+open.a sig.a case.a env.a stralloc.a alloc.a strerr.a substdio.a error.a str.a \

 fs.a auto_qmail.o socket.lib

-	./load qmail-smtpd rcpthosts.o commands.o timeoutread.o \

+	./load qmail-smtpd qregex.o rcpthosts.o commands.o timeoutread.o \

 	timeoutwrite.o ip.o ipme.o ipalloc.o control.o constmap.o \

 	received.o date822fmt.o now.o qmail.o cdb.a fd.a wait.a \

 	datetime.a getln.a open.a sig.a case.a env.a stralloc.a \

-	alloc.a substdio.a error.a str.a fs.a auto_qmail.o  `cat \

+	alloc.a strerr.a substdio.a error.a str.a fs.a auto_qmail.o  `cat \

 	socket.lib`

 qmail-smtpd.0: \

@@ -1696,6 +1696,10 @@ compile rcpthosts.c cdb.h uint32.h byte.

 compile rcpthosts.c cdb.h uint32.h byte.h open.h error.h control.h \

 constmap.h stralloc.h gen_alloc.h rcpthosts.h

 	./compile rcpthosts.c

+

+qregex.o: \

+compile qregex.c qregex.h

+	./compile qregex.c

 readsubdir.o: \

 compile readsubdir.c readsubdir.h direntry.h fmt.h scan.h str.h \

diff -r 1510847ae5bf README.qregex

--- /dev/null	Thu Jan 01 00:00:00 1970 +0000

+++ b/README.qregex	Thu Nov 01 16:24:02 2007 +0100

@@ -0,0 +1,203 @@

+QREGEX (v2) 20060423 - README April 23, 2006

+A Regular Expression matching patch for qmail 1.03 and netqmail

+

+

+OVERVIEW:

+

+qregex adds the ability to match address evelopes via Regular Expressions (REs)

+in the qmail-smtpd process. It has the abiltiy to match `helo/ehlo` (host name),

+`mail from` (envelope sender), and `rcpt to` (envelope recipient) commands.

+It follows all the base rules that are set out with qmail (ie using control

+files) so it makes for easy integretion into an existing setup (see the

+install instructions for more info). The v2 is specified because qregex was

+re-written to better conform to the security guarantee set forth by the author

+of qmail. The original version used stdio.h and stdlib.h for reading the

+control files whereas v2 now uses all stralloc functions which are much more

+regulated against buffer overruns and the like.

+See: http://cr.yp.to/qmail/guarantee.html

+

+

+FEATURES:

+

+Features of qregex include:

+

+1. Performs pattern matching on envelope senders and envelope

+   recipients against REs in the badmailfrom and badmailto control

+   files. Two additional control files, badmailfromnorelay and

+   badmailtonorelay, are used for pattern matching when the 

+   RELAYCLIENT environment variable is not set.

+

+2. Performs pattern matching on the helo/ehlo host name. Setting the

+   NOBADHELO environment variable prevents the host name from being

+   compared to the patterns in the badhelo control file.

+

+3. Matches to patterns are logged. Setting the LOGREGEX environment

+   variable causes the matched regex pattern to be included in the log. 

+

+4. Matching is case insensitive.

+

+5. qregex ignores empty envelope senders. An empty envelope sender is not

+   compared to the patterns in the badmailfrom and badmailfromnorelay

+   control files and is always accepted.

+

+

+PLATFORMS:

+

+qregex has been built and tested on the following platforms. I'm sure it won't

+have any problems on any platform that qmail will run on (providing they have

+a regex interface) but if you run into problems let me know.

+

+	- OpenBSD 3.x

+	- FreeBSD 4.x, 5.x

+	- Mandrake Linux 9.x

+	- SuSE Linux 8.x

+

+

+

+INSTALLATION INSTRUCTIONS:

+

+Installation is very simple, there is only one requirement. You need to use the

+GNU version of the patch utility (http://www.gnu.org/software/patch/patch.html).

+(For Solaris 8 users it is installed as 'gpatch')

+

+- If this is a new setup.

+Unpack the qmail archive, cd into the qmail-1.03 directory and run

+"patch < /path/to/qregex-<version>.patch". Follow the instructions as per the

+included qmail INSTALL file. Once you are done come back to this file and read

+the section on the control files.

+

+If you are using netqmail, then unpack the netqmail archive. Run the collate.sh

+script and cd into the resulting netqmail-<version> directory. From there, run

+"patch < /path/to/qregex-<version>.patch". Complete the netqmail installation

+normally. Once you are done, come back to this file and read the section on the

+control files.

+

+- If this is an existing setup.

+FIRST: create your control files (see below).

+cd into your existing qmail or netqmail source directory. Run

+"patch < /path/to/qregex-<version>.patch" then "make qmail-smtpd". Now run

+./qmail-smtpd and test your new rules to make sure they work as expected.

+

+Install the new binary by cd'ing to /var/qmail/bin and as root (in one command)

+copy the existing binary to 'qmail-smtpd.old' and copy the new binary from the

+source directory to 'qmail-smtpd'.

+(ex. cp qmail-smtpd qmail-smtpd.old && cp ~/qmail-1.03/qmail-smtpd qmail-smtpd)

+

+You can also optionally just run "make setup check" as it will install the

+updated documentation and man pages provided with this patch. Stopping qmail

+before doing the "make setup check" is always a good idea.

+

+

+LOGGING:

+

+qregex will log matches to the patterns in the various control files. Log

+messages will take these three forms depending on which control file was

+matched:

+

+badhelo

+qmail-smtpd: badhelo: <host> at <remote IP>

+

+badmailfrom and badmailfromnorelay

+qmail-smtpd: badmailfrom: <sender address> at <remote IP>

+

+badmailto and badmailtonorelay

+qmail-smtpd: badmailto: <rcpt address> at <remote IP>

+

+When the LOGREGEX environment variable is set, the matched pattern will

+be included in the log. Log messages will have the regex pattern appended

+to them. For example, a badhelo log message will look like this:

+

+qmail-smtpd: badhelo: <host> at <remote IP> matches pattern: <regex>

+

+

+CONTROL FILES:

+

+qregex provides you with five control files. None of these control files

+is mandatory and you can use them in any combination you choose in your setup.

+

+The "control/badmailfrom" and "control/badmailto" files contain your REs for

+matching against the 'mail from' (envelope sender) and 'rcpt to' (envelope

+recipient) smtp commands respectively.

+The "control/badmailfromnorelay" and "control/badmailtonorelay" match against

+the same commands but are read only when the RELAYCLIENT environment variable

+is not set.

+The "control/badhelo" file matches against the 'helo/ehlo' smtp command.

+

+If you prefer you can symlink the badmailfrom and badmailto control files

+(ln -s badmailfrom badmailto) and maintain fewer sets of rules. Beware

+this might cause problems in certain setups.

+        

+	Here's an example "badhelo" file.

+	-----------------------------------

+	# block host strings with no dot (not a FQDN)

+	!\.

+	-----------------------------------

+	

+	An example "badmailfrom" file.

+	-----------------------------------

+	# this will drop everything containing the string

+	# bad.domain.com or Bad.Domain.Com or BAD.domain.COM

+	bad\.domain\.com

+	# force users to fully qualify themselves

+	# (i.e. deny "user", accept "user@domain")

+	!@

+	-----------------------------------

+

+	And "badmailto" (a little more interesting)

+	-----------------------------------

+	# must not contain invalid characters, brakets or multiple @'s

+	[!%#:*^(){}]

+	@.*@

+	-----------------------------------

+

+You can use the non-RE character '!' to start an RE as a signal to qregex to

+negate the action. As used above in the badmailfrom file, by negating the '@'

+symbol qregex will signal qmail-smtpd to deny the 'mail from' command whenever

+the address doesn't contain an @ symbol. When used inside a bracket expression,

+the '!' character looses this special meaning. This is shown in the badmailto

+example.

+

+The norelay control files follow the same rules as the other control files but

+are intended to address two specific scenarios.

+The badmailfromnorelay file can be used to block mail trying to spoof a domain

+hosted on your mail server. It prevents a mail client that is not allowed to

+relay email through your server from using one of your hosted domains as its

+envelope sender.

+The badmailtonorelay file can be used to create email addresses that cannot

+receive mail from any source not allowed to relay email through your server.

+This is handy for creating email addresses for use only within your own 

+domain(s) that can't receive spam from the world at large.

+

+

+INTERNALS:

+

+qregex (or regexmatch as the function is called) will be called during the

+`helo/ehlo`, `rcpt to` and `mail from` handling routines in "qmail-smtpd.c".

+When called, it will read the proper control file then one by one compile and

+execute the regex on the string passed into qmail-smtpd. If the regex matches

+it returns TRUE (1) and the qmail-smtpd process will deny the user the ability

+to continue. If you change anything and think it betters this patch please

+send me a new diff file so I can take a peek.

+

+

+CONTACT:

+qregex is maintained by:

+	Andrew St. Jean

+	andrew@arda.homeunix.net

+	www.arda.homeunix.net/store/qmail/

+

+Contributers to qregex:

+	Jeremy Kitchen	

+	kitchen at scriptkitchen dot com

+	http://www.scriptkitchen.com/qmail

+

+	Alex Pleiner

+	alex@zeitform.de

+	zeitform Internet Dienste

+	http://www.zeitform.de/

+

+	Thanos Massias

+

+Original qregex patch written by:

+	Evan Borgstrom

+	evan at unixpimps dot org

diff -r 1510847ae5bf TARGETS

--- a/TARGETS	Thu Nov 01 16:23:16 2007 +0100

+++ b/TARGETS	Thu Nov 01 16:24:02 2007 +0100

@@ -252,6 +252,7 @@ qmail-qmtpd

 qmail-qmtpd

 qmail-smtpd.o

 qmail-smtpd

+qregex.o

 sendmail.o

 sendmail

 tcp-env.o

diff -r 1510847ae5bf hier.c

--- a/hier.c	Thu Nov 01 16:23:16 2007 +0100

+++ b/hier.c	Thu Nov 01 16:24:02 2007 +0100

@@ -76,6 +76,7 @@ void hier()

   c(auto_qmail,"boot","binm3+df",auto_uido,auto_gidq,0755);

   c(auto_qmail,"doc","FAQ",auto_uido,auto_gidq,0644);

+  c(auto_qmail,"doc","README.qregex",auto_uido,auto_gidq,0644);

   c(auto_qmail,"doc","UPGRADE",auto_uido,auto_gidq,0644);

   c(auto_qmail,"doc","SENDMAIL",auto_uido,auto_gidq,0644);

   c(auto_qmail,"doc","INSTALL",auto_uido,auto_gidq,0644);

diff -r 1510847ae5bf install-big.c

--- a/install-big.c	Thu Nov 01 16:23:16 2007 +0100

+++ b/install-big.c	Thu Nov 01 16:24:02 2007 +0100

@@ -76,6 +76,7 @@ void hier()

   c(auto_qmail,"boot","binm3+df",auto_uido,auto_gidq,0755);

   c(auto_qmail,"doc","FAQ",auto_uido,auto_gidq,0644);

+  c(auto_qmail,"doc","README.qregex",auto_uido,auto_gidq,0644);

   c(auto_qmail,"doc","UPGRADE",auto_uido,auto_gidq,0644);

   c(auto_qmail,"doc","SENDMAIL",auto_uido,auto_gidq,0644);

   c(auto_qmail,"doc","INSTALL",auto_uido,auto_gidq,0644);

diff -r 1510847ae5bf qmail-control.9

--- a/qmail-control.9	Thu Nov 01 16:23:16 2007 +0100

+++ b/qmail-control.9	Thu Nov 01 16:24:02 2007 +0100

@@ -20,7 +20,11 @@ other hostname-related control files.

 Comments are allowed

 in

+.IR badhelo ,

 .IR badmailfrom ,

+.IR badmailfromnorelay ,

+.IR badmailto ,

+.IR badmailtonorelay ,

 .IR locals ,

 .IR percenthack ,

 .IR qmqpservers ,

@@ -40,7 +44,11 @@ See the corresponding man pages for furt

 .ta 5c 10c

 control	default	used by

+.I badhelo	\fR(none)	\fRqmail-smtpd

 .I badmailfrom	\fR(none)	\fRqmail-smtpd

+.I badmailfromnorelay	\fR(none)	\fRqmail-smtpd

+.I badmailto	\fR(none)	\fRqmail-smtpd

+.I badmailtonorelay	\fR(none)	\fRqmail-smtpd

 .I bouncefrom	\fRMAILER-DAEMON	\fRqmail-send

 .I bouncehost	\fIme	\fRqmail-send

 .I concurrencylocal	\fR10	\fRqmail-send

diff -r 1510847ae5bf qmail-showctl.c

--- a/qmail-showctl.c	Thu Nov 01 16:23:16 2007 +0100

+++ b/qmail-showctl.c	Thu Nov 01 16:24:02 2007 +0100

@@ -214,7 +214,11 @@ void main()

     _exit(111);

   }

-  do_lst("badmailfrom","Any MAIL FROM is allowed.",""," not accepted in MAIL FROM.");

+  do_lst("badhelo","Any HELO host name is allowed.",""," HELO host name denied if it matches this pattern.");

+  do_lst("badmailfrom","Any MAIL FROM is allowed.",""," MAIL FROM denied if it matches this pattern.");

+  do_lst("badmailfromnorelay","Any MAIL FROM is allowed.",""," MAIL FROM denied if it matches this pattern and RELAYCLIENT is not set.");

+  do_lst("badmailto","No RCPT TO are specifically denied.",""," RCPT TO denied if it matches this pattern.");

+  do_lst("badmailtonorelay","No RCPT TO are specifically denied.",""," RCPT TO denied if it matches this pattern and RELAYCLIENT is not set.");

   do_str("bouncefrom",0,"MAILER-DAEMON","Bounce user name is ");

   do_str("bouncehost",1,"bouncehost","Bounce host name is ");

   do_int("concurrencylocal","10","Local concurrency is ","");

@@ -267,7 +271,11 @@ void main()

     if (str_equal(d->d_name,"..")) continue;

     if (str_equal(d->d_name,"bouncefrom")) continue;

     if (str_equal(d->d_name,"bouncehost")) continue;

+    if (str_equal(d->d_name,"badhelo")) continue;

     if (str_equal(d->d_name,"badmailfrom")) continue;

+    if (str_equal(d->d_name,"badmailfromnorelay")) continue;

+    if (str_equal(d->d_name,"badmailto")) continue;

+    if (str_equal(d->d_name,"badmailtonorelay")) continue;

     if (str_equal(d->d_name,"bouncefrom")) continue;

     if (str_equal(d->d_name,"bouncehost")) continue;

     if (str_equal(d->d_name,"concurrencylocal")) continue;

diff -r 1510847ae5bf qmail-smtpd.8

--- a/qmail-smtpd.8	Thu Nov 01 16:23:16 2007 +0100

+++ b/qmail-smtpd.8	Thu Nov 01 16:24:02 2007 +0100

@@ -37,11 +37,26 @@ even though such messages violate the SM

 even though such messages violate the SMTP protocol.

 .SH "CONTROL FILES"

 .TP 5

+.I badhelo

+Unacceptable HELO/EHLO host names.

+.B qmail-smtpd

+will reject every recipient address for a message if

+the host name is listed in, 

+or matches a POSIX regular expression pattern listed in,

+.IR badhelo .

+If the 

+.B NOBADHELO 

+environment variable is set, then the contents of 

+.IR badhelo 

+will be ignored.

+For more information, please have a look at doc/README.qregex.

+.TP 5

 .I badmailfrom

 Unacceptable envelope sender addresses.

 .B qmail-smtpd

 will reject every recipient address for a message

-if the envelope sender address is listed in

+if the envelope sender address is listed in, or matches a POSIX regular expression

+pattern listed in,

 .IR badmailfrom .

 A line in

 .I badmailfrom

@@ -49,6 +64,32 @@ may be of the form

 .BR @\fIhost ,

 meaning every address at

 .IR host .

+For more information, please have a look at doc/README.qregex.

+.TP 5

+.I badmailfromnorelay

+Functions the same as the

+.IR badmailfrom

+control file but is read only if the 

+.B RELAYCLIENT 

+environment variable is not set.

+For more information, please have a look at doc/README.qregex.

+.TP 5

+.I badmailto

+Unacceptable envelope recipient addresses.

+.B qmail-smtpd

+will reject every recipient address for a message if the recipient address

+is listed in,

+or matches a POSIX regular expression pattern listed in,

+.IR badmailto .

+For more information, please have a look at doc/README.qregex.

+.TP 5

+.I badmailtonorelay

+Functions the same as the

+.IR badmailto

+control file but is read only if the

+.B RELAYCLIENT

+environment variable is not set.

+For more information, please have a look at doc/README.qregex.

 .TP 5

 .I databytes

 Maximum number of bytes allowed in a message,

diff -r 1510847ae5bf qmail-smtpd.c

--- a/qmail-smtpd.c	Thu Nov 01 16:23:16 2007 +0100

+++ b/qmail-smtpd.c	Thu Nov 01 16:24:02 2007 +0100

@@ -23,6 +23,15 @@

 #include "timeoutread.h"

 #include "timeoutwrite.h"

 #include "commands.h"

+#include "qregex.h"

+#include "strerr.h"

+

+#define BMCHECK_BMF 0

+#define BMCHECK_BMFNR 1

+#define BMCHECK_BMT 2

+#define BMCHECK_BMTNR 3

+#define BMCHECK_BHELO 4

+    

 #define MAXHOPS 100

 unsigned int databytes = 0;

@@ -49,7 +58,9 @@ void die_ipme() { out("421 unable to fig

 void die_ipme() { out("421 unable to figure out my IP addresses (#4.3.0)\r\n"); flush(); _exit(1); }

 void straynewline() { out("451 See http://pobox.com/~djb/docs/smtplf.html.\r\n"); flush(); _exit(1); }

-void err_bmf() { out("553 sorry, your envelope sender is in my badmailfrom list (#5.7.1)\r\n"); }

+void err_bmf() { out("553 sorry, your envelope sender has been denied (#5.7.1)\r\n"); }

+void err_bmt() { out("553 sorry, your envelope recipient has been denied (#5.7.1)\r\n"); }

+void err_bhelo() { out("553 sorry, your HELO host name has been denied (#5.7.1)\r\n"); }

 void err_nogateway() { out("553 sorry, that domain isn't in my list of allowed rcpthosts (#5.7.1)\r\n"); }

 void err_unimpl() { out("502 unimplemented (#5.5.1)\r\n"); }

 void err_syntax() { out("555 syntax error (#5.5.4)\r\n"); }

@@ -93,9 +104,24 @@ void dohelo(arg) char *arg; {

 int liphostok = 0;

 stralloc liphost = {0};

+

 int bmfok = 0;

 stralloc bmf = {0};

-struct constmap mapbmf;

+

+int bmfnrok = 0;

+stralloc bmfnr = {0};

+

+int bmtok = 0;

+stralloc bmt = {0};

+

+int bmtnrok = 0;

+stralloc bmtnr = {0};

+

+int bhelook = 0;

+stralloc bhelo = {0};

+

+int logregex = 0;

+stralloc matchedregex = {0};

 void setup()

 {

@@ -114,8 +140,21 @@ void setup()

   bmfok = control_readfile(&bmf,"control/badmailfrom",0);

   if (bmfok == -1) die_control();

-  if (bmfok)

-    if (!constmap_init(&mapbmf,bmf.s,bmf.len,0)) die_nomem();

+

+  bmfnrok = control_readfile(&bmfnr,"control/badmailfromnorelay",0);

+  if (bmfnrok == -1) die_control();

+

+  bmtok = control_readfile(&bmt,"control/badmailto",0);

+  if (bmtok == -1) die_control();

+

+  bmtnrok = control_readfile(&bmtnr,"control/badmailtonorelay",0);

+  if (bmtnrok == -1) die_control();

+

+  bhelook = control_readfile(&bhelo, "control/badhelo",0);

+  if (bhelook == -1) die_control();

+  if (env_get("NOBADHELO")) bhelook = 0;

+

+  if (env_get("LOGREGEX")) logregex = 1;

   if (control_readint(&databytes,"control/databytes") == -1) die_control();

   x = env_get("DATABYTES");

@@ -197,14 +236,56 @@ char *arg;

   return 1;

 }

-int bmfcheck()

-{

-  int j;

-  if (!bmfok) return 0;

-  if (constmap(&mapbmf,addr.s,addr.len - 1)) return 1;

-  j = byte_rchr(addr.s,addr.len,'@');

-  if (j < addr.len)

-    if (constmap(&mapbmf,addr.s + j,addr.len - j - 1)) return 1;

+int bmcheck(which) int which;

+{

+  int i = 0;

+  int j = 0;

+  int x = 0;

+  int negate = 0;

+  static stralloc bmb = {0};

+  static stralloc curregex = {0};

+

+  if (which == BMCHECK_BMF) {

+    if (!stralloc_copy(&bmb,&bmf)) die_nomem();

+  } else if (which == BMCHECK_BMFNR) {

+    if (!stralloc_copy(&bmb,&bmfnr)) die_nomem();

+  } else if (which == BMCHECK_BMT) {

+    if (!stralloc_copy(&bmb,&bmt)) die_nomem();

+  } else if (which == BMCHECK_BMTNR) {

+    if (!stralloc_copy(&bmb,&bmtnr)) die_nomem();

+  } else if (which == BMCHECK_BHELO) {

+    if (!stralloc_copy(&bmb,&bhelo)) die_nomem();

+  } else {

+    die_control();

+  }

+

+  while (j < bmb.len) {

+    i = j;

+    while ((bmb.s[i] != '\0') && (i < bmb.len)) i++;

+    if (bmb.s[j] == '!') {

+      negate = 1;

+      j++;

+    }

+    if (!stralloc_copyb(&curregex,bmb.s + j,(i - j))) die_nomem();

+    if (!stralloc_0(&curregex)) die_nomem();

+    if (which == BMCHECK_BHELO) {

+      x = matchregex(helohost.s, curregex.s);

+    } else {

+      x = matchregex(addr.s, curregex.s);

+    }

+    if ((negate) && (x == 0)) {

+      if (!stralloc_copyb(&matchedregex,bmb.s + j - 1,(i - j + 1))) die_nomem();

+      if (!stralloc_0(&matchedregex)) die_nomem();      

+      return 1;