README revision 261194
1# Copyright (c) 1998-2004 Proofpoint, Inc. and its suppliers. 2# All rights reserved. 3# Copyright (c) 1983, 1995-1997 Eric P. Allman. All rights reserved. 4# Copyright (c) 1988 5# The Regents of the University of California. All rights reserved. 6# 7# By using this file, you agree to the terms and conditions set 8# forth in the LICENSE file which can be found at the top level of 9# the sendmail distribution. 10# 11# 12# $Id: README,v 8.393 2013/11/22 20:51:54 ca Exp $ 13# 14 15This directory contains the source files for sendmail(TM). 16 17 ******************************************************************* 18 !! Read sendmail/SECURITY for important installation information !! 19 ******************************************************************* 20 21 ********************************************************** 22 ** Read below for more details on building sendmail. ** 23 ********************************************************** 24 25************************************************************************** 26** IMPORTANT: Read the appropriate paragraphs in the section on ** 27** ``Operating System and Compile Quirks''. ** 28************************************************************************** 29 30For detailed instructions, please read the document ../doc/op/op.me: 31 32 cd ../doc/op ; make op.ps op.txt 33 34Sendmail is a trademark of Proofpoint, Inc. 35US Patent Numbers 6865671, 6986037. 36 37 38+-------------------+ 39| BUILDING SENDMAIL | 40+-------------------+ 41 42By far, the easiest way to compile sendmail is to use the "Build" 43script: 44 45 sh ./Build 46 47This uses the "uname" command to figure out what architecture you are 48on and creates a proper Makefile accordingly. It also creates a 49subdirectory per object format, so that multiarchitecture support is 50easy. In general this should be all you need. IRIX 6.x users should 51read the note below in the OPERATING SYSTEM AND COMPILE QUIRKS section. 52 53If you need to look at other include or library directories, use the 54-I or -L flags on the command line, e.g., 55 56 sh ./Build -I/usr/sww/include -L/usr/sww/lib 57 58It's also possible to create local site configuration in the file 59site.config.m4 (or another file settable with the -f flag). This 60file contains M4 definitions for various compilation values; the 61most useful are: 62 63confMAPDEF -D flags to specify database types to be included 64 (see below) 65confENVDEF -D flags to specify other environment information 66confINCDIRS -I flags for finding include files during compilation 67confLIBDIRS -L flags for finding libraries during linking 68confLIBS -l flags for selecting libraries during linking 69confLDOPTS other ld(1) linker options 70 71Others can be found by examining Makefile.m4. Please read 72../devtools/README for more information about the site.config.m4 73file. 74 75You can recompile from scratch using the -c flag with the Build 76command. This removes the existing compilation directory for the 77current platform and builds a new one. The -c flag must also 78be used if any site.*.m4 file in devtools/Site/ is changed. 79 80Porting to a new Unix-based system should be a matter of creating 81an appropriate configuration file in the devtools/OS/ directory. 82 83 84+----------------------+ 85| DATABASE DEFINITIONS | 86+----------------------+ 87 88There are several database formats that can be used for the alias files 89and for general maps. When used for alias files they interact in an 90attempt to be backward compatible. 91 92The options are: 93 94NEWDB The new Berkeley DB package. Some systems (e.g., BSD/OS and 95 Digital UNIX 4.0) have some version of this package 96 pre-installed. If your system does not have Berkeley DB 97 pre-installed, or the version installed is not version 2.0 98 or greater (e.g., is Berkeley DB 1.85 or 1.86), get the 99 current version from http://www.sleepycat.com/. DO NOT 100 use a version from any of the University of California, 101 Berkeley "Net" or other distributions. If you are still 102 running BSD/386 1.x, you will need to upgrade the included 103 Berkeley DB library to a current version. NEWDB is included 104 automatically if the Build script can find a library named 105 libdb.a or libdb.so. 106 See also OPERATING SYSTEM AND COMPILE QUIRKS about Berkeley 107 DB versions, e.g., DB 4.1.x. 108NDBM The older NDBM implementation -- the very old V7 DBM 109 implementation is no longer supported. 110NIS Network Information Services. To use this you must have 111 NIS support on your system. 112NISPLUS NIS+ (the revised NIS released with Solaris 2). You must 113 have NIS+ support on your system to use this flag. 114HESIOD Support for Hesiod (from the DEC/Athena distribution). You 115 must already have Hesiod support on your system for this to 116 work. You may be able to get this to work with the MIT/Athena 117 version of Hesiod, but that's likely to be a lot of work. 118 BIND 8.X also includes Hesiod support. 119LDAPMAP Lightweight Directory Access Protocol support. You will 120 have to install the UMich or OpenLDAP 121 (http://www.openldap.org/) ldap and lber libraries to use 122 this flag. 123MAP_REGEX Regular Expression support. You will need to use an 124 operating system which comes with the POSIX regex() 125 routines or install a regexp library such as libregex from 126 the Free Software Foundation. 127DNSMAP DNS map support. Requires NAMED_BIND. 128PH_MAP PH map support. You will need the libphclient library from 129 the nph package (http://www-dev.cites.uiuc.edu/ph/nph/). 130MAP_NSD nsd map support (IRIX 6.5 and later). 131SOCKETMAP Support for a trivial query protocol over UNIX domain or TCP 132 sockets. 133 134>>> NOTE WELL for NEWDB support: If you want to get ndbm support, for 135>>> Berkeley DB versions under 2.0, it is CRITICAL that you remove 136>>> ndbm.o from libdb.a before you install it and DO NOT install ndbm.h; 137>>> for Berkeley DB versions 2.0 through 2.3.14, remove dbm.o from libdb.a 138>>> before you install it. If you don't delete these, there is absolutely 139>>> no point to including -DNDBM, since it will just get you another 140>>> (inferior) API to the same format database. These files OVERRIDE 141>>> calls to ndbm routines -- in particular, if you leave ndbm.h in, 142>>> you can find yourself using the new db package even if you don't 143>>> define NEWDB. Berkeley DB versions later than 2.3.14 do not need 144>>> to be modified. Please also consult the README in the top level 145>>> directory of the sendmail distribution for other important information. 146>>> 147>>> Further note: DO NOT remove your existing /usr/include/ndbm.h -- 148>>> you need that one. But do not install an updated ndbm.h in 149>>> /usr/include, /usr/local/include, or anywhere else. 150 151If NEWDB and NDBM are defined (but not NIS), then sendmail will read 152NDBM format alias files, but the next time a newaliases is run the 153format will be converted to NEWDB; that format will be used forever 154more. This is intended as a transition feature. 155 156If NEWDB, NDBM, and NIS are all defined and the name of the file includes 157the string "/yp/", sendmail will rebuild BOTH the NEWDB and NDBM format 158alias files. However, it will only read the NEWDB file; the NDBM format 159file is used only by the NIS subsystem. This is needed because the NIS 160maps on an NIS server are built directly from the NDBM files. 161 162If NDBM and NIS are defined (regardless of the definition of NEWDB), 163and the filename includes the string "/yp/", sendmail adds the special 164tokens "YP_LAST_MODIFIED" and "YP_MASTER_NAME", both of which are 165required if the NDBM file is to be used as an NIS map. 166 167All of these flags are normally defined in a confMAPDEF setting in your 168site.config.m4. 169 170If you define NEWDB or HESIOD you get the User Database (USERDB) 171automatically. Generally you do want to have NEWDB for it to do 172anything interesting. See above for getting the Berkeley DB 173package (i.e., NEWDB). There is no separate "user database" 174package -- don't bother searching for it on the net. 175 176Hesiod and LDAP require libraries that may not be installed with your 177system. These are outside of my ability to provide support. See the 178"Quirks" section for more information. 179 180The regex map can be used to see if an address matches a certain regular 181expression. For example, all-numerics local parts are common spam 182addresses, so "^[0-9]+$" would match this. By using such a map in a 183check_* rule-set, you can block a certain range of addresses that would 184otherwise be considered valid. 185 186The socket map uses a simple request/reply protocol over TCP or 187UNIX domain sockets to query an external server. Both requests and 188replies are text based and encoded as netstrings. The socket map 189uses the same syntax as milters the specify the remote endpoint, 190e.g.: 191 192Ksocket mySocketMap inet:12345@127.0.0.1 193 194See doc/op/op.me for details. 195 196+---------------+ 197| COMPILE FLAGS | 198+---------------+ 199 200Wherever possible, I try to make sendmail pull in the correct 201compilation options needed to compile on various environments based on 202automatically defined symbols. Some machines don't seem to have useful 203symbols available, requiring that a compilation flag be defined in 204the Makefile; see the devtools/OS subdirectory for the supported 205architectures. 206 207If you are a system to which sendmail has already been ported you 208should not have to touch the following symbols. But if you are porting, 209you may have to tweak the following compilation flags in conf.h in order 210to get it to compile and link properly: 211 212SYSTEM5 Adjust for System V (not necessarily Release 4). 213SYS5SIGNALS Use System V signal semantics -- the signal handler 214 is automatically dropped when the signal is caught. 215 If this is not set, use POSIX/BSD semantics, where the 216 signal handler stays in force until an exec or an 217 explicit delete. Implied by SYSTEM5. 218SYS5SETPGRP Use System V setpgrp() semantics. Implied by SYSTEM5. 219HASNICE Define this to zero if you lack the nice(2) system call. 220HASRRESVPORT Define this to zero if you lack the rresvport(3) system call. 221HASFCHMOD Define this to one if you have the fchmod(2) system call. 222 This improves security. 223HASFCHOWN Define this to one if you have the fchown(2) system call. 224 This is required for the TrustedUser option if sendmail 225 must rebuild an (alias) map. 226HASFLOCK Set this if you prefer to use the flock(2) system call 227 rather than using fcntl-based locking. Fcntl locking 228 has some semantic gotchas, but many vendor systems 229 also interface it to lockd(8) to do NFS-style locking. 230 Unfortunately, may vendors implementations of fcntl locking 231 is just plain broken (e.g., locks are never released, 232 causing your sendmail to deadlock; when the kernel runs 233 out of locks your system crashes). For this reason, I 234 recommend always defining this unless you are absolutely 235 certain that your fcntl locking implementation really works. 236HASUNAME Set if you have the "uname" system call. Implied by 237 SYSTEM5. 238HASUNSETENV Define this if your system library has the "unsetenv" 239 subroutine. 240HASSETSID Define this if you have the setsid(2) system call. This 241 is implied if your system appears to be POSIX compliant. 242HASINITGROUPS Define this if you have the initgroups(3) routine. 243HASSETVBUF Define this if you have the setvbuf(3) library call. 244 If you don't, setlinebuf will be used instead. This 245 defaults on if your compiler defines __STDC__. 246HASSETREUID Define this if you have setreuid(2) ***AND*** root can 247 use setreuid to change to an arbitrary user. This second 248 condition is not satisfied on AIX 3.x. You may find that 249 your system has setresuid(2), (for example, on HP-UX) in 250 which case you will also have to #define setreuid(r, e) 251 to be the appropriate call. Some systems (such as Solaris) 252 have a compatibility routine that doesn't work properly, 253 but may have "saved user ids" properly implemented so you 254 can ``#define setreuid(r, e) seteuid(e)'' and have it work. 255 The important thing is that you have a call that will set 256 the effective uid independently of the real or saved uid 257 and be able to set the effective uid back again when done. 258 There's a test program in ../test/t_setreuid.c that will 259 try things on your system. Setting this improves the 260 security, since sendmail doesn't have to read .forward 261 and :include: files as root. There are certain attacks 262 that may be unpreventable without this call. 263USESETEUID Define this to 1 if you have a seteuid(2) system call that 264 will allow root to set only the effective user id to an 265 arbitrary value ***AND*** you have saved user ids. This is 266 preferable to HASSETREUID if these conditions are fulfilled. 267 These are the semantics of the to-be-released revision of 268 Posix.1. The test program ../test/t_seteuid.c will try 269 this out on your system. If you define both HASSETREUID 270 and USESETEUID, the former is ignored. 271HASSETEGID Define this if you have setegid(2) and it can be 272 used to set the saved gid. Please run t_dropgid in 273 test/ if you are not sure whether the call works. 274HASSETREGID Define this if you have setregid(2) and it can be 275 used to set the saved gid. Please run t_dropgid in 276 test/ if you are not sure whether the call works. 277HASSETRESGID Define this if you have setresgid(2) and it can be 278 used to set the saved gid. Please run t_dropgid in 279 test/ if you are not sure whether the call works. 280HASLSTAT Define this if you have symbolic links (and thus the 281 lstat(2) system call). This improves security. Unlike 282 most other options, this one is on by default, so you 283 need to #undef it in conf.h if you don't have symbolic 284 links (these days everyone does). 285HASSETRLIMIT Define this to 1 if you have the setrlimit(2) syscall. 286 You can define it to 0 to force it off. It is assumed 287 if you are running a BSD-like system. 288HASULIMIT Define this if you have the ulimit(2) syscall (System V 289 style systems). HASSETRLIMIT overrides, as it is more 290 general. 291HASWAITPID Define this if you have the waitpid(2) syscall. 292HASGETDTABLESIZE 293 Define this if you have the getdtablesize(2) syscall. 294HAS_ST_GEN Define this to 1 if your system has the st_gen field in 295 the stat structure (see stat(2)). 296HASSRANDOMDEV Define this if your system has the srandomdev(3) function 297 call. 298HASURANDOMDEV Define this if your system has /dev/urandom(4). 299HASSTRERROR Define this if you have the libc strerror(3) function (which 300 should be declared in <errno.h>), and it should be used 301 instead of sys_errlist. 302HASCLOSEFROM Define this if your system has closefrom(3). 303HASFDWALK Define this if your system has fdwalk(3). 304SM_CONF_GETOPT Define this as 0 if you need a reimplementation of getopt(3). 305 On some systems, getopt does very odd things if called 306 to scan the arguments twice. This flag will ask sendmail 307 to compile in a local version of getopt that works 308 properly. You may also need this if you build with 309 another library that introduces a non-standard getopt(3). 310NEEDSTRTOL Define this if your standard C library does not define 311 strtol(3). This will compile in a local version. 312NEEDFSYNC Define this if your standard C library does not define 313 fsync(2). This will try to simulate the operation using 314 fcntl(2); if that is not available it does nothing, which 315 isn't great, but at least it compiles and runs. 316HASGETUSERSHELL Define this to 1 if you have getusershell(3) in your 317 standard C library. If this is not defined, or is defined 318 to be 0, sendmail will scan the /etc/shells file (no 319 NIS-style support, defaults to /bin/sh and /bin/csh if 320 that file does not exist) to get a list of unrestricted 321 user shells. This is used to determine whether users 322 are allowed to forward their mail to a program or a file. 323NEEDPUTENV Define this if your system needs am emulation of the 324 putenv(3) call. Define to 1 to implement it in terms 325 of setenv(3) or to 2 to do it in terms of primitives. 326NOFTRUNCATE Define this if you don't have the ftruncate(2) syscall. 327 If you don't have this system call, there is an unavoidable 328 race condition that occurs when creating alias databases. 329GIDSET_T The type of entries in a gidset passed as the second 330 argument to getgroups(2). Historically this has been an 331 int, so this is the default, but some systems (such as 332 IRIX) pass it as a gid_t, which is an unsigned short. 333 This will make a difference, so it is important to get 334 this right! However, it is only an issue if you have 335 group sets. 336SLEEP_T The type returned by the system sleep() function. 337 Defaults to "unsigned int". Don't worry about this 338 if you don't have compilation problems. 339ARBPTR_T The type of an arbitrary pointer -- defaults to "void *". 340 If you are an very old compiler you may need to define 341 this to be "char *". 342SOCKADDR_LEN_T The type used for the third parameter to accept(2), 343 getsockname(2), and getpeername(2), representing the 344 length of a struct sockaddr. Defaults to int. 345SOCKOPT_LEN_T The type used for the fifth parameter to getsockopt(2) 346 and setsockopt(2), representing the length of the option 347 buffer. Defaults to int. 348LA_TYPE The type of load average your kernel supports. These 349 can be one of: 350 LA_ZERO (1) -- it always returns the load average as 351 "zero" (and does so on all architectures). 352 LA_INT (2) to read /dev/kmem for the symbol avenrun and 353 interpret as a long integer. 354 LA_FLOAT (3) same, but interpret the result as a floating 355 point number. 356 LA_SHORT (6) to interpret as a short integer. 357 LA_SUBR (4) if you have the getloadavg(3) routine in your 358 system library. 359 LA_MACH (5) to use MACH-style load averages (calls 360 processor_set_info()), 361 LA_PROCSTR (7) to read /proc/loadavg and interpret it 362 as a string representing a floating-point 363 number (Linux-style). 364 LA_READKSYM (8) is an implementation suitable for some 365 versions of SVr4 that uses the MIOC_READKSYM ioctl 366 call to read /dev/kmem. 367 LA_DGUX (9) is a special implementation for DG/UX that uses 368 the dg_sys_info system call. 369 LA_HPUX (10) is an HP-UX specific version that uses the 370 pstat_getdynamic system call. 371 LA_IRIX6 (11) is an IRIX 6.x specific version that adapts 372 to 32 or 64 bit kernels; it is otherwise very similar 373 to LA_INT. 374 LA_KSTAT (12) uses the (Solaris-specific) kstat(3k) 375 implementation. 376 LA_DEVSHORT (13) reads a short from a system file (default: 377 /dev/table/avenrun) and scales it in the same manner 378 as LA_SHORT. 379 LA_LONGLONG (17) to read /dev/kmem for the symbol avenrun and 380 interpret as a long long integer (e.g., for 64 bit 381 systems). 382 LA_INT, LA_SHORT, LA_FLOAT, and LA_READKSYM have several 383 other parameters that they try to divine: the name of your 384 kernel, the name of the variable in the kernel to examine, 385 the number of bits of precision in a fixed point load average, 386 and so forth. LA_DEVSHORT uses _PATH_AVENRUN to find the 387 device to be read to find the load average. 388 In desperation, use LA_ZERO. The actual code is in 389 conf.c -- it can be tweaked if you are brave. 390FSHIFT For LA_INT, LA_SHORT, and LA_READKSYM, this is the number 391 of bits of load average after the binary point -- i.e., 392 the number of bits to shift right in order to scale the 393 integer to get the true integer load average. Defaults to 8. 394_PATH_UNIX The path to your kernel. Needed only for LA_INT, LA_SHORT, 395 and LA_FLOAT. Defaults to "/unix" on System V, "/vmunix" 396 everywhere else. 397LA_AVENRUN For LA_INT, LA_SHORT, and LA_FLOAT, the name of the kernel 398 variable that holds the load average. Defaults to "avenrun" 399 on System V, "_avenrun" everywhere else. 400SFS_TYPE Encodes how your kernel can locate the amount of free 401 space on a disk partition. This can be set to SFS_NONE 402 (0) if you have no way of getting this information, 403 SFS_USTAT (1) if you have the ustat(2) system call, 404 SFS_4ARGS (2) if you have a four-argument statfs(2) 405 system call (and the include file is <sys/statfs.h>), 406 SFS_VFS (3), SFS_MOUNT (4), SFS_STATFS (5) if you have 407 the two-argument statfs(2) system call with includes in 408 <sys/vfs.h>, <sys/mount.h>, or <sys/statfs.h> respectively, 409 or SFS_STATVFS (6) if you have the two-argument statvfs(2) 410 call. The default if nothing is defined is SFS_NONE. 411SFS_BAVAIL with SFS_4ARGS you can also set SFS_BAVAIL to the field name 412 in the statfs structure that holds the useful information; 413 this defaults to f_bavail. 414SPT_TYPE Encodes how your system can display what a process is doing 415 on a ps(1) command (SPT stands for Set Process Title). Can 416 be set to: 417 SPT_NONE (0) -- Don't try to set the process title at all. 418 SPT_REUSEARGV (1) -- Pad out your argv with the information; 419 this is the default if none specified. 420 SPT_BUILTIN (2) -- The system library has setproctitle. 421 SPT_PSTAT (3) -- Use the PSTAT_SETCMD option to pstat(2) 422 to set the process title; this is used by HP-UX. 423 SPT_PSSTRINGS (4) -- Use the magic PS_STRINGS pointer (4.4BSD). 424 SPT_SYSMIPS (5) -- Use sysmips() supported by NEWS-OS 6. 425 SPT_SCO (6) -- Write kernel u. area. 426 SPT_CHANGEARGV (7) -- Write pointers to our own strings into 427 the existing argv vector. 428SPT_PADCHAR Character used to pad the process title; if undefined, 429 the space character (0x20) is used. This is ignored if 430 SPT_TYPE != SPT_REUSEARGV 431ERRLIST_PREDEFINED 432 If set, assumes that some header file defines sys_errlist. 433 This may be needed if you get type conflicts on this 434 variable -- otherwise don't worry about it. 435WAITUNION The wait(2) routine takes a "union wait" argument instead 436 of an integer argument. This is for compatibility with 437 old versions of BSD. 438SCANF You can set this to extend the F command to accept a 439 scanf string -- this gives you a primitive parser for 440 class definitions -- BUT it can make you vulnerable to 441 core dumps if the target file is poorly formed. 442SYSLOG_BUFSIZE You can define this to be the size of the buffer that 443 syslog accepts. If it is not defined, it assumes a 444 1024-byte buffer. If the buffer is very small (under 445 256 bytes) the log message format changes -- each 446 e-mail message will log many more messages, since it 447 will log each piece of information as a separate line 448 in syslog. 449BROKEN_RES_SEARCH 450 On Ultrix (and maybe other systems?) if you use the 451 res_search routine with an unknown host name, it returns 452 -1 but sets h_errno to 0 instead of HOST_NOT_FOUND. If 453 you set this, sendmail considers 0 to be the same as 454 HOST_NOT_FOUND. 455NAMELISTMASK If defined, values returned by nlist(3) are masked 456 against this value before use -- a common value is 457 0x7fffffff to strip off the top bit. 458BSD4_4_SOCKADDR If defined, socket addresses have an sa_len field that 459 defines the length of this address. 460SAFENFSPATHCONF Set this to 1 if and only if you have verified that a 461 pathconf(2) call with _PC_CHOWN_RESTRICTED argument on an 462 NFS filesystem where the underlying system allows users to 463 give away files to other users returns <= 0. Be sure you 464 try both on NFS V2 and V3. Some systems assume that their 465 local policy apply to NFS servers -- this is a bad 466 assumption! The test/t_pathconf.c program will try this 467 for you -- you have to run it in a directory that is 468 mounted from a server that allows file giveaway. 469SIOCGIFCONF_IS_BROKEN 470 Set this if your system has an SIOCGIFCONF ioctl defined, 471 but it doesn't behave the same way as "most" systems (BSD, 472 Solaris, SunOS, HP-UX, etc.) 473SIOCGIFNUM_IS_BROKEN 474 Set this if your system has an SIOCGIFNUM ioctl defined, 475 but it doesn't behave the same way as "most" systems 476 (Solaris, HP-UX). 477FAST_PID_RECYCLE 478 Set this if your system can reuse the same PID in the same 479 second. 480SO_REUSEADDR_IS_BROKEN 481 Set this if your system has a setsockopt() SO_REUSEADDR 482 flag but doesn't pay attention to it when trying to bind a 483 socket to a recently closed port. 484NEEDSGETIPNODE Set this if your system supports IPv6 but doesn't include 485 the getipnodeby{name,addr}() functions. Set automatically 486 for Linux's glibc. 487PIPELINING Support SMTP PIPELINING (set by default). 488USING_NETSCAPE_LDAP 489 Deprecated in favor of SM_CONF_LDAP_MEMFREE. See 490 libsm/README. 491NEEDLINK Set this if your system doesn't have a link() call. It 492 will create a copy of the file instead of a hardlink. 493USE_ENVIRON Set this to 1 to access process environment variables from 494 the external variable environ instead of the third 495 parameter of main(). 496USE_DOUBLE_FORK By default this is on (1). Set it to 0 to suppress the 497 extra fork() used to avoid intermediate zombies. 498ALLOW_255 Do not convert (char)0xff to (char)0x7f in headers etc. 499 This can also be done at runtime with the command line 500 option -d82.101. 501NEEDINTERRNO Set this if <errno.h> does not declare errno, i.e., if an 502 application needs to use 503 extern int errno; 504USE_TTYPATH Set this to 1 to enable ErrorMode=write. 505USESYSCTL Use sysctl(3) to determine the number of CPUs in a system. 506HASSNPRINTF Set this to 1 if your OS has a working snprintf(3), i.e., 507 it properly obeys the size of the buffer and returns the 508 number of characters that would have been printed if the 509 size were unlimited. 510LDAP_REFERRALS Set this if you want to use the -R flag (do not auto chase 511 referrals) for LDAP maps (requires -DLDAPMAP). 512MILTER_NO_NAGLE Turn off Nagle algorithm for communication with libmilter 513 ("cork" on Linux). On some operating systems this may 514 improve the interprocess communication performance. 515 516 517+-----------------------+ 518| COMPILE-TIME FEATURES | 519+-----------------------+ 520 521There are a bunch of features that you can decide to compile in, such 522as selecting various database packages and special protocol support. 523Several are assumed based on other compilation flags -- if you want to 524"un-assume" something, you probably need to edit conf.h. Compilation 525flags that add support for special features include: 526 527NDBM Include support for "new" DBM library for aliases and maps. 528 Normally defined in the Makefile. 529NEWDB Include support for Berkeley DB package (hash & btree) 530 for aliases and maps. Normally defined in the Makefile. 531 If the version of NEWDB you have is the old one that does 532 not include the "fd" call (this call was added in version 533 1.5 of the Berkeley DB code), you must upgrade to the 534 current version of Berkeley DB. 535NIS Define this to get NIS (YP) support for aliases and maps. 536 Normally defined in the Makefile. 537NISPLUS Define this to get NIS+ support for aliases and maps. 538 Normally defined in the Makefile. 539HESIOD Define this to get Hesiod support for aliases and maps. 540 Normally defined in the Makefile. 541NETINFO Define this to get NeXT NetInfo support for aliases and maps. 542 Normally defined in the Makefile. 543LDAPMAP Define this to get LDAP support for maps. 544PH_MAP Define this to get PH support for maps. 545MAP_NSD Define this to get nsd support for maps. 546USERDB Define this to 1 to include support for the User Information 547 Database. Implied by NEWDB or HESIOD. You can use 548 -DUSERDB=0 to explicitly turn it off. 549IDENTPROTO Define this as 1 to get IDENT (RFC 1413) protocol support. 550 This is assumed unless you are running on Ultrix or 551 HP-UX, both of which have a problem in the UDP 552 implementation. You can define it to be 0 to explicitly 553 turn off IDENT protocol support. If defined off, the code 554 is actually still compiled in, but it defaults off; you 555 can turn it on by setting the IDENT timeout in the 556 configuration file. 557IP_SRCROUTE Define this to 1 to get IP source routing information 558 displayed in the Received: header. This is assumed on 559 most systems, but some (e.g., Ultrix) apparently have a 560 broken version of getsockopt that doesn't properly 561 support the IP_OPTIONS call. You probably want this if 562 your OS can cope with it. Symptoms of failure will be that 563 it won't compile properly (that is, no support for fetching 564 IP_OPTIONs), or it compiles but source-routed TCP connections 565 either refuse to open or open and hang for no apparent reason. 566 Ultrix and AIX3 are known to fail this way. 567LOG Set this to get syslog(3) support. Defined by default 568 in conf.h. You want this if at all possible. 569NETINET Set this to get TCP/IP support. Defined by default 570 in conf.h. You probably want this. 571NETINET6 Set this to get IPv6 support. Other configuration may 572 be needed in conf.h for your particular operating system. 573 Also, DaemonPortOptions must be set appropriately for 574 sendmail to accept IPv6 connections. 575NETISO Define this to get ISO networking support. 576NETUNIX Define this to get Unix domain networking support. Defined 577 by default. A few bizarre systems (SCO, ISC, Altos) don't 578 support this networking domain. 579NETNS Define this to get NS networking support. 580NETX25 Define this to get X.25 networking support. 581NAMED_BIND If non-zero, include DNS (name daemon) support, including 582 MX support. The specs say you must use this if you run 583 SMTP. You don't have to be running a name server daemon 584 on your machine to need this -- any use of the DNS resolver, 585 including remote access to another machine, requires this 586 option. Defined by default in conf.h. Define it to zero 587 ONLY on machines that do not use DNS in any way. 588MATCHGECOS Permit fuzzy matching of user names against the full 589 name (GECOS) field in the /etc/passwd file. This should 590 probably be on, since you can disable it from the config 591 file if you want to. Defined by default in conf.h. 592MIME8TO7 If non-zero, include 8 to 7 bit MIME conversions. This 593 also controls advertisement of 8BITMIME in the ESMTP 594 startup dialogue. 595MIME7TO8_OLD If 0 then use an algorithm for MIME 7-bit quoted-printable 596 or base64 encoding to 8-bit text that has been introduced 597 in 8.12.3. There are some examples where that code fails, 598 but the old code works. If you have an example of improper 599 7 to 8 bit conversion please send it to sendmail-bugs. 600MIME7TO8 If non-zero, include 7 to 8 bit MIME conversions. 601HES_GETMAILHOST Define this to 1 if you are using Hesiod with the 602 hes_getmailhost() routine. This is included with the MIT 603 Hesiod distribution, but not with the DEC Hesiod distribution. 604XDEBUG Do additional internal checking. These don't cost too 605 much; you might as well leave this on. 606TCPWRAPPERS Turns on support for the TCP wrappers library (-lwrap). 607 See below for further information. 608SECUREWARE Enable calls to the SecureWare luid enabling/changing routines. 609 SecureWare is a C2 security package added to several UNIX's 610 (notably ConvexOS) to get a C2 Secure system. This 611 option causes mail delivery to be done with the luid of the 612 recipient. 613SHARE_V1 Support for the fair share scheduler, version 1. Setting to 614 1 causes final delivery to be done using the recipients 615 resource limitations. So far as I know, this is only 616 supported on ConvexOS. 617SASL Enables SMTP AUTH (RFC 2554). This requires the Cyrus SASL 618 library (ftp://ftp.andrew.cmu.edu/pub/cyrus-mail/). Please 619 install at least version 1.5.13. See below for further 620 information: SASL COMPILATION AND CONFIGURATION. If your 621 SASL library is older than 1.5.10, you have to set this 622 to its version number using a simple conversion: a.b.c 623 -> c + b*100 + a*10000, e.g. for 1.5.9 define SASL=10509. 624 Note: Using an older version than 1.5.5 of Cyrus SASL is 625 not supported. Starting with version 1.5.10, setting SASL=1 626 is sufficient. Any value other than 1 (or 0) will be 627 compared with the actual version found and if there is a 628 mismatch, compilation will fail. 629EGD Define this if your system has EGD installed, see 630 http://egd.sourceforge.net/ . It should be used to 631 seed the PRNG for STARTTLS if HASURANDOMDEV is not defined. 632STARTTLS Enables SMTP STARTTLS (RFC 2487). This requires OpenSSL 633 (http://www.OpenSSL.org/); use OpenSSL 0.9.5a or later 634 (if compatible with this version), do not use 0.9.3. 635 See STARTTLS COMPILATION AND CONFIGURATION for further 636 information. 637TLS_NO_RSA Turn off support for RSA algorithms in STARTTLS. 638MILTER Turn on support for external filters using the Milter API; 639 this option is set by default, to turn it off use 640 APPENDDEF(`conf_sendmail_ENVDEF', `-DMILTER=0') 641 in devtools/Site/site.config.m4 (see devtools/README). 642 See libmilter/README for more information about milter. 643REQUIRES_DIR_FSYNC Turn on support for file systems that require to 644 call fsync() for a directory if the meta-data in it has 645 been changed. This should be turned on at least for older 646 versions of ReiserFS; it is enabled by default for Linux. 647 According to some information this flag is not needed 648 anymore for kernel 2.4.16 and newer. We would appreciate 649 feedback about the semantics of the various file systems 650 available for Linux. 651 An alternative to this compile time flag is to mount the 652 queue directory without the -async option, or using 653 chattr +S on Linux. 654DBMMODE The default file permissions to use when creating new 655 database files for maps and aliases. Defaults to 0640. 656 657Generic notice: If you enable a compile time option that needs 658libraries or include files that don't come with sendmail or are 659installed in a location that your C compiler doesn't use by default 660you should set confINCDIRS and confLIBDIRS as explained in the 661first section: BUILDING SENDMAIL. 662 663 664+---------------------+ 665| DNS/RESOLVER ISSUES | 666+---------------------+ 667 668Many systems have old versions of the resolver library. At a minimum, 669you should be running BIND 4.8.3; older versions may compile, but they 670have known bugs that should give you pause. 671 672Common problems in old versions include "undefined" errors for 673dn_skipname. 674 675Some people have had a problem with BIND 4.9; it uses some routines 676that it expects to be externally defined such as strerror(). It may 677help to link with "-l44bsd" to solve this problem. This has apparently 678been fixed in later versions of BIND, starting around 4.9.3. In other 679words, if you use 4.9.0 through 4.9.2, you need -l44bsd; for earlier or 680later versions, you do not. 681 682!PLEASE! be sure to link with the same version of the resolver as 683the header files you used -- some people have used the 4.9 headers 684and linked with BIND 4.8 or vice versa, and it doesn't work. 685Unfortunately, it doesn't fail in an obvious way -- things just 686subtly don't work. 687 688WILDCARD MX RECORDS ARE A BAD IDEA! The only situation in which they 689work reliably is if you have two versions of DNS, one in the real world 690which has a wildcard pointing to your firewall, and a completely 691different version of the database internally that does not include 692wildcard MX records that match your domain. ANYTHING ELSE WILL GIVE 693YOU HEADACHES! 694 695When attempting to canonify a hostname, some broken name servers will 696return SERVFAIL (a temporary failure) on T_AAAA (IPv6) lookups. If you 697want to excuse this behavior, include WorkAroundBrokenAAAA in 698ResolverOptions. However, instead, we recommend catching the problem and 699reporting it to the name server administrator so we can rid the world of 700broken name servers. 701 702 703+----------------------------------------+ 704| STARTTLS COMPILATION AND CONFIGURATION | 705+----------------------------------------+ 706 707Please read the documentation accompanying the OpenSSL library. You 708have to compile and install the OpenSSL libraries before you can compile 709sendmail. See devtools/README how to set the correct compile time 710parameters; you should at least set the following variables: 711 712APPENDDEF(`conf_sendmail_ENVDEF', `-DSTARTTLS') 713APPENDDEF(`conf_sendmail_LIBS', `-lssl -lcrypto') 714 715If you have installed the OpenSSL libraries and include files in 716a location that your C compiler doesn't use by default you should 717set confINCDIRS and confLIBDIRS as explained in the first section: 718BUILDING SENDMAIL. 719 720Configuration information can be found in doc/op/op.me (required 721certificates) and cf/README (how to tell sendmail about certificates). 722 723To perform an initial test, connect to your sendmail daemon 724(telnet localhost 25) and issue a EHLO localhost and see whether 725250-STARTTLS 726is in the response. If it isn't, run the daemon with 727-O LogLevel=14 728and try again. Then take a look at the logfile and see whether 729there are any problems listed about permissions (unsafe files) 730or the validity of X.509 certificates. 731 732From: Garrett Wollman <wollman@lcs.mit.edu> 733 734 If your certificate authority is hierarchical, and you only include 735 the top-level CA certificate in the CACertFile file, some mail clients 736 may be unable to infer the proper certificate chain when selecting a 737 client certificate. Including the bottom-level CA certificate(s) in 738 the CACertFile file will allow these clients to work properly. This 739 is not necessary if you are not using client certificates for 740 authentication, or if all your clients are running Sendmail or other 741 programs using the OpenSSL library (which get it right automatically). 742 In addition, some mail clients are totally incapable of using 743 certificate authentication -- even some of those which already support 744 SSL/TLS for confidentiality. 745 746Further information can be found via: 747http://www.sendmail.org/tips/ 748 749 750+------------------------------------+ 751| SASL COMPILATION AND CONFIGURATION | 752+------------------------------------+ 753 754Please read the documentation accompanying the Cyrus SASL library 755(INSTALL and README). If you use Berkeley DB for Cyrus SASL then 756you must compile sendmail with the same version of Berkeley DB. 757See devtools/README for how to set the correct compile time parameters; 758you should at least set the following variables: 759 760APPENDDEF(`conf_sendmail_ENVDEF', `-DSASL') 761APPENDDEF(`conf_sendmail_LIBS', `-lsasl') 762 763If you have installed the Cyrus SASL library and include files in 764a location that your C compiler doesn't use by default you should 765set confINCDIRS and confLIBDIRS as explained in the first section: 766BUILDING SENDMAIL. 767 768You have to select and install authentication mechanisms and tell 769sendmail where to find the sasl library and the include files (see 770devtools/README for the parameters to set). Set up the required 771users and passwords as explained in the SASL documentation. See 772also cf/README for authentication related options (especially 773DefaultAuthInfo if you want authentication between MTAs). 774 775To perform an initial test, connect to your sendmail daemon 776(telnet localhost 25) and issue a EHLO localhost and see whether 777250-AUTH .... 778is in the response. If it isn't, run the daemon with 779-O LogLevel=14 780and try again. Then take a look at the logfile and see whether 781there are any security related problems listed (unsafe files). 782 783Further information can be found via: 784http://www.sendmail.org/tips/ 785 786 787+-------------------------------------+ 788| OPERATING SYSTEM AND COMPILE QUIRKS | 789+-------------------------------------+ 790 791GCC problems 792 When compiling with "gcc -O -Wall" specify "-DSM_OMIT_BOGUS_WARNINGS" 793 too (see include/sm/cdefs.h for more info). 794 795 ***************************************************************** 796 ** IMPORTANT: DO NOT USE OPTIMIZATION (``-O'') IF YOU ARE ** 797 ** RUNNING GCC 2.4.x or 2.5.x. THERE IS A BUG IN THE GCC ** 798 ** OPTIMIZER THAT CAUSES SENDMAIL COMPILES TO FAIL MISERABLY. ** 799 ***************************************************************** 800 801 Jim Wilson of Cygnus believes he has found the problem -- it will 802 probably be fixed in GCC 2.5.6 -- but until this is verified, be 803 very suspicious of gcc -O. This problem is reported to have been 804 fixed in gcc 2.6. 805 806 A bug in gcc 2.5.5 caused problems compiling sendmail 8.6.5 with 807 optimization on a Sparc. If you are using gcc 2.5.5, youi should 808 upgrade to the latest version of gcc. 809 810 Apparently GCC 2.7.0 on the Pentium processor has optimization 811 problems. I recommend against using -O on that architecture. This 812 has been seen on FreeBSD 2.0.5 RELEASE. 813 814 Solaris 2.X users should use version 2.7.2.3 over 2.7.2. 815 816 We have been told there are problems with gcc 2.8.0. If you are 817 using this version, you should upgrade to 2.8.1 or later. 818 819Berkeley DB 820 Berkeley DB 4.1.x with x <= 24 does not work with sendmail. 821 You need at least 4.1.25. 822 823GDBM GDBM does not work with sendmail because the additional 824 security checks and file locking cause problems. Unfortunately, 825 gdbm does not provide a compile flag in its version of ndbm.h so 826 the code can adapt. Until the GDBM authors can fix these problems, 827 GDBM will not be supported. Please use Berkeley DB instead. 828 829Configuration file location 830 Up to 8.6, sendmail tried to find the sendmail.cf file in the same 831 place as the vendors had put it, even when this was obviously 832 stupid. As of 8.7, sendmail ALWAYS looks for /etc/sendmail.cf. 833 Beginning with 8.10, sendmail uses /etc/mail/sendmail.cf. 834 You can get sendmail to use the stupid vendor .cf location by 835 adding -DUSE_VENDOR_CF_PATH during compilation, but this may break 836 support programs and scripts that need to find sendmail.cf. You 837 are STRONGLY urged to use symbolic links if you want to use the 838 vendor location rather than changing the location in the sendmail 839 binary. 840 841 NETINFO systems use NETINFO to determine the location of 842 sendmail.cf. The full path to sendmail.cf is stored as the value of 843 the "sendmail.cf" property in the "/locations/sendmail" 844 subdirectory of NETINFO. Set the value of this property to 845 "/etc/mail/sendmail.cf" (without the quotes) to use this new 846 default location for Sendmail 8.10.0 and higher. 847 848ControlSocket permissions 849 Paraphrased from BIND 8.2.1's README: 850 851 Solaris and other pre-4.4BSD kernels do not respect ownership or 852 protections on UNIX-domain sockets. The short term fix for this is to 853 override the default path and put such control sockets into root- 854 owned directories which do not permit non-root to r/w/x through them. 855 The long term fix is for all kernels to upgrade to 4.4BSD semantics. 856 857HP MPE/iX 858 The MPE-specific code within sendmail emulates a set-user-id root 859 environment for the sendmail binary. But there is no root uid 0 on 860 MPE, nor is there any support for set-user-id programs. Even when 861 sendmail thinks it is running as uid 0, it will still have the file 862 access rights of the underlying non-zero uid, but because sendmail is 863 an MPE priv-mode program it will still be able to call setuid() to 864 successfully switch to a new uid. 865 866 MPE setgid() semantics don't quite work the way sendmail expects, so 867 special emulation is done here also. 868 869 This uid/gid emulation is enabled via the setuid/setgid file mode bits 870 which are not currently used by MPE. Code in libsm/mpeix.c examines 871 these bits and enables emulation if they have been set, i.e., 872 chmod u+s,g+s /SENDMAIL/CURRENT/SENDMAIL. 873 874SunOS 4.x (Solaris 1.x) 875 You may have to use -lresolv on SunOS. However, beware that 876 this links in a new version of gethostbyname that does not 877 understand NIS, so you must have all of your hosts in DNS. 878 879 Some people have reported problems with the SunOS version of 880 -lresolv and/or in.named, and suggest that you get a newer 881 version. The symptoms are delays when you connect to the 882 SMTP server on a SunOS machine or having your domain added to 883 addresses inappropriately. There is a version of BIND 884 version 4.9 on gatekeeper.DEC.COM in pub/BSD/bind/4.9. 885 886 There is substantial disagreement about whether you can make 887 this work with resolv+, which allows you to specify a search-path 888 of services. Some people report that it works fine, others 889 claim it doesn't work at all (including causing sendmail to 890 drop core when it tries to do multiple resolv+ lookups for a 891 single job). I haven't tried resolv+, as we use DNS exclusively. 892 893 Should you want to try resolv+, it is on ftp.uu.net in 894 /networking/ip/dns. 895 896 Apparently getservbyname() can fail under moderate to high 897 load under some circumstances. This will exhibit itself as 898 the message ``554 makeconnection: service "smtp" unknown''. 899 The problem has been traced to one or more blank lines in 900 /etc/services on the NIS server machine. Delete these 901 and it should work. This info is thanks to Brian Bartholomew 902 <bb@math.ufl.edu> of I-Kinetics, Inc. 903 904 NOTE: The SunOS 4.X linker uses library paths specified during 905 compilation using -L for run-time shared library searches. 906 Therefore, it is vital that relative and unsafe directory paths not 907 be used when compiling sendmail. 908 909SunOS 4.0.2 (Sun 386i) 910 Date: Fri, 25 Aug 1995 11:13:58 +0200 (MET DST) 911 From: teus@oce.nl 912 913 Sendmail 8.7.Beta.12 compiles and runs nearly out of the box with the 914 following changes: 915 * Don't use /usr/5bin in your PATH, but make /usr/5bin/uname 916 available as "uname" command. 917 * Use the defines "-DBSD4_3 -DNAMED_BIND=0" in 918 devtools/OS/SunOS.4.0, which is selected via the "uname" command. 919 I recommend to make available the db-library on the system first 920 (and change the Makefile to use this library). 921 Note that the sendmail.cf and aliases files are found in /etc. 922 923SunOS 4.1.3, 4.1.3_U1 924 Sendmail causes crashes on SunOS 4.1.3 and 4.1.3_U1. According 925 to Sun bug number 1077939: 926 927 If an application does a getsockopt() on a SOCK_STREAM (TCP) socket 928 after the other side of the connection has sent a TCP RESET for 929 the stream, the kernel gets a Bus Trap in the tcp_ctloutput() or 930 ip_ctloutput() routine. 931 932 For 4.1.3, this is fixed in patch 100584-08, available on the 933 Sunsolve 2.7.1 or later CDs. For 4.1.3_U1, this was fixed in patch 934 101790-01 (SunOS 4.1.3_U1: TCP socket and reset problems), later 935 obsoleted by patch 102010-05. 936 937 Sun patch 100584-08 is not currently publicly available on their 938 ftp site but a user has reported it can be found at other sites 939 using a web search engine. 940 941Solaris 2.x (SunOS 5.x) 942 To compile for Solaris, the Makefile built by Build must 943 include a SOLARIS definition which reflects the Solaris version 944 (i.e. -DSOLARIS=20400 for 2.4 or -DSOLARIS=20501 for 2.5.1). 945 If you are using gcc, make sure -I/usr/include is not used (or 946 it might complain about TopFrame). If you are using Sun's cc, 947 make sure /opt/SUNWspro/bin/cc is used instead of /usr/ucb/cc 948 (or it might complain about tm_zone). 949 950 The Solaris 2.x (x <= 3) "syslog" function is apparently limited 951 to something about 90 characters because of a kernel limitation. 952 If you have source code, you can probably up this number. You 953 can get patches that fix this problem: the patch ids are: 954 955 Solaris 2.1 100834 956 Solaris 2.2 100999 957 Solaris 2.3 101318 958 959 Be sure you have the appropriate patch installed or you won't 960 see system logging. 961 962Solaris 2.4 (SunOS 5.4) 963 If you include /usr/lib at the end of your LD_LIBRARY_PATH you run 964 the risk of getting the wrong libraries under some circumstances. 965 This is because of a new feature in Solaris 2.4, described by 966 Rod.Evans@Eng.Sun.COM: 967 968 >> Prior to SunOS 5.4, any LD_LIBRARY_PATH setting was ignored by the 969 >> runtime linker if the application was setxid (secure), thus your 970 >> applications search path would be: 971 >> 972 >> /usr/local/lib LD_LIBRARY_PATH component - IGNORED 973 >> /usr/lib LD_LIBRARY_PATH component - IGNORED 974 >> /usr/local/lib RPATH - honored 975 >> /usr/lib RPATH - honored 976 >> 977 >> the effect is that path 3 would be the first used, and this would 978 >> satisfy your resolv.so lookup. 979 >> 980 >> In SunOS 5.4 we made the LD_LIBRARY_PATH a little more flexible. 981 >> People who developed setxid applications wanted to be able to alter 982 >> the library search path to some degree to allow for their own 983 >> testing and debugging mechanisms. It was decided that the only 984 >> secure way to do this was to allow a `trusted' path to be used in 985 >> LD_LIBRARY_PATH. The only trusted directory we presently define 986 >> is /usr/lib. Thus a set-user-ID root developer could play with some 987 >> alternative shared object implementations and place them in 988 >> /usr/lib (being root we assume they'ed have access to write in this 989 >> directory). This change was made as part of 1155380 - after a 990 >> *huge* amount of discussion regarding the security aspect of things. 991 >> 992 >> So, in SunOS 5.4 your applications search path would be: 993 >> 994 >> /usr/local/lib from LD_LIBRARY_PATH - IGNORED (untrustworthy) 995 >> /usr/lib from LD_LIBRARY_PATH - honored (trustworthy) 996 >> /usr/local/lib from RPATH - honored 997 >> /usr/lib from RPATH - honored 998 >> 999 >> here, path 2 would be the first used. 1000 1001Solaris 2.5.1 (SunOS 5.5.1) and 2.6 (SunOS 5.6) 1002 Apparently Solaris 2.5.1 patch 103663-01 installs a new 1003 /usr/include/resolv.h file that defines the __P macro without 1004 checking to see if it is already defined. This new resolv.h is also 1005 included in the Solaris 2.6 distribution. This causes compile 1006 warnings such as: 1007 1008 In file included from daemon.c:51: 1009 /usr/include/resolv.h:208: warning: `__P' redefined 1010 cdefs.h:58: warning: this is the location of the previous definition 1011 1012 These warnings can be safely ignored or you can create a resolv.h 1013 file in the obj.SunOS.5.5.1.* or obj.SunOS.5.6.* directory that reads: 1014 1015 #undef __P 1016 #include "/usr/include/resolv.h" 1017 1018 This problem was fixed in Solaris 7 (Sun bug ID 4081053). 1019 1020Solaris 7 (SunOS 5.7) 1021 Solaris 7 includes LDAP libraries but the implementation was 1022 lacking a few things. The following settings can be placed in 1023 devtools/Site/site.SunOS.5.7.m4 if you plan on using those 1024 libraries. 1025 1026 APPENDDEF(`confMAPDEF', `-DLDAPMAP') 1027 APPENDDEF(`confENVDEF', `-DLDAP_VERSION_MAX=3') 1028 APPENDDEF(`confLIBS', `-lldap') 1029 1030 Also, Sun's patch 107555 is needed to prevent a crash in the call 1031 to ldap_set_option for LDAP_OPT_REFERRALS in ldapmap_setopts if 1032 LDAP support is compiled in sendmail. 1033 1034Solaris 8 and later (SunOS 5.8 and later) 1035 Solaris 8 and later can optionally install LDAP support. If you 1036 have installed the Entire Distribution meta-cluster, you can use 1037 the following in devtools/Site/site.SunOS.5.8.m4 (or other 1038 appropriately versioned file) to enable LDAP: 1039 1040 APPENDDEF(`confMAPDEF', `-DLDAPMAP') 1041 APPENDDEF(`confLIBS', `-lldap') 1042 1043Solaris 9 and later (SunOS 5.9 and later) 1044 Solaris 9 and later have a revised LDAP library, libldap.so.5, 1045 which is derived from a Netscape implementation, thus requiring 1046 that SM_CONF_LDAP_MEMFREE be defined in conjunction with LDAPMAP: 1047 1048 APPENDDEF(`confMAPDEF', `-DLDAPMAP') 1049 APPENDDEF(`confENVDEF', `-DSM_CONF_LDAP_MEMFREE') 1050 APPENDDEF(`confLIBS', `-lldap') 1051 1052Solaris 1053 If you are using dns for hostname resolution on Solaris, make sure 1054 that the 'dns' entry is last on the hosts line in 1055 '/etc/nsswitch.conf'. For example, use: 1056 1057 hosts: nisplus files dns 1058 1059 Do not use: 1060 1061 hosts: nisplus dns [NOTFOUND=return] files 1062 1063 Note that 'nisplus' above is an illustration. The same comment 1064 applies no matter what naming services you are using. If you have 1065 anything other than dns last, even after "[NOTFOUND=return]", 1066 sendmail may not be able to determine whether an error was 1067 temporary or permanent. The error returned by the solaris 1068 gethostbyname() is the error for the last lookup used, and other 1069 naming services do not have the same concept of temporary failure. 1070 1071Ultrix 1072 By default, the IDENT protocol is turned off on Ultrix. If you 1073 are running Ultrix 4.4 or later, or if you have included patch 1074 CXO-8919 for Ultrix 4.2 or 4.3 to fix the TCP problem, you can turn 1075 IDENT on in the configuration file by setting the "ident" timeout. 1076 1077 The Ultrix 4.5 Y2K patch (ULTV45-022-1) has changed the resolver 1078 included in libc.a. Unfortunately, the __RES symbol hasn't changed 1079 and therefore, sendmail can no longer automatically detect the 1080 newer version. If you get a compiler error: 1081 1082 /lib/libc.a(gethostent.o): local_hostname_length: multiply defined 1083 1084 Then rebuild with this in devtools/Site/site.ULTRIX.m4: 1085 1086 APPENDDEF(`conf_sendmail_ENVDEF', `-DNEEDLOCAL_HOSTNAME_LENGTH=0') 1087 1088Digital UNIX (formerly DEC OSF/1) 1089 If you are compiling on OSF/1 (DEC Alpha), you must use 1090 -L/usr/shlib (otherwise it core dumps on startup). You may also 1091 need -mld to get the nlist() function, although some versions 1092 apparently don't need this. 1093 1094 Also, the enclosed makefile removed /usr/sbin/smtpd; if you need 1095 it, just create the link to the sendmail binary. 1096 1097 On DEC OSF/1 3.2 or earlier, the MatchGECOS option doesn't work 1098 properly due to a bug in the getpw* routines. If you want to use 1099 this, use -DDEC_OSF_BROKEN_GETPWENT=1. The problem is fixed in 3.2C. 1100 1101 Digital's mail delivery agent, /bin/mail (aka /bin/binmail), will 1102 only preserve the envelope sender in the "From " header if 1103 DefaultUserID is set to daemon. Setting this to mailnull will 1104 cause all mail to have the header "From mailnull ...". To use 1105 a different DefaultUserID, you will need to use a different mail 1106 delivery agent (such as mail.local found in the sendmail 1107 distribution). 1108 1109 On Digital UNIX 4.0 and later, Berkeley DB 1.85 is included with the 1110 operating system and already has the ndbm.o module removed. However, 1111 Digital has modified the original Berkeley DB db.h include file. 1112 This results in the following warning while compiling map.c and udb.c: 1113 1114 cc: Warning: /usr/include/db.h, line 74: The redefinition of the macro 1115 "__signed" conflicts with a current definition because the replacement 1116 lists differ. The redefinition is now in effect. 1117 #define __signed signed 1118 ------------------------^ 1119 1120 This warning can be ignored. 1121 1122 Digital UNIX's linker checks /usr/ccs/lib/ before /usr/lib/. 1123 If you have installed a new version of BIND in /usr/include 1124 and /usr/lib, you will experience difficulties as Digital ships 1125 libresolv.a in /usr/ccs/lib/ as well. Be sure to replace both 1126 copies of libresolv.a. 1127 1128IRIX 1129 The header files on SGI IRIX are completely prototyped, and as 1130 a result you can sometimes get some warning messages during 1131 compilation. These can be ignored. There are two errors in 1132 deliver only if you are using gcc, both of the form ``warning: 1133 passing arg N of `execve' from incompatible pointer type''. 1134 Also, if you compile with -DNIS, you will get a complaint 1135 about a declaration of struct dom_binding in a prototype 1136 when compiling map.c; this is not important because the 1137 function being prototyped is not used in that file. 1138 1139 In order to compile sendmail you will have had to install 1140 the developers' option in order to get the necessary include 1141 files. 1142 1143 If you compile with -lmalloc (the fast memory allocator), you may 1144 get warning messages such as the following: 1145 1146 ld32: WARNING 85: definition of _calloc in /usr/lib32/libmalloc.so 1147 preempts that definition in /usr/lib32/mips3/libc.so. 1148 ld32: WARNING 85: definition of _malloc in /usr/lib32/libmalloc.so 1149 preempts that definition in /usr/lib32/mips3/libc.so. 1150 ld32: WARNING 85: definition of _realloc in /usr/lib32/libmalloc.so 1151 preempts that definition in /usr/lib32/mips3/libc.so. 1152 ld32: WARNING 85: definition of _free in /usr/lib32/libmalloc.so 1153 preempts that definition in /usr/lib32/mips3/libc.so. 1154 ld32: WARNING 85: definition of _cfree in /usr/lib32/libmalloc.so 1155 preempts that definition in /usr/lib32/mips3/libc.so. 1156 1157 These are unavoidable and innocuous -- just ignore them. 1158 1159 According to Dave Sill <de5@ornl.gov>, there is a version of the 1160 Berkeley DB library patched to run on Irix 6.2 available from 1161 http://reality.sgi.com/ariel/freeware/#db . 1162 1163IRIX 6.x 1164 If you are using XFS filesystem, avoid using the -32 ABI switch to 1165 the cc compiler if possible. 1166 1167 Broken inet_aton and inet_ntoa on IRIX using gcc: There's 1168 a problem with gcc on IRIX, i.e., gcc can't pass structs 1169 less than 16 bits long unless they are 8 bits; IRIX 6.2 has 1170 some other sized structs. See 1171 http://www.bitmechanic.com/mail-archives/mysql/current/0418.html 1172 This problem seems to be fixed by gcc v2.95.2, gcc v2.8.1 1173 is reported as broken. Check your gcc version for this bug 1174 before installing sendmail. 1175 1176IRIX 6.4 1177 The IRIX 6.5.4 version of /bin/m4 does not work properly with 1178 sendmail. Either install fw_m4.sw.m4 off the Freeware_May99 CD and 1179 use /usr/freeware/bin/m4 or install and use GNU m4. 1180 1181NeXT or NEXTSTEP 1182 NEXTSTEP 3.3 and earlier ship with the old DBM library. Also, 1183 Berkeley DB does not currently run on NEXTSTEP. 1184 1185 If you are compiling on NEXTSTEP, you will have to create an 1186 empty file "unistd.h" and create a file "dirent.h" containing: 1187 1188 #include <sys/dir.h> 1189 #define dirent direct 1190 1191 (devtools/OS/NeXT should try to do both of these for you.) 1192 1193 Apparently, there is a bug in getservbyname on Nextstep 3.0 1194 that causes it to fail under some circumstances with the 1195 message "SYSERR: service "smtp" unknown" logged. You should 1196 be able to work around this by including the line: 1197 1198 OOPort=25 1199 1200 in your .cf file. 1201 1202BSDI (BSD/386) 1.0, NetBSD 0.9, FreeBSD 1.0 1203 The "m4" from BSDI won't handle the config files properly. 1204 I haven't had a chance to test this myself. 1205 1206 The M4 shipped in FreeBSD and NetBSD 0.9 don't handle the config 1207 files properly. One must use either GNU m4 1.1 or the PD-M4 1208 recently posted in comp.os.386bsd.bugs (and maybe others). 1209 NetBSD-current includes the PD-M4 (as stated in the NetBSD file 1210 CHANGES). 1211 1212 FreeBSD 1.0 RELEASE has uname(2) now. Use -DUSEUNAME in order to 1213 use it (look into devtools/OS/FreeBSD). NetBSD-current may have 1214 it too but it has not been verified. 1215 1216 The latest version of Berkeley DB uses a different naming 1217 scheme than the version that is supplied with your release. This 1218 means you will be able to use the current version of Berkeley DB 1219 with sendmail as long you use the new db.h when compiling 1220 sendmail and link it against the new libdb.a or libdb.so. You 1221 should probably keep the original db.h in /usr/include and the 1222 new db.h in /usr/local/include. 1223 12244.3BSD 1225 If you are running a "virgin" version of 4.3BSD, you'll have 1226 a very old resolver and be missing some header files. The 1227 header files are simple -- create empty versions and everything 1228 will work fine. For the resolver you should really port a new 1229 version (4.8.3 or later) of the resolver; 4.9 is available on 1230 gatekeeper.DEC.COM in pub/BSD/bind/4.9. If you are really 1231 determined to continue to use your old, buggy version (or as 1232 a shortcut to get sendmail working -- I'm sure you have the 1233 best intentions to port a modern version of BIND), you can 1234 copy ../contrib/oldbind.compat.c into sendmail and add the 1235 following to devtools/Site/site.config.m4: 1236 1237 APPENDDEF(`confOBJADD', `oldbind.compat.o') 1238 1239OpenBSD (up to 2.9 Release), NetBSD, FreeBSD (up to 4.3-RELEASE) 1240 m4 from *BSD won't handle libsm/Makefile.m4 properly, since the 1241 maximum length for strings is too short. You need to use GNU m4 1242 or patch m4, see for example: 1243 http://FreeBSD.org/cgi/cvsweb.cgi/src/usr.bin/m4/eval.c.diff?r1=1.11&r2=1.12 1244 1245A/UX 1246 Date: Tue, 12 Oct 1993 18:28:28 -0400 (EDT) 1247 From: "Eric C. Hagberg" <hagberg@med.cornell.edu> 1248 Subject: Fix for A/UX ndbm 1249 1250 I guess this isn't really a sendmail bug, however, it is something 1251 that A/UX users should be aware of when compiling sendmail 8.6. 1252 1253 Apparently, the calls that sendmail is using to the ndbm routines 1254 in A/UX 3.0.x contain calls to "broken" routines, in that the 1255 aliases database will break when it gets "just a little big" 1256 (sorry I don't have exact numbers here, but it broke somewhere 1257 around 20-25 aliases for me.), making all aliases non-functional 1258 after exceeding this point. 1259 1260 What I did was to get the gnu-dbm-1.6 package, compile it, and 1261 then re-compile sendmail with "-lgdbm", "-DNDBM", and using the 1262 ndbm.h header file that comes with the gnu-package. This makes 1263 things behave properly. 1264 [NOTE: see comment above about GDBM] 1265 1266 I suppose porting the New Berkeley DB package is another route, 1267 however, I made a quick attempt at it, and found it difficult 1268 (not easy at least); the gnu-dbm package "configured" and 1269 compiled easily. 1270 1271 [NOTE: Berkeley DB version 2.X runs on A/UX and can be used for 1272 database maps.] 1273 1274SCO Unix 1275 From: Thomas Essebier <tom@stallion.oz.au> 1276 Organisation: Stallion Technologies Pty Ltd. 1277 1278 It will probably help those who are trying to configure sendmail 8.6.9 1279 to know that if they are on SCO, they had better set 1280 OI-dnsrch 1281 or they will core dump as soon as they try to use the resolver. 1282 i.e., although SCO has _res.dnsrch defined, and is kinda BIND 4.8.3, 1283 it does not inititialise it, nor does it understand 'search' in 1284 /etc/named.boot. 1285 - sigh - 1286 1287 According to SCO, the m4 which ships with UnixWare 2.1.2 is broken. 1288 We recommend installing GNU m4 before attempting to build sendmail. 1289 1290 On some versions a bogus error value is listed if connections 1291 time out (large negative number). To avoid this explicitly set 1292 Timeout.connect to a reasonable value (several minutes). 1293 1294DG/UX 1295 Doug Anderson <dlander@afterlife.ncsc.mil> has successfully run 1296 V8 on the DG/UX 5.4.2 and 5.4R3.x platforms under heavy usage. 1297 Originally, the DG /bin/mail program wasn't compatible with 1298 the V8 sendmail, since the DG /bin/mail requires the environment 1299 variable "_FORCE_MAIL_LOCAL_=yes" be set. Version 8.7 now includes 1300 this in the environment before invoking the local mailer. Some 1301 have used procmail to avoid this problem in the past. It works 1302 but some have experienced file locking problems with their DG/UX 1303 ports of procmail. 1304 1305Apollo DomainOS 1306 If you are compiling on Apollo, you will have to create an empty 1307 file "unistd.h" (for DomainOS 10.3 and earlier) and create a file 1308 "dirent.h" containing: 1309 1310 #include <sys/dir.h> 1311 #define dirent direct 1312 1313 (devtools/OS/DomainOS will attempt to do both of these for you.) 1314 1315HP-UX 8.00 1316 Date: Mon, 24 Jan 1994 13:25:45 +0200 1317 From: Kimmo Suominen <Kimmo.Suominen@lut.fi> 1318 Subject: 8.6.5 w/ HP-UX 8.00 on s300 1319 1320 Just compiled and fought with sendmail 8.6.5 on a HP9000/360 (i.e., 1321 a series 300 machine) running HP-UX 8.00. 1322 1323 I was getting segmentation fault when delivering to a local user. 1324 With debugging I saw it was faulting when doing _free@libc... *sigh* 1325 It seems the new implementation of malloc on s300 is buggy as of 8.0, 1326 so I tried out the one in -lmalloc (malloc(3X)). With that it seems 1327 to work just dandy. 1328 1329 When linking, you will get the following error: 1330 1331 ld: multiply defined symbol _freespace in file /usr/lib/libmalloc.a 1332 1333 but you can just ignore it. You might want to add this info to the 1334 README file for the future... 1335 1336Linux 1337 Something broke between versions 0.99.13 and 0.99.14 of Linux: the 1338 flock() system call gives errors. If you are running .14, you must 1339 not use flock. You can do this with -DHASFLOCK=0. We have also 1340 been getting complaints since version 2.4.X was released. 1341 sendmail 8.13 has changed the default locking method to fcntl() 1342 for Linux kernel version 2.4 and later. Be sure to update other 1343 sendmail related programs to match locking techniques (some 1344 examples, besides makemap and mail.local, include procmail, mailx, 1345 mutt, elm, etc). 1346 1347 Around the inclusion of bind-4.9.3 & Linux libc-4.6.20, the 1348 initialization of the _res structure changed. If /etc/hosts.conf 1349 was configured as "hosts, bind" the resolver code could return 1350 "Name server failure" errors. This is supposedly fixed in 1351 later versions of libc (>= 4.6.29?), and later versions of 1352 sendmail (> 8.6.10) try to work around the problem. 1353 1354 Some older versions (< 4.6.20?) of the libc/include files conflict 1355 with sendmail's version of cdefs.h. Deleting sendmail's version 1356 on those systems should be non-harmful, and new versions don't care. 1357 1358 NOTE ON LINUX & BIND: By default, the Makefile generated for Linux 1359 includes header files in /usr/local/include and libraries in 1360 /usr/local/lib. If you've installed BIND on your system, the header 1361 files typically end up in the search path and you need to add 1362 "-lresolv" to the LIBS line in your Makefile. Really old versions 1363 may need to include "-l44bsd" as well (particularly if the link phase 1364 complains about missing strcasecmp, strncasecmp or strpbrk). 1365 Complaints about an undefined reference to `__dn_skipname' in 1366 domain.o are a sure sign that you need to add -lresolv to LIBS. 1367 Newer versions of Linux are basically threaded BIND, so you may or 1368 may not see complaints if you accidentally mix BIND 1369 headers/libraries with virginal libc. If you have BIND headers in 1370 /usr/local/include (resolv.h, etc) you *should* be adding -lresolv 1371 to LIBS. Data structures may change and you'd be asking for a 1372 core dump. 1373 1374 A number of problems have been reported regarding the Linux 2.2.0 1375 kernel. So far, these problems have been tracked down to syslog() 1376 and DNS resolution. We believe the problem is with the poll() 1377 implementation in the Linux 2.2.0 kernel and poll()-aware versions 1378 of glib (at least up to 2.0.111). 1379 1380glibc 1381 glibc 2.2.1 (and possibly other versions) changed the value of 1382 __RES in resolv.h but failed to actually provide the IPv6 API 1383 changes that the change implied. Therefore, compiling with 1384 -DNETINET6 fails. 1385 1386 Workarounds: 1387 1) Compile without -DNETINET6 1388 2) Build against a real BIND 8.2.2 include/lib tree 1389 3) Wait for glibc to fix it 1390 1391AIX 4.X 1392 The AIX 4.X linker uses library paths specified during compilation 1393 using -L for run-time shared library searches. Therefore, it is 1394 vital that relative and unsafe directory paths not be using when 1395 compiling sendmail. Because of this danger, by default, compiles 1396 on AIX use the -blibpath option to limit shared libraries to 1397 /usr/lib and /lib. If you need to allow more directories, such as 1398 /usr/local/lib, modify your devtools/Site/site.AIX.4.2.m4, 1399 site.AIX.4.3.m4, and/or site.AIX.4.x.m4 file(s) and set confLDOPTS 1400 appropriately. For example: 1401 1402 define(`confLDOPTS', `-blibpath:/usr/lib:/lib:/usr/local/lib') 1403 1404 Be sure to only add (safe) system directories. 1405 1406 The AIX version of GNU ld also exhibits this problem. If you are 1407 using that version, instead of -blibpath, use its -rpath option. 1408 For example: 1409 1410 gcc -Wl,-rpath /usr/lib -Wl,-rpath /lib -Wl,-rpath /usr/local/lib 1411 1412AIX 4.X If the test program t-event (and most others) in libsm fails, 1413 check your compiler settings. It seems that the flags -qnoro or 1414 -qnoroconst on some AIX versions trigger a compiler bug. Check 1415 your compiler settings or use cc instead of xlc. 1416 1417AIX 4.0-4.2, maybe some AIX 4.3 versions 1418 The AIX m4 implements a different mechanism for ifdef which is 1419 inconsistent with other versions of m4. Therefore, it will not 1420 work properly with the sendmail Build architecture or m4 1421 configuration method. To work around this problem, please use 1422 GNU m4 from ftp://ftp.gnu.org/pub/gnu/. 1423 The problem seems to be solved in AIX 4.3.3 at least. 1424 1425AIX 4.3.3 1426 From: Valdis.Kletnieks@vt.edu 1427 Date: Sun, 02 Jul 2000 03:58:02 -0400 1428 1429 Under AIX 4.3.3, after applying bos.adt.include 4.3.3.12 to close the 1430 BIND 8.2.2 security holes, you can no longer build with -DNETINET6 1431 because they changed the value of __RES in resolv.h but failed to 1432 actually provide the API changes that the change implied. 1433 1434 Workarounds: 1435 1) Compile without -DNETINET6 1436 2) Build against a real BIND 8.2.2 include/lib tree 1437 3) Wait for IBM to fix it 1438 1439AIX 3.x 1440 This version of sendmail does not support MB, MG, and MR resource 1441 records, which are supported by AIX sendmail. 1442 1443 Several people have reported that the IBM-supplied named returns 1444 fairly random results -- the named should be replaced. It is not 1445 necessary to replace the resolver, which will simplify installation. 1446 A new BIND resolver can be found at http://www.isc.org/isc/. 1447 1448AIX 3.1.x 1449 The supplied load average code only works correctly for AIX 3.2.x. 1450 For 3.1, use -DLA_TYPE=LA_SUBR and get the latest ``monitor'' 1451 package by Jussi Maki <jmaki@hut.fi> from ftp.funet.fi in the 1452 directory pub/unix/AIX/rs6000/monitor-1.12.tar.Z; use the loadavgd 1453 daemon, and the getloadavg subroutine supplied with that package. 1454 If you don't care about load average throttling, just turn off 1455 load average checking using -DLA_TYPE=LA_ZERO. 1456 1457RISC/os 1458 RISC/os from MIPS is a merged AT&T/Berkeley system. When you 1459 compile on that platform you will get duplicate definitions 1460 on many files. You can ignore these. 1461 1462System V Release 4 Based Systems 1463 There is a single devtools OS that is intended for all SVR4-based 1464 systems (built from devtools/OS/SVR4). It defines __svr4__, 1465 which is predefined by some compilers. If your compiler already 1466 defines this compile variable, you can delete the definition from 1467 the generated Makefile or create a devtools/Site/site.config.m4 1468 file. 1469 1470 It's been tested on Dell Issue 2.2. 1471 1472DELL SVR4 1473 Date: Mon, 06 Dec 1993 10:42:29 EST 1474 From: "Kimmo Suominen" <kim@grendel.lut.fi> 1475 Message-ID: <2d0352f9.lento29@lento29.UUCP> 1476 To: eric@cs.berkeley.edu 1477 Cc: sendmail@cs.berkeley.edu 1478 Subject: Notes for DELL SVR4 1479 1480 Eric, 1481 1482 Here are some notes for compiling Sendmail 8.6.4 on DELL SVR4. I ran 1483 across these things when helping out some people who contacted me by 1484 e-mail. 1485 1486 1) Use gcc 2.4.5 (or later?). Dell distributes gcc 2.1 with their 1487 Issue 2.2 Unix. It is too old, and gives you problems with 1488 clock.c, because sigset_t won't get defined in <sys/signal.h>. 1489 This is due to a problematic protection rule in there, and is 1490 fixed with gcc 2.4.5. 1491 1492 2) If you don't use the new Berkeley DB (-DNEWDB), then you need 1493 to add "-lc -lucb" to the libraries to link with. This is because 1494 the -ldbm distributed by Dell needs the bcopy, bcmp and bzero 1495 functions. It is important that you specify both libraries in 1496 the given order to be sure you only get the BSTRING functions 1497 from the UCB library (and not the signal routines etc.). 1498 1499 3) Don't leave out "-lelf" even if compiling with "-lc -lucb". 1500 The UCB library also has another copy of the nlist routines, 1501 but we do want the ones from "-lelf". 1502 1503 If anyone needs a compiled gcc 2.4.5 and/or a ported DB library, they 1504 can use anonymous ftp to fetch them from lut.fi in the /kim directory. 1505 They are copies of what I use on grendel.lut.fi, and offering them 1506 does not imply that I would also support them. I have sent the DB 1507 port for SVR4 back to Keith Bostic for inclusion in the official 1508 distribution, but I haven't heard anything from him as of today. 1509 1510 - gcc-2.4.5-svr4.tar.gz (gcc 2.4.5 and the corresponding libg++) 1511 - db-1.72.tar.gz (with source, objects and a installed copy) 1512 1513 Cheers 1514 + Kim 1515 -- 1516 * Kimmo.Suominen@lut.fi * SysVr4 enthusiast at GRENDEL.LUT.FI * 1517 * KIM@FINFILES.BITNET * Postmaster and Hostmaster at LUT.FI * 1518 * + 358 200 865 718 * Unix area moderator at NIC.FUNET.FI * 1519 1520ConvexOS 10.1 and below 1521 In order to use the name server, you must create the file 1522 /etc/use_nameserver. If this file does not exist, the call 1523 to res_init() will fail and you will have absolutely no 1524 access to DNS, including MX records. 1525 1526Amdahl UTS 2.1.5 1527 In order to get UTS to work, you will have to port BIND 4.9. 1528 The vendor's BIND is reported to be ``totally inadequate.'' 1529 See sendmail/contrib/AmdahlUTS.patch for the patches necessary 1530 to get BIND 4.9 compiled for UTS. 1531 1532UnixWare 1533 According to Alexander Kolbasov <sasha@unitech.gamma.ru>, 1534 the m4 on UnixWare 2.0 (still in Beta) will core dump on the 1535 config files. GNU m4 and the m4 from UnixWare 1.x both work. 1536 1537 According to Larry Rosenman <ler@lerami.lerctr.org>: 1538 1539 UnixWare 2.1.[23]'s m4 chokes (not obviously) when 1540 processing the 8.9.0 cf files. 1541 1542 I had a LOCAL_RULE_0 that wound up AFTER the 1543 SBasic_check_rcpt rules using the SCO supplied M4. 1544 GNU M4 works fine. 1545 1546UNICOS 8.0.3.4 1547 Some people have reported that the -O flag on UNICOS can cause 1548 problems. You may want to turn this off if you have problems 1549 running sendmail. Reported by Jerry G. DeLapp <jgd@acl.lanl.gov>. 1550 1551Darwin/Mac OS X (10.X.X) 1552 The linker errors produced regarding getopt() and its associated 1553 variables can safely be ignored. 1554 1555 From Mike Zimmerman <zimmy@torrentnet.com>: 1556 1557 From scratch here is what Darwin users need to do to the standard 1558 10.0.0, 10.0.1 install to get sendmail working. 1559 From http://www.macosx.com/forums/showthread.php?s=6dac0e9e1f3fd118a4870a8a9b559491&threadid=2242: 1560 1. chmod g-w / /private /private/etc 1561 2. Properly set HOSTNAME in /etc/hostconfig to your FQDN: 1562 HOSTNAME=-my.domain.com- 1563 3. Edit /etc/rc.boot: 1564 hostname my.domain.com 1565 domainname domain.com 1566 4. Edit /System/Library/StartupItems/Sendmail/Sendmail: 1567 Remove the "&" after the sendmail command: 1568 /usr/sbin/sendmail -bd -q1h 1569 1570 From Carsten Klapp <carsten.klapp@home.com>: 1571 1572 The easiest workaround is to remove the group-writable permission 1573 for the root directory and the symbolic /etc inherits this 1574 change. While this does fix sendmail, the unfortunate side-effect 1575 is the OS X admin will no longer be able to manipulate icons in the 1576 top level of the Startup disk unless logged into the GUI as the 1577 superuser. 1578 1579 In applying the alternate workaround, care must be taken while 1580 swapping the symlink /etc with the directory /private/etc. In all 1581 likelihood any admin who is concerned with this sendmail error has 1582 enough experience to not accidentally harm anything in the process. 1583 1584 a. Swap the /etc symlink with /private/etc (as superuser): 1585 rm /etc 1586 mv /private/etc /etc 1587 ln -s /etc /private/etc 1588 1589 b. Set / to group unwritable (as superuser): 1590 chmod g-w / 1591 1592Darwin/Mac OS X (10.1.5) 1593 Apple's upgrade to sendmail 8.12 is incorrectly configured. You 1594 will need to manually fix it up by doing the following: 1595 1596 1. chown smmsp:smmsp /var/spool/clientmqueue 1597 2. chmod 2770 /var/spool/clientmqueue 1598 3. chgrp smmsp /usr/sbin/sendmail 1599 4. chmod g+s /usr/sbin/sendmail 1600 1601 From Daniel J. Luke <dluke@geeklair.net>: 1602 1603 It appears that setting the sendmail.cf property in 1604 /locations/sendmail in NetInfo on Mac OS X 10.1.5 with sendmail 1605 8.12.4 causes 'bad things' to happen. 1606 1607 Specifically sendmail instances that should be getting their config 1608 from /etc/mail/submit.cf don't (so mail/mutt/perl scripts which 1609 open pipes to sendmail stop working as sendmail tries to write to 1610 /var/spool/mqueue and cannot as sendmail is no longer suid root). 1611 1612 Removing the entry from NetInfo fixes this problem. 1613 1614GNU getopt 1615 I'm told that GNU getopt has a problem in that it gets confused 1616 by the double call. Use the version in conf.c instead. 1617 1618BIND 4.9.2 and Ultrix 1619 If you are running on Ultrix, be sure you read conf/Info.Ultrix 1620 in the BIND distribution very carefully -- there is information 1621 in there that you need to know in order to avoid errors of the 1622 form: 1623 1624 /lib/libc.a(gethostent.o): sethostent: multiply defined 1625 /lib/libc.a(gethostent.o): endhostent: multiply defined 1626 /lib/libc.a(gethostent.o): gethostbyname: multiply defined 1627 /lib/libc.a(gethostent.o): gethostbyaddr: multiply defined 1628 1629 during the link stage. 1630 1631BIND 8.X 1632 BIND 8.X returns HOST_NOT_FOUND instead of TRY_AGAIN on temporary 1633 DNS failures when trying to find the hostname associated with an IP 1634 address (gethostbyaddr()). This can cause problems as 1635 $&{client_name} based lookups in class R ($=R) and the access 1636 database won't succeed. 1637 1638 This will be fixed in BIND 8.2.1. For earlier versions, this can 1639 be fixed by making "dns" the last name service queried for host 1640 resolution in /etc/irs.conf: 1641 1642 hosts local continue 1643 hosts dns 1644 1645strtoul 1646 Some compilers (notably gcc) claim to be ANSI C but do not 1647 include the ANSI-required routine "strtoul". If your compiler 1648 has this problem, you will get an error in srvrsmtp.c on the 1649 code: 1650 1651 # ifdef defined(__STDC__) && !defined(BROKEN_ANSI_LIBRARY) 1652 e->e_msgsize = strtoul(vp, (char **) NULL, 10); 1653 # else 1654 e->e_msgsize = strtol(vp, (char **) NULL, 10); 1655 # endif 1656 1657 You can use -DBROKEN_ANSI_LIBRARY to get around this problem. 1658 1659Listproc 6.0c 1660 Date: 23 Sep 1995 23:56:07 GMT 1661 Message-ID: <95925101334.~INN-AUMa00187.comp-news@dl.ac.uk> 1662 From: alansz@mellers1.psych.berkeley.edu (Alan Schwartz) 1663 Subject: Listproc 6.0c + Sendmail 8.7 [Helpful hint] 1664 1665 Just upgraded to sendmail 8.7, and discovered that listproc 6.0c 1666 breaks, because it, by default, sends a blank "HELO" rather than 1667 a "HELO hostname" when using the 'system' or 'telnet' mail method. 1668 1669 The fix is to include -DZMAILER in the compilation, which will 1670 cause it to use "HELO hostname" (which Z-mail apparently requires 1671 as well. :) 1672 1673OpenSSL 1674 OpenSSL versions prior to 0.9.6 use a macro named Free which 1675 conflicts with existing macro names on some platforms, such as 1676 AIX. 1677 Do not use 0.9.3, but OpenSSL 0.9.5a or later if compatible with 1678 0.9.5a. 1679 1680PH 1681 PH support is provided by Mark Roth <roth@uiuc.edu>. The map is 1682 described at http://www-dev.cites.uiuc.edu/sendmail/ . 1683 1684 NOTE: The "spacedname" pseudo-field which was used by earlier 1685 versions of the PH map code is no longer supported! See the URL 1686 listed above for more information. 1687 1688 Please contact Mark Roth for support and questions regarding the 1689 map. 1690 1691TCP Wrappers 1692 If you are using -DTCPWRAPPERS to get TCP Wrappers support you will 1693 also need to install libwrap.a and modify your site.config.m4 file 1694 or the generated Makefile to include -lwrap in the LIBS line 1695 (make sure that INCDIRS and LIBDIRS point to where the tcpd.h and 1696 libwrap.a can be found). 1697 1698 TCP Wrappers is available at ftp://ftp.porcupine.org/pub/security/. 1699 1700 If you have alternate MX sites for your site, be sure that all of 1701 your MX sites reject the same set of hosts. If not, a bad guy whom 1702 you reject will connect to your site, fail, and move on to the next 1703 MX site, which will accept the mail for you and forward it on to you. 1704 1705Regular Expressions (MAP_REGEX) 1706 If sendmail linking fails with: 1707 1708 undefined reference to 'regcomp' 1709 1710 or sendmail gives an error about a regular expression with: 1711 1712 pattern-compile-error: : Operation not applicable 1713 1714 Your libc does not include a running version of POSIX-regex. Use 1715 librx or regex.o from the GNU Free Software Foundation, 1716 ftp://ftp.gnu.org/pub/gnu/rx-?.?.tar.gz or 1717 ftp://ftp.gnu.org/pub/gnu/regex-?.?.tar.gz. 1718 You can also use the regex-lib by Henry Spencer, 1719 ftp://ftp.funet.fi/pub/languages/C/spencer/regex.shar.gz 1720 Make sure, your compiler reads regex.h from the distribution, 1721 not from /usr/include, otherwise sendmail will dump a core. 1722 1723Fedora Core 5, 64 bit version 1724 If the ld stage fails with undefined functions like 1725 __res_querydomain, __dn_expand 1726 then add these lines to devtools/Site/site.config.m4 1727 1728 APPENDDEF(`confLIBDIRS', `-L/usr/lib64') 1729 APPENDDEF(`confINCDIRS', `-I/usr/include/bind9') 1730 1731 and rebuild (sh ./Build -c). 1732 1733 Problem noted by Daniel Krones, solution suggested by 1734 Anthony Howe. 1735 1736+--------------+ 1737| MANUAL PAGES | 1738+--------------+ 1739 1740The manual pages have been written against the -man macros, and 1741should format correctly with any reasonable *roff. 1742 1743 1744+-----------------+ 1745| DEBUGGING HOOKS | 1746+-----------------+ 1747 1748As of 8.6.5, sendmail daemons will catch a SIGUSR1 signal and log 1749some debugging output (logged at LOG_DEBUG severity). The 1750information dumped is: 1751 1752 * The value of the $j macro. 1753 * A warning if $j is not in the set $=w. 1754 * A list of the open file descriptors. 1755 * The contents of the connection cache. 1756 * If ruleset 89 is defined, it is evaluated and the results printed. 1757 1758This allows you to get information regarding the runtime state of the 1759daemon on the fly. This should not be done too frequently, since 1760the process of rewriting may lose memory which will not be recovered. 1761Also, ruleset 89 may call non-reentrant routines, so there is a small 1762non-zero probability that this will cause other problems. It is 1763really only for debugging serious problems. 1764 1765A typical formulation of ruleset 89 would be: 1766 1767 R$* $@ $>0 some test address 1768 1769 1770+-----------------------------+ 1771| DESCRIPTION OF SOURCE FILES | 1772+-----------------------------+ 1773 1774The following list describes the files in this directory: 1775 1776Build Shell script for building sendmail. 1777Makefile A convenience for calling ./Build. 1778Makefile.m4 A template for constructing a makefile based on the 1779 information in the devtools directory. 1780README This file. 1781TRACEFLAGS My own personal list of the trace flags -- not guaranteed 1782 to be particularly up to date. 1783alias.c Does name aliasing in all forms. 1784aliases.5 Man page describing the format of the aliases file. 1785arpadate.c A subroutine which creates ARPANET standard dates. 1786bf.c Routines to implement memory-buffered file system using 1787 hooks provided by libsm now (formerly Torek stdio library). 1788bf.h Buffered file I/O function declarations and 1789 data structure and function declarations for bf.c. 1790collect.c The routine that actually reads the mail into a temp 1791 file. It also does a certain amount of parsing of 1792 the header, etc. 1793conf.c The configuration file. This contains information 1794 that is presumed to be quite static and non- 1795 controversial, or code compiled in for efficiency 1796 reasons. Most of the configuration is in sendmail.cf. 1797conf.h Configuration that must be known everywhere. 1798control.c Routines to implement control socket. 1799convtime.c A routine to sanely process times. 1800daemon.c Routines to implement daemon mode. 1801deliver.c Routines to deliver mail. 1802domain.c Routines that interface with DNS (the Domain Name 1803 System). 1804envelope.c Routines to manipulate the envelope structure. 1805err.c Routines to print error messages. 1806headers.c Routines to process message headers. 1807helpfile An example helpfile for the SMTP HELP command and -bt mode. 1808macro.c The macro expander. This is used internally to 1809 insert information from the configuration file. 1810mailq.1 Man page for the mailq command. 1811main.c The main routine to sendmail. This file also 1812 contains some miscellaneous routines. 1813makesendmail A convenience for calling ./Build. 1814map.c Support for database maps. 1815mci.c Routines that handle mail connection information caching. 1816milter.c MTA portions of the mail filter API. 1817mime.c MIME conversion routines. 1818newaliases.1 Man page for the newaliases command. 1819parseaddr.c The routines which do address parsing. 1820queue.c Routines to implement message queueing. 1821readcf.c The routine that reads the configuration file and 1822 translates it to internal form. 1823recipient.c Routines that manipulate the recipient list. 1824sasl.c Routines to interact with Cyrys-SASL. 1825savemail.c Routines which save the letter on processing errors. 1826sendmail.8 Man page for the sendmail command. 1827sendmail.h Main header file for sendmail. 1828sfsasl.c I/O interface between SASL/TLS and the MTA. 1829sfsasl.h Header file for sfsasl.c. 1830shmticklib.c Routines for shared memory counters. 1831sm_resolve.c Routines for DNS lookups (for DNS map type). 1832sm_resolve.h Header file for sm_resolve.c. 1833srvrsmtp.c Routines to implement server SMTP. 1834stab.c Routines to manage the symbol table. 1835stats.c Routines to collect and post the statistics. 1836statusd_shm.h Data structure and function declarations for shmticklib.c. 1837sysexits.c List of error messages associated with error codes 1838 in sysexits.h. 1839sysexits.h List of error codes for systems that lack their own. 1840timers.c Routines to provide microtimers. 1841timers.h Data structure and function declarations for timers.h. 1842tls.c Routines for TLS. 1843trace.c The trace package. These routines allow setting and 1844 testing of trace flags with a high granularity. 1845udb.c The user database interface module. 1846usersmtp.c Routines to implement user SMTP. 1847util.c Some general purpose routines used by sendmail. 1848version.c The version number and information about this 1849 version of sendmail. 1850 1851(Version $Revision: 8.393 $, last update $Date: 2013/11/22 20:51:54 $ ) 1852