README revision 38032
1# Copyright (c) 1998 Sendmail, Inc. All rights reserved. 2# Copyright (c) 1983, 1995-1997 Eric P. Allman. All rights reserved. 3# Copyright (c) 1988 4# The Regents of the University of California. All rights reserved. 5# 6# By using this file, you agree to the terms and conditions set 7# forth in the LICENSE file which can be found at the top level of 8# the sendmail distribution. 9# 10# 11# @(#)README 8.206 (Berkeley) 6/30/98 12# 13 14This directory contains the source files for sendmail(TM). 15 16********************* 17!! DO NOT USE MAKE !! in this directory to compile sendmail -- 18********************* instead, use the "Build" script located in 19the src directory. It will build an appropriate Makefile, and 20create an appropriate obj.* subdirectory so that multiplatform 21support works easily. 22 23 ********************************************************** 24 ** Read below for more details on building sendmail. ** 25 ********************************************************** 26 27************************************************************************** 28** IMPORTANT: Read the appropriate paragraphs in the section on ** 29** ``Operating System and Compile Quirks''. ** 30************************************************************************** 31 32For detailed instructions, please read the document ../doc/op/op.me: 33 34 eqn ../doc/op/op.me | pic | ditroff -me 35 36Sendmail is a trademark of Sendmail, Inc. 37 38 39+-------------------+ 40| BUILDING SENDMAIL | 41+-------------------+ 42 43By far, the easiest way to compile sendmail is to use the "Build" 44script: 45 46 sh Build 47 48This uses the "uname" command to figure out what architecture you are 49on and creates a proper Makefile accordingly. It also creates a 50subdirectory per object format, so that multiarchitecture support is 51easy. In general this should be all you need. IRIX 6.x users should 52read the note below in the OPERATING SYSTEM AND COMPILE QUIRKS section. 53 54If you need to look at other include or library directories, use the 55-I or -L flags on the command line, e.g., 56 57 sh Build -I/usr/sww/include -L/usr/sww/lib 58 59It's also possible to create local site configuration in the file 60site.config.m4 (or another file settable with the -f flag). This 61file contains M4 definitions for various compilation values; the 62most useful are: 63 64confMAPDEF -D flags to specify database types to be included 65 (see below) 66confENVDEF -D flags to specify other environment information 67confINCDIRS -I flags for finding include files during compilation 68confLIBDIRS -L flags for finding libraries during linking 69confLIBS -l flags for selecting libraries during linking 70confLDOPTS other ld(1) linker options 71 72Others can be found by examining Makefile.m4. Please read 73../BuildTools/README for more information about the site.config.m4 74file. 75 76You can recompile from scratch using the -c flag with the Build 77command. This removes the existing compilation directory for the 78current platform and builds a new one. 79 80Porting to a new Unix-based system should be a matter of creating 81an appropriate configuration file in the BuildTools/OS/ directory. 82 83 84 85+----------------------+ 86| DATABASE DEFINITIONS | 87+----------------------+ 88 89There are several database formats that can be used for the alias files 90and for general maps. When used for alias files they interact in an 91attempt to be backward compatible. 92 93The options are: 94 95NEWDB The new Berkeley DB package. Some systems (e.g., BSD/OS and 96 Digital UNIX 4.0) have some version of this package 97 pre-installed. If your system does not have Berkeley DB 98 pre-installed, or the version installed is not version 2.0 99 or greater (e.g., is Berkeley DB 1.85 or 1.86), get the 100 current version from http://www.sleepycat.com/. DO NOT 101 use a version from any of the University of California, 102 Berkeley "Net" or other distributions. If you are still 103 running BSD/386 1.x, you will need to upgrade the included 104 Berkeley DB library to a current version. NEWDB is included 105 automatically if the Build script can find a library named 106 libdb.a. 107NDBM The older NDBM implementation -- the very old V7 DBM 108 implementation is no longer supported. 109NIS Network Information Services. To use this you must have 110 NIS support on your system. 111NISPLUS NIS+ (the revised NIS released with Solaris 2). You must 112 have NIS+ support on your system to use this flag. 113HESIOD Support for Hesiod (from the DEC/Athena distribution). You 114 must already have Hesiod support on your system for this to 115 work. You may be able to get this to work with the MIT/Athena 116 version of Hesiod, but that's likely to be a lot of work. 117LDAPMAP Lightweight Directory Lookup Protocol support. You will 118 have to install the UMich ldap and lber libraries to use 119 this flag. 120MAP_REGEX Regular Expression support. You will need to use an 121 operating system which comes with the POSIX regex() 122 routines or install a regexp library such as libregex from 123 the Free Software Foundation. 124 125>>> NOTE WELL for NEWDB support: If you want to get ndbm support, for 126>>> Berkeley DB versions under 2.0, it is CRITICAL that you remove 127>>> ndbm.o from libdb.a before you install it and DO NOT install ndbm.h; 128>>> for Berkeley DB versions 2.0 through 2.3.14, remove dbm.o from libdb.a 129>>> before you install it. If you don't delete these, there is absolutely 130>>> no point to including -DNDBM, since it will just get you another 131>>> (inferior) API to the same format database. These files OVERRIDE 132>>> calls to ndbm routines -- in particular, if you leave ndbm.h in, 133>>> you can find yourself using the new db package even if you don't 134>>> define NEWDB. Berkeley DB versions later than 2.3.14 do not need 135>>> to be modified. Please also consult the README in the top level 136>>> directory of the sendmail distribution for other important information. 137>>> 138>>> Further note: DO NOT remove your existing /usr/include/ndbm.h -- 139>>> you need that one. But do not install an updated ndbm.h in 140>>> /usr/include, /usr/local/include, or anywhere else. 141 142If NEWDB and NDBM are defined (but not NIS), then sendmail will read 143NDBM format alias files, but the next time a newaliases is run the 144format will be converted to NEWDB; that format will be used forever 145more. This is intended as a transition feature. 146 147If NEWDB, NDBM, and NIS are all defined and the name of the file includes 148the string "/yp/", sendmail will rebuild BOTH the NEWDB and NDBM format 149alias files. However, it will only read the NEWDB file; the NDBM format 150file is used only by the NIS subsystem. This is needed because the NIS 151maps on an NIS server are built directly from the NDBM files. 152 153If NDBM and NIS are defined (regardless of the definition of NEWDB), 154and the filename includes the string "/yp/", sendmail adds the special 155tokens "YP_LAST_MODIFIED" and "YP_MASTER_NAME", both of which are 156required if the NDBM file is to be used as an NIS map. 157 158All of these flags are normally defined in the DBMDEF line in the 159Makefile. 160 161If you define NEWDB or HESIOD you get the User Database (USERDB) 162automatically. Generally you do want to have NEWDB for it to do 163anything interesting. See above for getting the Berkeley DB 164package (i.e., NEWDB). There is no separate "user database" 165package -- don't bother searching for it on the net. 166 167Hesiod and LDAP require libraries that may not be installed with your 168system. These are outside of my ability to provide support. See the 169"Quirks" section for more information. 170 171The regex map can be used to see if an address matches a certain regular 172expression. For example, all-numerics local parts are common spam 173addresses, so "^[0-9]+$" would match this. By using such a map in a 174check_* rule-set, you can block a certain range of addresses that would 175otherwise be considered valid. 176 177+---------------+ 178| COMPILE FLAGS | 179+---------------+ 180 181Wherever possible, I try to make sendmail pull in the correct 182compilation options needed to compile on various environments based on 183automatically defined symbols. Some machines don't seem to have useful 184symbols available, requiring that a compilation flag be defined in 185the Makefile; see the Buildtools/OS subdirectory for the supported 186architectures. 187 188If you are a system to which sendmail has already been ported you 189should not have to touch the following symbols. But if you are porting, 190you may have to tweak the following compilation flags in conf.h in order 191to get it to compile and link properly: 192 193SYSTEM5 Adjust for System V (not necessarily Release 4). 194SYS5SIGNALS Use System V signal semantics -- the signal handler 195 is automatically dropped when the signal is caught. 196 If this is not set, use POSIX/BSD semantics, where the 197 signal handler stays in force until an exec or an 198 explicit delete. Implied by SYSTEM5. 199SYS5SETPGRP Use System V setpgrp() semantics. Implied by SYSTEM5. 200HASFCHMOD Define this to one if you have the fchmod(2) system call. 201 This improves security. 202HASFLOCK Set this if you prefer to use the flock(2) system call 203 rather than using fcntl-based locking. Fcntl locking 204 has some semantic gotchas, but many vendor systems 205 also interface it to lockd(8) to do NFS-style locking. 206 Unfortunately, may vendors implementations of fcntl locking 207 is just plain broken (e.g., locks are never released, 208 causing your sendmail to deadlock; when the kernel runs 209 out of locks your system crashes). For this reason, I 210 recommend always defining this unless you are absolutely 211 certain that your fcntl locking implementation really works. 212HASUNAME Set if you have the "uname" system call. Implied by 213 SYSTEM5. 214HASUNSETENV Define this if your system library has the "unsetenv" 215 subroutine. 216HASSETSID Define this if you have the setsid(2) system call. This 217 is implied if your system appears to be POSIX compliant. 218HASINITGROUPS Define this if you have the initgroups(3) routine. 219HASSETVBUF Define this if you have the setvbuf(3) library call. 220 If you don't, setlinebuf will be used instead. This 221 defaults on if your compiler defines __STDC__. 222HASSETREUID Define this if you have setreuid(2) ***AND*** root can 223 use setreuid to change to an arbitrary user. This second 224 condition is not satisfied on AIX 3.x. You may find that 225 your system has setresuid(2), (for example, on HP-UX) in 226 which case you will also have to #define setreuid(r, e) 227 to be the appropriate call. Some systems (such as Solaris) 228 have a compatibility routine that doesn't work properly, 229 but may have "saved user ids" properly implemented so you 230 can ``#define setreuid(r, e) seteuid(e)'' and have it work. 231 The important thing is that you have a call that will set 232 the effective uid independently of the real or saved uid 233 and be able to set the effective uid back again when done. 234 There's a test program in ../test/t_setreuid.c that will 235 try things on your system. Setting this improves the 236 security, since sendmail doesn't have to read .forward 237 and :include: files as root. There are certain attacks 238 that may be unpreventable without this call. 239USESETEUID Define this to 1 if you have a seteuid(2) system call that 240 will allow root to set only the effective user id to an 241 arbitrary value ***AND*** you have saved user ids. This is 242 preferable to HASSETREUID if these conditions are fulfilled. 243 These are the semantics of the to-be-released revision of 244 Posix.1. The test program ../test/t_seteuid.c will try 245 this out on your system. If you define both HASSETREUID 246 and USESETEUID, the former is ignored. 247HASLSTAT Define this if you have symbolic links (and thus the 248 lstat(2) system call). This improves security. Unlike 249 most other options, this one is on by default, so you 250 need to #undef it in conf.h if you don't have symbolic 251 links (these days everyone does). 252HASSETRLIMIT Define this to 1 if you have the setrlimit(2) syscall. 253 You can define it to 0 to force it off. It is assumed 254 if you are running a BSD-like system. 255HASULIMIT Define this if you have the ulimit(2) syscall (System V 256 style systems). HASSETRLIMIT overrides, as it is more 257 general. 258HASWAITPID Define this if you have the waitpid(2) syscall. 259HASGETDTABLESIZE 260 Define this if you have the getdtablesize(2) syscall. 261HAS_ST_GEN Define this to 1 if your system has the st_gen field in 262 the stat structure (see stat(2)). 263USESTRERROR Define this if you have the libc strerror function (which 264 should be declared in <errno.h>), and it should be used 265 instead of sys_errlist. 266NEEDGETOPT Define this if you need a reimplementation of getopt(3). 267 On some systems, getopt does very odd things if called 268 to scan the arguments twice. This flag will ask sendmail 269 to compile in a local version of getopt that works 270 properly. 271NEEDSTRTOL Define this if your standard C library does not define 272 strtol(3). This will compile in a local version. 273NEEDVPRINTF Define this if your standard C library does not define 274 vprintf(3). Note that the resulting fake implementation 275 is not very elegant and may not even work on some 276 architectures. 277NEEDFSYNC Define this if your standard C library does not define 278 fsync(2). This will try to simulate the operation using 279 fcntl(2); if that is not available it does nothing, which 280 isn't great, but at least it compiles and runs. 281HASGETUSERSHELL Define this to 1 if you have getusershell(3) in your 282 standard C library. If this is not defined, or is defined 283 to be 0, sendmail will scan the /etc/shells file (no 284 NIS-style support, defaults to /bin/sh and /bin/csh if 285 that file does not exist) to get a list of unrestricted 286 user shells. This is used to determine whether users 287 are allowed to forward their mail to a program or a file. 288NEEDPUTENV Define this if your system needs am emulation of the 289 putenv(3) call. Define to 1 to implement it in terms 290 of setenv(3) or to 2 to do it in terms of primitives. 291NOFTRUNCATE Define this if you don't have the ftruncate(2) syscall. 292 If you don't have this system call, there is an unavoidable 293 race condition that occurs when creating alias databases. 294GIDSET_T The type of entries in a gidset passed as the second 295 argument to getgroups(2). Historically this has been an 296 int, so this is the default, but some systems (such as 297 IRIX) pass it as a gid_t, which is an unsigned short. 298 This will make a difference, so it is important to get 299 this right! However, it is only an issue if you have 300 group sets. 301SLEEP_T The type returned by the system sleep() function. 302 Defaults to "unsigned int". Don't worry about this 303 if you don't have compilation problems. 304ARBPTR_T The type of an arbitrary pointer -- defaults to "void *". 305 If you are an very old compiler you may need to define 306 this to be "char *". 307SOCKADDR_LEN_T The type used for the third parameter to accept(2), 308 getsockname(2), and getpeername(2), representing the 309 length of a struct sockaddr. Defaults to int. 310SOCKOPT_LEN_T The type used for the fifth parameter to getsockopt(2) 311 and setsockopt(2), representing the length of the option 312 buffer. Defaults to int. 313LA_TYPE The type of load average your kernel supports. These 314 can be one of: 315 LA_ZERO (1) -- it always returns the load average as 316 "zero" (and does so on all architectures). 317 LA_INT (2) to read /dev/kmem for the symbol avenrun and 318 interpret as a long integer. 319 LA_FLOAT (3) same, but interpret the result as a floating 320 point number. 321 LA_SHORT (6) to interpret as a short integer. 322 LA_SUBR (4) if you have the getloadavg(3) routine in your 323 system library. 324 LA_MACH (5) to use MACH-style load averages (calls 325 processor_set_info()), 326 LA_PROCSTR (7) to read /proc/loadavg and interpret it 327 as a string representing a floating-point 328 number (Linux-style). 329 LA_READKSYM (8) is an implementation suitable for some 330 versions of SVr4 that uses the MIOC_READKSYM ioctl 331 call to read /dev/kmem. 332 LA_DGUX (9) is a special implementation for DG/UX that uses 333 the dg_sys_info system call. 334 LA_HPUX (10) is an HP-UX specific version that uses the 335 pstat_getdynamic system call. 336 LA_IRIX6 (11) is an IRIX 6.x specific version that adapts 337 to 32 or 64 bit kernels; it is otherwise very similar 338 to LA_INT. 339 LA_KSTAT (12) uses the (Solaris-specific) kstat(3k) 340 implementation. 341 LA_DEVSHORT (13) reads a short from a system file (default: 342 /dev/table/avenrun) and scales it in the same manner 343 as LA_SHORT. 344 LA_INT, LA_SHORT, LA_FLOAT, and LA_READKSYM have several 345 other parameters that they try to divine: the name of your 346 kernel, the name of the variable in the kernel to examine, 347 the number of bits of precision in a fixed point load average, 348 and so forth. LA_DEVSHORT uses _PATH_AVENRUN to find the 349 device to be read to find the load average. 350 In desperation, use LA_ZERO. The actual code is in 351 conf.c -- it can be tweaked if you are brave. 352FSHIFT For LA_INT, LA_SHORT, and LA_READKSYM, this is the number 353 of bits of load average after the binary point -- i.e., 354 the number of bits to shift right in order to scale the 355 integer to get the true integer load average. Defaults to 8. 356_PATH_UNIX The path to your kernel. Needed only for LA_INT, LA_SHORT, 357 and LA_FLOAT. Defaults to "/unix" on System V, "/vmunix" 358 everywhere else. 359LA_AVENRUN For LA_INT, LA_SHORT, and LA_FLOAT, the name of the kernel 360 variable that holds the load average. Defaults to "avenrun" 361 on System V, "_avenrun" everywhere else. 362SFS_TYPE Encodes how your kernel can locate the amount of free 363 space on a disk partition. This can be set to SFS_NONE 364 (0) if you have no way of getting this information, 365 SFS_USTAT (1) if you have the ustat(2) system call, 366 SFS_4ARGS (2) if you have a four-argument statfs(2) 367 system call (and the include file is <sys/statfs.h>), 368 SFS_VFS (3), SFS_MOUNT (4), SFS_STATFS (5) if you have 369 the two-argument statfs(2) system call with includes in 370 <sys/vfs.h>, <sys/mount.h>, or <sys/statfs.h> respectively, 371 or SFS_STATVFS (6) if you have the two-argument statvfs(2) 372 call. The default if nothing is defined is SFS_NONE. 373SFS_BAVAIL with SFS_4ARGS you can also set SFS_BAVAIL to the field name 374 in the statfs structure that holds the useful information; 375 this defaults to f_bavail. 376SPT_TYPE Encodes how your system can display what a process is doing 377 on a ps(1) command (SPT stands for Set Process Title). Can 378 be set to: 379 SPT_NONE (0) -- Don't try to set the process title at all. 380 SPT_REUSEARGV (1) -- Pad out your argv with the information; 381 this is the default if none specified. 382 SPT_BUILTIN (2) -- The system library has setproctitle. 383 SPT_PSTAT (3) -- Use the PSTAT_SETCMD option to pstat(2) 384 to set the process title; this is used by HP-UX. 385 SPT_PSSTRINGS (4) -- Use the magic PS_STRINGS pointer (4.4BSD). 386 SPT_SYSMIPS (5) -- Use sysmips() supported by NEWS-OS 6. 387 SPT_SCO (6) -- Write kernel u. area. 388 SPT_CHANGEARGV (7) -- Write pointers to our own strings into 389 the existing argv vector. 390SPT_PADCHAR Character used to pad the process title; if undefined, 391 the space character (0x20) is used. This is ignored if 392 SPT_TYPE != SPT_REUSEARGV 393ERRLIST_PREDEFINED 394 If set, assumes that some header file defines sys_errlist. 395 This may be needed if you get type conflicts on this 396 variable -- otherwise don't worry about it. 397WAITUNION The wait(2) routine takes a "union wait" argument instead 398 of an integer argument. This is for compatibility with 399 old versions of BSD. 400SCANF You can set this to extend the F command to accept a 401 scanf string -- this gives you a primitive parser for 402 class definitions -- BUT it can make you vulnerable to 403 core dumps if the target file is poorly formed. 404SYSLOG_BUFSIZE You can define this to be the size of the buffer that 405 syslog accepts. If it is not defined, it assumes a 406 1024-byte buffer. If the buffer is very small (under 407 256 bytes) the log message format changes -- each 408 e-mail message will log many more messages, since it 409 will log each piece of information as a separate line 410 in syslog. 411BROKEN_RES_SEARCH 412 On Ultrix (and maybe other systems?) if you use the 413 res_search routine with an unknown host name, it returns 414 -1 but sets h_errno to 0 instead of HOST_NOT_FOUND. If 415 you set this, sendmail considers 0 to be the same as 416 HOST_NOT_FOUND. 417NAMELISTMASK If defined, values returned by nlist(3) are masked 418 against this value before use -- a common value is 419 0x7fffffff to strip off the top bit. 420BSD4_4_SOCKADDR If defined, socket addresses have an sa_len field that 421 defines the length of this address. 422SAFENFSPATHCONF Set this to 1 if and only if you have verified that a 423 pathconf(2) call with _PC_CHOWN_RESTRICTED argument on an 424 NFS filesystem where the underlying system allows users to 425 give away files to other users returns <= 0. Be sure you 426 try both on NFS V2 and V3. Some systems assume that their 427 local policy apply to NFS servers -- this is a bad 428 assumption! The test/t_pathconf.c program will try this 429 for you -- you have to run it in a directory that is 430 mounted from a server that allows file giveaway. 431SIOCGIFCONF_IS_BROKEN 432 Set this if your system has an SIOCGIFCONF ioctl defined, 433 but it doesn't behave the same way as "most" systems (BSD, 434 Solaris, SunOS, HP-UX, etc.) 435SIOCGIFNUM_IS_BROKEN 436 Set this if your system has an SIOCGIFNUM ioctl defined, 437 but it doesn't behave the same way as "most" systems 438 (Solaris, HP-UX). 439NEED_PERCENTQ Set this if your system doesn't support the printf 440 format strings %lld or %llu. If this is set, %qd and 441 %qu are used instead. 442 443 444 445+-----------------------+ 446| COMPILE-TIME FEATURES | 447+-----------------------+ 448 449There are a bunch of features that you can decide to compile in, such 450as selecting various database packages and special protocol support. 451Several are assumed based on other compilation flags -- if you want to 452"un-assume" something, you probably need to edit conf.h. Compilation 453flags that add support for special features include: 454 455NDBM Include support for "new" DBM library for aliases and maps. 456 Normally defined in the Makefile. 457NEWDB Include support for Berkeley DB package (hash & btree) 458 for aliases and maps. Normally defined in the Makefile. 459 If the version of NEWDB you have is the old one that does 460 not include the "fd" call (this call was added in version 461 1.5 of the Berkeley DB code), you must upgrade to the 462 current version of Berkeley DB. 463NIS Define this to get NIS (YP) support for aliases and maps. 464 Normally defined in the Makefile. 465NISPLUS Define this to get NIS+ support for aliases and maps. 466 Normally defined in the Makefile. 467HESIOD Define this to get Hesiod support for aliases and maps. 468 Normally defined in the Makefile. 469NETINFO Define this to get NeXT NetInfo support for aliases and maps. 470 Normally defined in the Makefile. 471USERDB Define this to 1 to include support for the User Information 472 Database. Implied by NEWDB or HESIOD. You can use 473 -DUSERDB=0 to explicitly turn it off. 474IDENTPROTO Define this as 1 to get IDENT (RFC 1413) protocol support. 475 This is assumed unless you are running on Ultrix or 476 HP-UX, both of which have a problem in the UDP 477 implementation. You can define it to be 0 to explicitly 478 turn off IDENT protocol support. If defined off, the code 479 is actually still compiled in, but it defaults off; you 480 can turn it on by setting the IDENT timeout to 30s in the 481 configuration file. 482IP_SRCROUTE Define this to 1 to get IP source routing information 483 displayed in the Received: header. This is assumed on 484 most systems, but some (e.g., Ultrix) apparently have a 485 broken version of getsockopt that doesn't properly 486 support the IP_OPTIONS call. You probably want this if 487 your OS can cope with it. Symptoms of failure will be that 488 it won't compile properly (that is, no support for fetching 489 IP_OPTIONs), or it compiles but source-routed TCP connections 490 either refuse to open or open and hang for no apparent reason. 491 Ultrix and AIX3 are known to fail this way. 492LOG Set this to get syslog(3) support. Defined by default 493 in conf.h. You want this if at all possible. 494NETINET Set this to get TCP/IP support. Defined by default 495 in conf.h. You probably want this. 496NETISO Define this to get ISO networking support. 497NETUNIX Define this to get Unix domain networking support. Defined 498 by default. A few bizarre systems (SCO, ISC, Altos) don't 499 support this networking domain. 500SMTP Define this to get the SMTP code. Implied by NETINET 501 or NETISO. 502NAMED_BIND If non-zero, include DNS (name daemon) support, including 503 MX support. The specs say you must use this if you run 504 SMTP. You don't have to be running a name server daemon 505 on your machine to need this -- any use of the DNS resolver, 506 including remote access to another machine, requires this 507 option. Defined by default in conf.h. Define it to zero 508 ONLY on machines that do not use DNS in any way. 509QUEUE Define this to get queueing code. Implied by NETINET 510 or NETISO; required by SMTP. This gives you other good 511 stuff -- it should be on. 512DAEMON Define this to get general network support. Implied by 513 NETINET or NETISO. Defined by default in conf.h. You 514 almost certainly want it on. 515MATCHGECOS Permit fuzzy matching of user names against the full 516 name (GECOS) field in the /etc/passwd file. This should 517 probably be on, since you can disable it from the config 518 file if you want to. Defined by default in conf.h. 519MIME8TO7 If non-zero, include 8 to 7 bit MIME conversions. This 520 also controls advertisement of 8BITMIME in the ESMTP 521 startup dialogue. 522MIME7TO8 If non-zero, include 7 to 8 bit MIME conversions. 523HES_GETMAILHOST Define this to 1 if you are using Hesiod with the 524 hes_getmailhost() routine. This is included with the MIT 525 Hesiod distribution, but not with the DEC Hesiod distribution. 526XDEBUG Do additional internal checking. These don't cost too 527 much; you might as well leave this on. 528TCPWRAPPERS Turns on support for the TCP wrappers library (-lwrap). 529 See below for further information. 530SECUREWARE Enable calls to the SecureWare luid enabling/changing routines. 531 SecureWare is a C2 security package added to several UNIX's 532 (notably ConvexOS) to get a C2 Secure system. This 533 option causes mail delivery to be done with the luid of the 534 recipient. 535SHARE_V1 Support for the fair share scheduler, version 1. Setting to 536 1 causes final delivery to be done using the recipients 537 resource limitations. So far as I know, this is only 538 supported on ConvexOS. 539 540 541+---------------------+ 542| DNS/RESOLVER ISSUES | 543+---------------------+ 544 545Many systems have old versions of the resolver library. At a minimum, 546you should be running BIND 4.8.3; older versions may compile, but they 547have known bugs that should give you pause. 548 549Common problems in old versions include "undefined" errors for 550dn_skipname. 551 552Some people have had a problem with BIND 4.9; it uses some routines 553that it expects to be externally defined such as strerror(). It may 554help to link with "-l44bsd" to solve this problem. This has apparently 555been fixed in later versions of BIND, starting around 4.9.3. In other 556words, if you use 4.9.0 through 4.9.2, you need -l44bsd; for earlier or 557later versions, you do not. 558 559!PLEASE! be sure to link with the same version of the resolver as 560the header files you used -- some people have used the 4.9 headers 561and linked with BIND 4.8 or vice versa, and it doesn't work. 562Unfortunately, it doesn't fail in an obvious way -- things just 563subtly don't work. 564 565WILDCARD MX RECORDS ARE A BAD IDEA! The only situation in which they 566work reliably is if you have two versions of DNS, one in the real world 567which has a wildcard pointing to your firewall, and a completely 568different version of the database internally that does not include 569wildcard MX records that match your domain. ANYTHING ELSE WILL GIVE 570YOU HEADACHES! 571 572 573+-------------------------------------+ 574| OPERATING SYSTEM AND COMPILE QUIRKS | 575+-------------------------------------+ 576 577GCC problems 578 ***************************************************************** 579 ** IMPORTANT: DO NOT USE OPTIMIZATION (``-O'') IF YOU ARE ** 580 ** RUNNING GCC 2.4.x or 2.5.x. THERE IS A BUG IN THE GCC ** 581 ** OPTIMIZER THAT CAUSES SENDMAIL COMPILES TO FAIL MISERABLY. ** 582 ***************************************************************** 583 584 Jim Wilson of Cygnus believes he has found the problem -- it will 585 probably be fixed in GCC 2.5.6 -- but until this is verified, be 586 very suspicious of gcc -O. This problem is reported to have been 587 fixed in gcc 2.6. 588 589 A bug in gcc 2.5.5 caused problems compiling sendmail 8.6.5 with 590 optimization on a Sparc. If you are using gcc 2.5.5, youi should 591 upgrade to the latest version of gcc. 592 593 Apparently GCC 2.7.0 on the Pentium processor has optimization 594 problems. I recommend against using -O on that architecture. This 595 has been seen on FreeBSD 2.0.5 RELEASE. 596 597 Solaris 2.X users should use version 2.7.2.3 over 2.7.2. 598 599 We have been told there are problems with gcc 2.8.0. If you are 600 using this version, you should upgrade to 2.8.1 or later. 601 602GDBM GDBM does not work with sendmail 8.8 because the additional 603 security checks and file locking cause problems. Unfortunately, 604 gdbm does not provide a compile flag in its version of ndbm.h so 605 the code can adapt. Until the GDBM authors can fix these problems, 606 GDBM will not be supported. Please use Berkeley DB instead. 607 608Configuration file location 609 Up to 8.6, sendmail tried to find the sendmail.cf file in the same 610 place as the vendors had put it, even when this was obviously 611 stupid. As of 8.7, sendmail ALWAYS looks for /etc/sendmail.cf. 612 Beginning with 8.10, sendmail will use /etc/mail/sendmail.cf. 613 You can get sendmail to use the stupid vendor .cf location by 614 adding -DUSE_VENDOR_CF_PATH during compilation, but this may break 615 support programs and scripts that need to find sendmail.cf. You 616 are STRONGLY urged to use symbolic links if you want to use the 617 vendor location rather than changing the location in the sendmail 618 binary. 619 620SunOS 4.x (Solaris 1.x) 621 You may have to use -lresolv on SunOS. However, beware that 622 this links in a new version of gethostbyname that does not 623 understand NIS, so you must have all of your hosts in DNS. 624 625 Some people have reported problems with the SunOS version of 626 -lresolv and/or in.named, and suggest that you get a newer 627 version. The symptoms are delays when you connect to the 628 SMTP server on a SunOS machine or having your domain added to 629 addresses inappropriately. There is a version of BIND 630 version 4.9 on gatekeeper.DEC.COM in pub/BSD/bind/4.9. 631 632 There is substantial disagreement about whether you can make 633 this work with resolv+, which allows you to specify a search-path 634 of services. Some people report that it works fine, others 635 claim it doesn't work at all (including causing sendmail to 636 drop core when it tries to do multiple resolv+ lookups for a 637 single job). I haven't tried resolv+, as we use DNS exclusively. 638 639 Should you want to try resolv+, it is on ftp.uu.net in 640 /networking/ip/dns. 641 642 Apparently getservbyname() can fail under moderate to high 643 load under some circumstances. This will exhibit itself as 644 the message ``554 makeconnection: service "smtp" unknown''. 645 The problem has been traced to one or more blank lines in 646 /etc/services on the NIS server machine. Delete these 647 and it should work. This info is thanks to Brian Bartholomew 648 <bb@math.ufl.edu> of I-Kinetics, Inc. 649 650SunOS 4.0.2 (Sun 386i) 651 Date: Fri, 25 Aug 1995 11:13:58 +0200 (MET DST) 652 From: teus@oce.nl 653 654 Sendmail 8.7.Beta.12 compiles and runs nearly out of the box with the 655 following changes: 656 * Don't use /usr/5bin in your PATH, but make /usr/5bin/uname 657 available as "uname" command. 658 * Use the defines "-DBSD4_3 -DNAMED_BIND=0" in 659 BuildTools/OS/SunOS.4.0, which is selected via the "uname" command. 660 I recommend to make available the db-library on the system first 661 (and change the Makefile to use this library). 662 Note that the sendmail.cf and aliases files are found in /etc. 663 664SunOS 4.1.3, 4.1.3_U1 665 Sendmail causes crashes on SunOS 4.1.3 and 4.1.3_U1. According 666 to Sun bug number 1077939: 667 668 If an application does a getsockopt() on a SOCK_STREAM (TCP) socket 669 after the other side of the connection has sent a TCP RESET for 670 the stream, the kernel gets a Bus Trap in the tcp_ctloutput() or 671 ip_ctloutput() routine. 672 673 For 4.1.3, this is fixed in patch 100584-08, available on the 674 Sunsolve 2.7.1 or later CDs. For 4.1.3_U1, this was fixed in patch 675 101790-01 (SunOS 4.1.3_U1: TCP socket and reset problems), later 676 obsoleted by patch 102010-05. 677 678 Sun patch 100584-08 is not currently publicly available on their 679 ftp site but a user has reported it can be found at other sites 680 using a web search engine. 681 682Solaris 2.x (SunOS 5.x) 683 To compile for Solaris, the Makefile built by Build must 684 include a SOLARIS definition which reflects the Solaris version 685 (i.e. -DSOLARIS=20400 for 2.4 or -DSOLARIS=20501 for 2.5.1). 686 If you are using gcc, make sure -I/usr/include is not used (or 687 it might complain about TopFrame). If you are using Sun's cc, 688 make sure /opt/SUNWspro/bin/cc is used instead of /usr/ucb/cc 689 (or it might complain about tm_zone). 690 691 To the best of my knowledge, Solaris does not have the 692 gethostbyname problem described above. However, it does 693 have another one: 694 695 From a correspondent: 696 697 For solaris 2.2, I have 698 699 hosts: files dns 700 701 in /etc/nsswitch.conf and /etc/hosts has to have the fully 702 qualified host name. I think "files" has to be before "dns" 703 in /etc/nsswitch.conf during bootup. 704 705 From another correspondent: 706 707 When running sendmail under Solaris, the gethostbyname() 708 hack in conf.c which should perform proper canonicalization 709 of host names could fail. Result: the host name is not 710 canonicalized despite the hack, and you'll have to define $j 711 and $m in sendmail.cf somewhere. 712 713 The reason could be that /etc/nsswitch.conf is improperly 714 configured (at least from sendmail's point of view). For 715 example, the line 716 717 hosts: files nisplus dns 718 719 will make gethostbyname() look in /etc/hosts first, then ask 720 nisplus, then dns. However, if /etc/hosts does not contain 721 the full canonicalized hostname, then no amount of 722 gethostbyname()s will work. 723 724 Solution (or rather, a workaround): Ask nisplus first, then 725 dns, then local files: 726 727 hosts: nisplus dns [NOTFOUND=return] files 728 729 The Solaris "syslog" function is apparently limited to something 730 about 90 characters because of a kernel limitation. If you have 731 source code, you can probably up this number. You can get patches 732 that fix this problem: the patch ids are: 733 734 Solaris 2.1 100834 735 Solaris 2.2 100999 736 Solaris 2.3 101318 737 738 Be sure you have the appropriate patch installed or you won't 739 see system logging. 740 741Solaris 2.4 (SunOS 5.4) 742 If you include /usr/lib at the end of your LD_LIBRARY_PATH you run 743 the risk of getting the wrong libraries under some circumstances. 744 This is because of a new feature in Solaris 2.4, described by 745 Rod.Evans@Eng.Sun.COM: 746 747 >> Prior to SunOS 5.4, any LD_LIBRARY_PATH setting was ignored by the 748 >> runtime linker if the application was setxid (secure), thus your 749 >> applications search path would be: 750 >> 751 >> /usr/local/lib LD_LIBRARY_PATH component - IGNORED 752 >> /usr/lib LD_LIBRARY_PATH component - IGNORED 753 >> /usr/local/lib RPATH - honored 754 >> /usr/lib RPATH - honored 755 >> 756 >> the effect is that path 3 would be the first used, and this would 757 >> satisfy your resolv.so lookup. 758 >> 759 >> In SunOS 5.4 we made the LD_LIBRARY_PATH a little more flexible. 760 >> People who developed setxid applications wanted to be able to alter 761 >> the library search path to some degree to allow for their own 762 >> testing and debugging mechanisms. It was decided that the only 763 >> secure way to do this was to allow a `trusted' path to be used in 764 >> LD_LIBRARY_PATH. The only trusted directory we presently define 765 >> is /usr/lib. Thus a setuid root developer could play with some 766 >> alternative shared object implementations and place them in 767 >> /usr/lib (being root we assume they'ed have access to write in this 768 >> directory). This change was made as part of 1155380 - after a 769 >> *huge* amount of discussion regarding the security aspect of things. 770 >> 771 >> So, in SunOS 5.4 your applications search path would be: 772 >> 773 >> /usr/local/lib from LD_LIBRARY_PATH - IGNORED (untrustworthy) 774 >> /usr/lib from LD_LIBRARY_PATH - honored (trustworthy) 775 >> /usr/local/lib from RPATH - honored 776 >> /usr/lib from RPATH - honored 777 >> 778 >> here, path 2 would be the first used. 779 780Solaris 2.6 (SunOS 5.6) 781 If you built sendmail 8.8.1 through 8.8.4 inclusive on a Solaris 2.5 782 system, that binary will not run on Solaris 2.6, due to problems with 783 incompatible snprintf(3s) calls. This problem is fixed in sendmail 784 8.8.5. 785 786Solaris 2.5.1 (SunOS 5.5.1) and 2.6 (SunOS 5.6) 787 Apparently Solaris 2.5.1 patch 103663-01 installs a new 788 /usr/include/resolv.h file that defines the __P macro without 789 checking to see if it is already defined. This new resolv.h is also 790 included in the Solaris 2.6 distribution. This causes compile 791 warnings such as: 792 793 In file included from daemon.c:51: 794 /usr/include/resolv.h:208: warning: `__P' redefined 795 cdefs.h:58: warning: this is the location of the previous definition 796 797 These warnings can be safely ignored or you can create a resolv.h 798 file in the obj.SunOS.5.5.1.* or obj.SunOS.5.6.* directory that reads: 799 800 #undef __P 801 #include "/usr/include/resolv.h" 802 803 Sun is aware of the problem (Sun bug ID 4081053) and it will be fixed 804 in Solaris 2.7. 805 806Ultrix 807 By default, the IDENT protocol is turned off on Ultrix. If you 808 are running Ultrix 4.4 or later, or if you have included patch 809 CXO-8919 for Ultrix 4.2 or 4.3 to fix the TCP problem, you can turn 810 IDENT on in the configuration file by setting the "ident" timeout 811 to 30 seconds. 812 813Digital UNIX (formerly DEC OSF/1) 814 If you are compiling on OSF/1 (DEC Alpha), you must use 815 -L/usr/shlib (otherwise it core dumps on startup). You may also 816 need -mld to get the nlist() function, although some versions 817 apparently don't need this. 818 819 Also, the enclosed makefile removed /usr/sbin/smtpd; if you need 820 it, just create the link to the sendmail binary. 821 822 On DEC OSF/1 3.2 or earlier, the MatchGECOS option doesn't work 823 properly due to a bug in the getpw* routines. If you want to use 824 this, use -DDEC_OSF_BROKEN_GETPWENT=1. The problem is fixed in 3.2C. 825 826 Digital's mail delivery agent, /bin/mail (aka /bin/binmail), will 827 only preserve the envelope sender in the "From " header if 828 DefaultUserID is set to daemon. Setting this to mailnull will 829 cause all mail to have the header "From mailnull ...". To use 830 a different DefaultUserID, you will need to use a different mail 831 delivery agent (such as mail.local found in the sendmail 832 distribution). 833 834 On Digital UNIX 4.0 and later, Berkeley DB 1.85 is included with the 835 operating system and already has the ndbm.o module removed. However, 836 Digital has modified the original Berkeley DB db.h include file. 837 This results in the following warning while compiling map.c and udb.c: 838 839 cc: Warning: /usr/include/db.h, line 74: The redefinition of the macro 840 "__signed" conflicts with a current definition because the replacement 841 lists differ. The redefinition is now in effect. 842 #define __signed signed 843 ------------------------^ 844 845 This warning can be ignored. 846 847IRIX 848 The header files on SGI IRIX are completely prototyped, and as 849 a result you can sometimes get some warning messages during 850 compilation. These can be ignored. There are two errors in 851 deliver only if you are using gcc, both of the form ``warning: 852 passing arg N of `execve' from incompatible pointer type''. 853 Also, if you compile with -DNIS, you will get a complaint 854 about a declaration of struct dom_binding in a prototype 855 when compiling map.c; this is not important because the 856 function being prototyped is not used in that file. 857 858 In order to compile sendmail you will have had to install 859 the developers' option in order to get the necessary include 860 files. 861 862 If you compile with -lmalloc (the fast memory allocator), you may 863 get warning messages such as the following: 864 865 ld32: WARNING 85: definition of _calloc in /usr/lib32/libmalloc.so 866 preempts that definition in /usr/lib32/mips3/libc.so. 867 ld32: WARNING 85: definition of _malloc in /usr/lib32/libmalloc.so 868 preempts that definition in /usr/lib32/mips3/libc.so. 869 ld32: WARNING 85: definition of _realloc in /usr/lib32/libmalloc.so 870 preempts that definition in /usr/lib32/mips3/libc.so. 871 ld32: WARNING 85: definition of _free in /usr/lib32/libmalloc.so 872 preempts that definition in /usr/lib32/mips3/libc.so. 873 ld32: WARNING 85: definition of _cfree in /usr/lib32/libmalloc.so 874 preempts that definition in /usr/lib32/mips3/libc.so. 875 876 These are unavoidable and innocuous -- just ignore them. 877 878 According to Dave Sill <de5@ornl.gov>, there is a version of the 879 Berkeley DB library patched to run on Irix 6.2 available from 880 http://reality.sgi.com/ariel/freeware/#db . 881 882IRIX 6.x 883 It is important that on IRIX 6.x you give used ABI in command 884 line of Build, otherwise configuration script does not work 885 correctly, e.g., 886 887 sh Build -E ABI=-n32 888 889 If you are using XFS filesystem, avoid using ABI=-32 if possible. 890 891NeXT or NEXTSTEP 892 NEXTSTEP 3.3 and earlier ship with the old DBM library. Also, 893 Berkeley DB does not currently run on NEXTSTEP. 894 895 If you are compiling on NEXTSTEP, you will have to create an 896 empty file "unistd.h" and create a file "dirent.h" containing: 897 898 #include <sys/dir.h> 899 #define dirent direct 900 901 (BuildTools/OS/NeXT should try to do both of these for you.) 902 903 Apparently, there is a bug in getservbyname on Nextstep 3.0 904 that causes it to fail under some circumstances with the 905 message "SYSERR: service "smtp" unknown" logged. You should 906 be able to work around this by including the line: 907 908 OOPort=25 909 910 in your .cf file. 911 912 You may have to use -DNeXT. 913 914BSDI (BSD/386) 1.0, NetBSD 0.9, FreeBSD 1.0 915 The "m4" from BSDI won't handle the config files properly. 916 I haven't had a chance to test this myself. 917 918 The M4 shipped in FreeBSD and NetBSD 0.9 don't handle the config 919 files properly. One must use either GNU m4 1.1 or the PD-M4 920 recently posted in comp.os.386bsd.bugs (and maybe others). 921 NetBSD-current includes the PD-M4 (as stated in the NetBSD file 922 CHANGES). 923 924 FreeBSD 1.0 RELEASE has uname(2) now. Use -DUSEUNAME in order to 925 use it (look into BuildTools/OS/FreeBSD). NetBSD-current may have 926 it too but it has not been verified. 927 928 The latest version of Berkeley DB uses a different naming 929 scheme than the version that is supplied with your release. This 930 means you will be able to use the current version of Berkeley DB 931 with sendmail as long you use the new db.h when compiling 932 sendmail and link it against the new libdb.a. You should probably 933 keep the original db.h in /usr/include and the new db.h in 934 /usr/local/include. 935 9364.3BSD 937 If you are running a "virgin" version of 4.3BSD, you'll have 938 a very old resolver and be missing some header files. The 939 header files are simple -- create empty versions and everything 940 will work fine. For the resolver you should really port a new 941 version (4.8.3 or later) of the resolver; 4.9 is available on 942 gatekeeper.DEC.COM in pub/BSD/bind/4.9. If you are really 943 determined to continue to use your old, buggy version (or as 944 a shortcut to get sendmail working -- I'm sure you have the 945 best intentions to port a modern version of BIND), you can 946 copy ../contrib/oldbind.compat.c into src and add 947 oldbind.compat.o to OBJADD in the Makefile. 948 949A/UX 950 Date: Tue, 12 Oct 1993 18:28:28 -0400 (EDT) 951 From: "Eric C. Hagberg" <hagberg@med.cornell.edu> 952 Subject: Fix for A/UX ndbm 953 954 I guess this isn't really a sendmail bug, however, it is something 955 that A/UX users should be aware of when compiling sendmail 8.6. 956 957 Apparently, the calls that sendmail is using to the ndbm routines 958 in A/UX 3.0.x contain calls to "broken" routines, in that the 959 aliases database will break when it gets "just a little big" 960 (sorry I don't have exact numbers here, but it broke somewhere 961 around 20-25 aliases for me.), making all aliases non-functional 962 after exceeding this point. 963 964 What I did was to get the gnu-dbm-1.6 package, compile it, and 965 then re-compile sendmail with "-lgdbm", "-DNDBM", and using the 966 ndbm.h header file that comes with the gnu-package. This makes 967 things behave properly. 968 [NOTE: see comment above about GDBM] 969 970 I suppose porting the New Berkeley DB package is another route, 971 however, I made a quick attempt at it, and found it difficult 972 (not easy at least); the gnu-dbm package "configured" and 973 compiled easily. 974 975 [NOTE: Berkeley DB version 2.X runs on A/UX and can be used for 976 database maps.] 977 978SCO Unix 979 From: Thomas Essebier <tom@stallion.oz.au> 980 Organisation: Stallion Technologies Pty Ltd. 981 982 It will probably help those who are trying to configure sendmail 8.6.9 983 to know that if they are on SCO, they had better set 984 OI-dnsrch 985 or they will core dump as soon as they try to use the resolver. 986 ie. although SCO has _res.dnsrch defined, and is kinda BIND 4.8.3, it 987 does not inititialise it, nor does it understand 'search' in 988 /etc/named.boot. 989 - sigh - 990 991 According to SCO, the m4 which ships with UnixWare 2.1.2 is broken. 992 We recommend installing GNU m4 before attempting to build sendmail. 993 994DG/UX 995 Doug Anderson <dlander@afterlife.ncsc.mil> has successfully run 996 V8 on the DG/UX 5.4.2 and 5.4R3.x platforms under heavy usage. 997 Originally, the DG /bin/mail program wasn't compatible with 998 the V8 sendmail, since the DG /bin/mail requires the environment 999 variable "_FORCE_MAIL_LOCAL_=yes" be set. Version 8.7 now includes 1000 this in the environment before invoking the local mailer. Some 1001 have used procmail to avoid this problem in the past. It works 1002 but some have experienced file locking problems with their DG/UX 1003 ports of procmail. 1004 1005Apollo DomainOS 1006 If you are compiling on Apollo, you will have to create an empty 1007 file "unistd.h" (for DomainOS 10.3 and earlier) and create a file 1008 "dirent.h" containing: 1009 1010 #include <sys/dir.h> 1011 #define dirent direct 1012 1013 (BuildTools/OS/DomainOS will attempt to do both of these for you.) 1014 1015HP-UX 8.00 1016 Date: Mon, 24 Jan 1994 13:25:45 +0200 1017 From: Kimmo Suominen <Kimmo.Suominen@lut.fi> 1018 Subject: 8.6.5 w/ HP-UX 8.00 on s300 1019 1020 Just compiled and fought with sendmail 8.6.5 on a HP9000/360 (ie. a 1021 series 300 machine) running HP-UX 8.00. 1022 1023 I was getting segmentation fault when delivering to a local user. 1024 With debugging I saw it was faulting when doing _free@libc... *sigh* 1025 It seems the new implementation of malloc on s300 is buggy as of 8.0, 1026 so I tried out the one in -lmalloc (malloc(3X)). With that it seems 1027 to work just dandy. 1028 1029 When linking, you will get the following error: 1030 1031 ld: multiply defined symbol _freespace in file /usr/lib/libmalloc.a 1032 1033 but you can just ignore it. You might want to add this info to the 1034 README file for the future... 1035 1036Linux 1037 Something broke between versions 0.99.13 and 0.99.14 of Linux: 1038 the flock() system call gives errors. If you are running .14, 1039 you must not use flock. You can do this with -DHASFLOCK=0. 1040 1041 Around the inclusion of bind-4.9.3 & Linux libc-4.6.20, the 1042 initialization of the _res structure changed. If /etc/hosts.conf 1043 was configured as "hosts, bind" the resolver code could return 1044 "Name server failure" errors. This is supposedly fixed in 1045 later versions of libc (>= 4.6.29?), and later versions of 1046 sendmail (> 8.6.10) try to work around the problem. 1047 1048 Some older versions (< 4.6.20?) of the libc/include files conflict 1049 with sendmail's version of cdefs.h. Deleting sendmail's version 1050 on those systems should be non-harmful, and new versions don't care. 1051 1052 Sendmail assumes that libc has snprintf, which has been true since 1053 libc 4.7.0. If you are running an older version, you will need to 1054 use -DHASSNPRINTF=0 in the Makefile. If may be able to use -lbsd 1055 (which includes snprintf) instead of turning this off on versions 1056 of libc between 4.4.4 and 4.7.0 (snprintf improves security, so 1057 you want to use this if at all possible). 1058 1059 NOTE ON LINUX & BIND: By default, the Makefile generated for Linux 1060 includes header files in /usr/local/include and libraries in 1061 /usr/local/lib. If you've installed BIND on your system, the header 1062 files typically end up in the search path and you need to add 1063 "-lresolv" to the LIBS line in your Makefile. Really old versions 1064 may need to include "-l44bsd" as well (particularly if the link phase 1065 complains about missing strcasecmp, strncasecmp or strpbrk). 1066 Complaints about an undefined reference to `__dn_skipname' in 1067 domain.o are a sure sign that you need to add -lresolv to LIBS. 1068 Newer versions of Linux are basically threaded BIND, so you may or 1069 may not see complaints if you accidentally mix BIND 1070 headers/libraries with virginal libc. If you have BIND headers in 1071 /usr/local/include (resolv.h, etc) you *should* be adding -lresolv 1072 to LIBS. Data structures may change and you'd be asking for a 1073 core dump. 1074 1075AIX 3.x 1076 This version of sendmail does not support MB, MG, and MR resource 1077 records, which are supported by AIX sendmail. 1078 1079 Several people have reported that the IBM-supplied named returns 1080 fairly random results -- the named should be replaced. It is not 1081 necessary to replace the resolver, which will simplify installation. 1082 A new BIND resolver can be found at http://www.isc.org/isc/. 1083 1084AIX 3.1.x 1085 The supplied load average code only works correctly for AIX 3.2.x. 1086 For 3.1, use -DLA_TYPE=LA_SUBR and get the latest ``monitor'' 1087 package by Jussi Maki <jmaki@hut.fi> from ftp.funet.fi in the 1088 directory pub/unix/AIX/rs6000/monitor-1.12.tar.Z; use the loadavgd 1089 daemon, and the getloadavg subroutine supplied with that package. 1090 If you don't care about load average throttling, just turn off 1091 load average checking using -DLA_TYPE=LA_ZERO. 1092 1093AIX 2.2.1 1094 Date: Mon Dec 4 14:14:56 CST 1995 1095 From: Mark Whetzel <markw@antimatr.houston.tx.us> 1096 Subject: Porting sendmail 8.7.2 to AIX V2 on the RT. 1097 1098 This version of sendmail does not support MB, MG, and MR resource 1099 records, which are supported by AIX sendmail. 1100 1101 AIX V2 on the RT does not have 'paths.h'. Create a null 1102 file in the 'obj' directory to remove this compile error. 1103 1104 A patch file is needed to get the BSD 'db' library to compile 1105 for AIX/RT. I have sent the necessary updates to the author, 1106 but they may not be immediately available. 1107 [NOTE: Berkeley DB version 2.X runs on AIX/RT.] 1108 1109 The original AIX/RT resolver libraries are very old, and you 1110 should get the latest BIND to replace it. The 4.8.3 version 1111 has been tested, but 4.9.x is out and should work. 1112 1113 To make the load average code work correctly requires an 1114 external routine, as the kernel does not maintain system 1115 load averages, similar to AIX V3.1.x. A reverse port of the 1116 older 1.05 'monitor' load average daemon code written by 1117 Jussi Maki that will work on AIX V2 for the RT is available 1118 by E-mail to Mark Whetzel <markw@antimatr.houston.tx.us>. 1119 That code depends on an external daemon to collect system 1120 load information, and the external routine 'getloadavg', 1121 that will return that information. The 'LA_SUBR' define 1122 will handle this for AIX V2 on the RT. 1123 1124 Note: You will have to change BuildTools/OS/AIX.2 to correctly 1125 point to the locatons of the updated BIND source tree and 1126 the location of the 'newdb' tree and library location. 1127 You will also have to change BuildTools/OS/AIX.2 to know 1128 about the location of the 'getloadavg' routine if you use 1129 the LA_SUBR define. 1130 1131 1132 Manual pages will format correctly if given the mandoc macros 1133 and used with nroff. I have not tried groff. 1134 1135RISC/os 1136 RISC/os from MIPS is a merged AT&T/Berkeley system. When you 1137 compile on that platform you will get duplicate definitions 1138 on many files. You can ignore these. 1139 1140System V Release 4 Based Systems 1141 There is a single BuildTools OS that is intended for all SVR4-based 1142 systems (built from BuildTools/OS/SVR4). It defines __svr4__, 1143 which is predefined by some compilers. If your compiler already 1144 defines this compile variable, you can delete the definition from 1145 the generated Makefile or create a BuildTools/Site/site.config.m4 1146 file. 1147 1148 It's been tested on Dell Issue 2.2. 1149 1150DELL SVR4 1151 Date: Mon, 06 Dec 1993 10:42:29 EST 1152 From: "Kimmo Suominen" <kim@grendel.lut.fi> 1153 Message-ID: <2d0352f9.lento29@lento29.UUCP> 1154 To: eric@cs.berkeley.edu 1155 Cc: sendmail@cs.berkeley.edu 1156 Subject: Notes for DELL SVR4 1157 1158 Eric, 1159 1160 Here are some notes for compiling Sendmail 8.6.4 on DELL SVR4. I ran 1161 across these things when helping out some people who contacted me by 1162 e-mail. 1163 1164 1) Use gcc 2.4.5 (or later?). Dell distributes gcc 2.1 with their 1165 Issue 2.2 Unix. It is too old, and gives you problems with 1166 clock.c, because sigset_t won't get defined in <sys/signal.h>. 1167 This is due to a problematic protection rule in there, and is 1168 fixed with gcc 2.4.5. 1169 1170 2) If you don't use the new Berkeley DB (-DNEWDB), then you need 1171 to add "-lc -lucb" to the libraries to link with. This is because 1172 the -ldbm distributed by Dell needs the bcopy, bcmp and bzero 1173 functions. It is important that you specify both libraries in 1174 the given order to be sure you only get the BSTRING functions 1175 from the UCB library (and not the signal routines etc.). 1176 1177 3) Don't leave out "-lelf" even if compiling with "-lc -lucb". 1178 The UCB library also has another copy of the nlist routines, 1179 but we do want the ones from "-lelf". 1180 1181 If anyone needs a compiled gcc 2.4.5 and/or a ported DB library, they 1182 can use anonymous ftp to fetch them from lut.fi in the /kim directory. 1183 They are copies of what I use on grendel.lut.fi, and offering them 1184 does not imply that I would also support them. I have sent the DB 1185 port for SVR4 back to Keith Bostic for inclusion in the official 1186 distribution, but I haven't heard anything from him as of today. 1187 1188 - gcc-2.4.5-svr4.tar.gz (gcc 2.4.5 and the corresponding libg++) 1189 - db-1.72.tar.gz (with source, objects and a installed copy) 1190 1191 Cheers 1192 + Kim 1193 -- 1194 * Kimmo.Suominen@lut.fi * SysVr4 enthusiast at GRENDEL.LUT.FI * 1195 * KIM@FINFILES.BITNET * Postmaster and Hostmaster at LUT.FI * 1196 * + 358 200 865 718 * Unix area moderator at NIC.FUNET.FI * 1197 1198ConvexOS 10.1 and below 1199 In order to use the name server, you must create the file 1200 /etc/use_nameserver. If this file does not exist, the call 1201 to res_init() will fail and you will have absolutely no 1202 access to DNS, including MX records. 1203 1204Amdahl UTS 2.1.5 1205 In order to get UTS to work, you will have to port BIND 4.9. 1206 The vendor's BIND is reported to be ``totally inadequate.'' 1207 See sendmail/contrib/AmdahlUTS.patch for the patches necessary 1208 to get BIND 4.9 compiled for UTS. 1209 1210UnixWare 1211 According to Alexander Kolbasov <sasha@unitech.gamma.ru>, 1212 the m4 on UnixWare 2.0 (still in Beta) will core dump on the 1213 config files. GNU m4 and the m4 from UnixWare 1.x both work. 1214 1215 According to Larry Rosenman <ler@lerami.lerctr.org>: 1216 1217 UnixWare 2.1.[23]'s m4 chokes (not obviously) when 1218 processing the 8.9.0 cf files. 1219 1220 I had a LOCAL_RULE_0 that wound up AFTER the 1221 SBasic_check_rcpt rules using the SCO supplied M4. 1222 GNU M4 works fine. 1223 1224UNICOS 8.0.3.4 1225 Some people have reported that the -O flag on UNICOS can cause 1226 problems. You may want to turn this off if you have problems 1227 running sendmail. Reported by Jerry G. DeLapp <jgd@acl.lanl.gov>. 1228 1229GNU getopt 1230 I'm told that GNU getopt has a problem in that it gets confused 1231 by the double call. Use the version in conf.c instead. 1232 1233BIND 4.9.2 and Ultrix 1234 If you are running on Ultrix, be sure you read conf/Info.Ultrix 1235 in the BIND distribution very carefully -- there is information 1236 in there that you need to know in order to avoid errors of the 1237 form: 1238 1239 /lib/libc.a(gethostent.o): sethostent: multiply defined 1240 /lib/libc.a(gethostent.o): endhostent: multiply defined 1241 /lib/libc.a(gethostent.o): gethostbyname: multiply defined 1242 /lib/libc.a(gethostent.o): gethostbyaddr: multiply defined 1243 1244 during the link stage. 1245 1246strtoul 1247 Some compilers (notably gcc) claim to be ANSI C but do not 1248 include the ANSI-required routine "strtoul". If your compiler 1249 has this problem, you will get an error in srvrsmtp.c on the 1250 code: 1251 1252 # ifdef defined(__STDC__) && !defined(BROKEN_ANSI_LIBRARY) 1253 e->e_msgsize = strtoul(vp, (char **) NULL, 10); 1254 # else 1255 e->e_msgsize = strtol(vp, (char **) NULL, 10); 1256 # endif 1257 1258 You can use -DBROKEN_ANSI_LIBRARY to get around this problem. 1259 1260Listproc 6.0c 1261 Date: 23 Sep 1995 23:56:07 GMT 1262 Message-ID: <95925101334.~INN-AUMa00187.comp-news@dl.ac.uk> 1263 From: alansz@mellers1.psych.berkeley.edu (Alan Schwartz) 1264 Subject: Listproc 6.0c + Sendmail 8.7 [Helpful hint] 1265 1266 Just upgraded to sendmail 8.7, and discovered that listproc 6.0c 1267 breaks, because it, by default, sends a blank "HELO" rather than 1268 a "HELO hostname" when using the 'system' or 'telnet' mailmethod. 1269 1270 The fix is to include -DZMAILER in the compilation, which will 1271 cause it to use "HELO hostname" (which Z-mail apparently requires 1272 as well. :) 1273 1274LDAP 1275 LDAP was provided by Booker Bense <bbense+ldap@stanford.edu> of 1276 Stanford University. From Booker: 1277 1278 - The patch attached to this message implements an Ldap map class. 1279 Currently we are using this at stanford to support campus-wide 1280 email addressing. More information can be found at 1281 http://www.stanford.edu/~bbense/Inst.html. 1282 1283 - Currently we are using the ldap map as follows: 1284 1285 Kluser ldapx 1286 -h"localhost borax.stanford.edu borate.stanford.edu boron.stanford.edu" 1287 -k"mailacceptinggeneralid=%s" -v maildrop 1288 1289 and in Rule set S5 1290 1291 # Now attempt to lookup in luser (ldap map) 1292 R< $L > $+ $: < $L > $( luser $1 $) 1293 R< $* > $+ @ $+ $: < $3 > $2 Rewrite if forward 1294 1295 - The map definition supports most of the standard Map args plus most 1296 of the command line options of ldapsearch. The software is currently 1297 limited to only accepting the first entry returned. It expects that 1298 the map defines an ldap filter that returns at most 1 valid entry. 1299 It requires the ldap and lber libraries from the Umich Ldap3.2 1300 release. 1301 1302 The software has been in production on Solaris.2.5.1 at Stanford 1303 for over 2 years. 1304 1305TCP Wrappers 1306 If you are using -DTCPWRAPPERS to get TCP Wrappers support you will 1307 also need to install libwrap.a and modify your site.config.m4 file 1308 or the generated Makefile to include -lwrap in the LIBS line 1309 (make sure that INCDIRS and LIBDIRS point to where the tcpd.h and 1310 libwrap.a can be found). 1311 1312 TCP Wrappers is available on ftp.win.tue.nl in /pub/security; 1313 grab tcp_wrappers_<VER>.tar.gz (where <VER> is the highest 1314 numbered version). 1315 1316 If you have alternate MX sites for your site, be sure that all of 1317 your MX sites reject the same set of hosts. If not, a bad guy whom 1318 you reject will connect to your site, fail, and move on to the next 1319 MX site, which will accept the mail for you and forward it on to you. 1320 1321Regular Expressions (MAP_REGEX) 1322 If sendmail linking fails with: 1323 1324 undefined reference to 'regcomp' 1325 1326 or sendmail gives an error about a regular expression with: 1327 1328 pattern-compile-error: : Operation not applicable 1329 1330 Your libc does not include a running version of POSIX-regex. Use 1331 librx or regex.o from the GNU Free Software Foundation, 1332 ftp://ftp.gnu.org/pub/gnu/rx-?.?.tar.gz or 1333 ftp://ftp.gnu.org/pub/gnu/regex-?.?.tar.gz. 1334 You can also use the regex-lib by Henry Spencer, 1335 ftp://ftp.funet.fi/pub/languages/C/spencer/regex.shar.gz 1336 Make sure, your compiler reads regex.h from the distribution, 1337 not from /usr/include, otherwise sendmail will dump a core. 1338 1339 1340+--------------+ 1341| MANUAL PAGES | 1342+--------------+ 1343 1344The manual pages have been written against the -mandoc macros 1345instead of the -man macros. The latest version of groff has them 1346included. You can also get a copy from FTP.UU.NET in the directory 1347/systems/unix/bsd-sources/share/tmac. groff is available from 1348ftp.gnu.org in the /pub/gnu directory. 1349 1350 1351+-----------------+ 1352| DEBUGGING HOOKS | 1353+-----------------+ 1354 1355As of 8.6.5, sendmail daemons will catch a SIGUSR1 signal and log 1356some debugging output (logged at LOG_DEBUG severity). The 1357information dumped is: 1358 1359 * The value of the $j macro. 1360 * A warning if $j is not in the set $=w. 1361 * A list of the open file descriptors. 1362 * The contents of the connection cache. 1363 * If ruleset 89 is defined, it is evaluated and the results printed. 1364 1365This allows you to get information regarding the runtime state of the 1366daemon on the fly. This should not be done too frequently, since 1367the process of rewriting may lose memory which will not be recovered. 1368Also, ruleset 89 may call non-reentrant routines, so there is a small 1369non-zero probability that this will cause other problems. It is 1370really only for debugging serious problems. 1371 1372A typical formulation of ruleset 89 would be: 1373 1374 R$* $@ $>0 some test address 1375 1376 1377+-----------------------------+ 1378| DESCRIPTION OF SOURCE FILES | 1379+-----------------------------+ 1380 1381The following list describes the files in this directory: 1382 1383Makefile.m4 A template for constructing a makefile based on the 1384 information in the BuildTools directory. 1385README This file. 1386TRACEFLAGS My own personal list of the trace flags -- not guaranteed 1387 to be particularly up to date. 1388alias.c Does name aliasing in all forms. 1389arpadate.c A subroutine which creates ARPANET standard dates. 1390clock.c Routines to implement real-time oriented functions 1391 in sendmail -- e.g., timeouts. 1392collect.c The routine that actually reads the mail into a temp 1393 file. It also does a certain amount of parsing of 1394 the header, etc. 1395conf.c The configuration file. This contains information 1396 that is presumed to be quite static and non- 1397 controversial, or code compiled in for efficiency 1398 reasons. Most of the configuration is in sendmail.cf. 1399conf.h Configuration that must be known everywhere. 1400convtime.c A routine to sanely process times. 1401daemon.c Routines to implement daemon mode. This version is 1402 specifically for Berkeley 4.1 IPC. 1403deliver.c Routines to deliver mail. 1404domain.c Routines that interface with DNS (the Domain Name 1405 System). 1406err.c Routines to print error messages. 1407envelope.c Routines to manipulate the envelope structure. 1408headers.c Routines to process message headers. 1409macro.c The macro expander. This is used internally to 1410 insert information from the configuration file. 1411main.c The main routine to sendmail. This file also 1412 contains some miscellaneous routines. 1413map.c Support for database maps. 1414mci.c Routines that handle mail connection information caching. 1415mime.c MIME conversion routines. 1416parseaddr.c The routines which do address parsing. 1417queue.c Routines to implement message queueing. 1418readcf.c The routine that reads the configuration file and 1419 translates it to internal form. 1420recipient.c Routines that manipulate the recipient list. 1421safefile.c Routines to do careful checking of file modes and permissions 1422 when opening or creating files. 1423savemail.c Routines which save the letter on processing errors. 1424sendmail.h Main header file for sendmail. 1425snprintf.c Routines to manipulate strings but prevent buffer overflows. 1426srvrsmtp.c Routines to implement server SMTP. 1427stab.c Routines to manage the symbol table. 1428stats.c Routines to collect and post the statistics. 1429sysexits.c List of error messages associated with error codes 1430 in sysexits.h. 1431trace.c The trace package. These routines allow setting and 1432 testing of trace flags with a high granularity. 1433udb.c The user database interface module. 1434usersmtp.c Routines to implement user SMTP. 1435util.c Some general purpose routines used by sendmail. 1436version.c The version number and information about this 1437 version of sendmail. Theoretically, this gets 1438 modified on every change. 1439 1440Eric Allman 1441 1442(Version 8.206, last update 6/30/98 22:08:36) 1443