1# Copyright (c) 1983 Eric P. Allman 2# Copyright (c) 1988 The Regents of the University of California. 3# All rights reserved. 4# 5# %sccs.include.redist.sh% 6# 7# @(#)READ_ME 8.40 (Berkeley) 12/24/93 8# 9 10This directory contains the source files for sendmail. 11 12For detailed instructions, please read the document ../doc/op.me: 13 14 eqn ../doc/op.me | pic | ditroff -me 15 16The Makefile is for the new (4.4BSD) Berkeley make, available from 17ftp.uu.net in the directory /systems/unix/bsd-sources/usr.bin/make. 18(Paul Southworth <pauls@umich.edu> published a description of porting 19this make in comp.unix.bsd.) This Makefile has assumptions about the 204.4 file system layout built in. 21 22There is also a Makefile.dist which is much less clever, but works on 23the old traditional make. You can use this using: 24 25 make -f Makefile.dist 26 27 <<<<<<<<<<<<<<<<<<<< IMPORTANT >>>>>>>>>>>>>>>>>>>> 28There are a bunch of other Makefiles for other systems -- these are 29the ones that I use, they have "Berkeley quirks" in them, and I don't 30guarantee that they will work unmodified in your environment. However, 31they are all designed for the old make and can be used to help you get 32started. They have names like "Makefile.HPUX". Many of them include 33-I/usr/sww/include/db and -L/usr/sww/lib -- this is Berkeley's 34location for the new database libraries, described below. You don't 35have to remove these definitions if you don't have these directories. 36Please look for an appropriate Makefile before you start trying to 37compile with Makefile or Makefile.dist. 38 <<<<<<<<<<<<<<<<<<<<<<<<<<*>>>>>>>>>>>>>>>>>>>>>>>>>> 39 40There is also a shell script (makesendmail) that tries to be clever 41about using object subdirectories. It's pretty straightforward, and 42may help if you share a source tree among different architectures. 43 44************************************************************************** 45** IMPORTANT: DO NOT USE OPTIMIZATION (``-O'') IF YOU ARE RUNNING ** 46** GCC 2.4.x or 2.5.x. THERE IS A BUG IN THE GCC OPTIMIZER THAT ** 47** CAUSES SENDMAIL COMPILES TO FAIL MISERABLY. ** 48************************************************************************** 49 50Jim Wilson of Cygnus believes he has found the problem -- it will 51probably be fixed in GCC 2.5.6 -- but until this is verified, be 52very suspicious of gcc -O. 53 54************************************************************************** 55** IMPORTANT: Read the appropriate paragraphs in the section on ** 56** ``Operating System and Compile Quirks''. ** 57************************************************************************** 58 59 60+----------------------+ 61| DATABASE DEFINITIONS | 62+----------------------+ 63 64There are several database formats that can be used for the alias files 65and for general maps. When used for alias files they interact in an 66attempt to be back compatible. 67 68The three options are NEWDB (the new Berkeley DB package), NDBM (the 69older DBM implementation -- the very old V7 implementation is no 70longer supported), and NIS (Network Information Services). Used alone 71these just include the support they indicate. [If you are using NEWDB, 72get the latest version from FTP.CS.Berkeley.EDU in /ucb/4bsd. DO NOT 73use the version from the Net2 distribution! However, if you are on 74BSD/386 or 386BSD-based systems, use the one that already exists 75on your system. You may need to define OLD_NEWDB to do this.] 76 77If NEWDB and NDBM are defined (but not NIS), then sendmail will read 78NDBM format alias files, but the next time a newaliases is run the 79format will be converted to NEWDB; that format will be used forever 80more. This is intended as a transition feature. [Note however that 81the NEWDB library also catches and maps NDBM calls; you will have to 82back out this feature to get this to work. See ``Quirks'' section 83below for details.] 84 85If all three are defined, sendmail operates as described above, and also 86looks for the file /var/yp/Makefile. If it exists, newaliases will 87build BOTH the NEWDB and NDBM format alias files. However, it will 88only use the NEWDB file; the NDBM format file is used only by the 89NIS subsystem. 90 91If NDBM and NIS are defined (regardless of the definition of NEWDB 92or the existance of /var/yp/Makefile), sendmail adds the special 93tokens "YP_LAST_MODIFIED" and "YP_MASTER_NAME", both of which are 94required if the NDBM file is to be used as an NIS map. 95 96All of -DNEWDB, -DNDBM, and -DNIS are normally defined in the DBMDEF 97line in the Makefile. 98 99 100+---------------+ 101| COMPILE FLAGS | 102+---------------+ 103 104Whereever possible, I try to make sendmail pull in the correct 105compilation options needed to compile on various environments based on 106automatically defined symbols. Some machines don't seem to have useful 107symbols availble, requiring the following compilation flags in the 108Makefile: 109 110SOLARIS Define this if you are running Solaris 2.0 or higher. 111SOLARIS_2_3 Define this if you are running Solaris 2.3 or higher. 112SUNOS403 Define this if you are running SunOS 4.0.3. 113NeXT Define this if you are on a NeXT box. (This one may 114 be pre-defined for you.) There are other hacks you 115 have to make -- see below. 116_AIX3 Define this if you are IBM AIX 3.x. 117RISCOS Define this if you are running RISC/os from MIPS. 118_SCO_unix_ Define this if you are on SCO UNIX. 119_SCO_unix_4_2 Define this if you are on SCO Open Server 3.2v4. 120 121If you are a system that sendmail has already been ported to, you 122probably won't have to touch these. But if you are porting, you may 123have to tweak the following compilation flags in conf.h in order to 124get it to compile and link properly: 125 126SYSTEM5 Adjust for System V (not necessarily Release 4). 127SYS5SIGNALS Use System V signal semantics -- the signal handler 128 is automatically dropped when the signal is caught. 129 If this is not set, use POSIX/BSD semantics, where the 130 signal handler stays in force until an exec or an 131 explicit delete. Implied by SYSTEM5. 132SYS5SETPGRP Use System V setpgrp() semantics. Implied by SYSTEM5. 133HASFLOCK Set this if you prefer to use the flock(2) system call 134 rather than using fcntl-based locking. Fcntl locking 135 has some semantic gotchas, but many vendor systems 136 also interface it to lockd(8) to do NFS-style locking. 137 For this reason, this should not be set unless you 138 don't have an alternative. 139HASUNAME Set if you have the "uname" system call. Implied by 140 SYSTEM5. 141HASUNSETENV Define this if your system library has the "unsetenv" 142 subroutine. 143HASSTATFS Define this if you have the statfs(2) system call. It's 144 not a disaster to get this wrong -- but you do lose the 145 queue free space code. 146HASUSTAT Define this if you have the ustat(2) system call. It's 147 not a disaster to get this wrong -- but you do lose the 148 queue free space code. 149HASSETSID Define this if you have the setsid(2) system call. This 150 is implied if your system appears to be POSIX compliant. 151HASINITGROUPS Define this if you have the initgroups(3) routine. 152HASSETVBUF Define this if you have the setvbuf(3) library call. 153 If you don't, setlinebuf will be used instead. This 154 defaults on if your compiler defines __STDC__. 155HASSETREUID Define this if you have setreuid(2) ***AND*** root can 156 use setreuid to change to an arbitrary user. This second 157 condition is not satisfied on AIX 3.x. You may find that 158 your system has setresuid(2), (for example, on HP-UX) in 159 which case you will also have to #define setreuid(r, e) 160 to be the appropriate call. Some systems (such as Solaris) 161 have a compatibility routine that doesn't work properly, 162 but may have "saved user ids" properly implemented so you 163 can ``#define setreuid(r, e) seteuid(e)'' and have it work. 164 The important thing is that you have a call that will set 165 the effective uid independently of the real or saved uid 166 and be able to set the effective uid back again when done. 167 There's a test program in ../test/t_setreuid.c that will 168 try things on your system. Setting this improves the 169 security, since sendmail doesn't have to read .forward 170 and :include: files as root. There are certain attacks 171 that may be unpreventable without this call. 172HASLSTAT Define this if you have symbolic links (and thus the 173 lstat(2) system call). This improves security. Unlike 174 most other options, this one is on by default, so you 175 need to #undef it in conf.h if you don't have symbolic 176 links (these days everyone does). 177NEEDGETOPT Define this if you need a reimplementation of getopt(3). 178 On some systems, getopt does very odd things if called 179 to scan the arguments twice. This flag will ask sendmail 180 to compile in a local version of getopt that works 181 properly. 182NEEDSTRTOL Define this if your standard C library does not define 183 strtol(3). This will compile in a local version. 184NEEDVPRINTF Define this if your standard C library does not define 185 vprintf(3). Note that the resulting fake implementation 186 is not very elegant and may not even work on some 187 architectures. 188NEEDGETUSERSHELL 189 Define this if you do not have getusershell(3) in your 190 standard C library. Currently it does not compile in 191 a new version -- instead, it just deletes the call; 192 a future version will supply a replacement implementation. 193GIDSET_T The type of entries in a gidset passed as the second 194 argument to getgroups(2). Historically this has been an 195 int, so this is the default, but some systems (such as 196 IRIX) pass it as a gid_t, which is an unsigned short. 197 This will make a difference, so it is important to get 198 this right! However, it is only an issue if you have 199 group sets. 200SLEEP_T The type returned by the system sleep() function. 201 Defaults to "unsigned int". Don't worry about this 202 if you don't have compilation problems. 203ARBPTR_T The type of an arbitrary pointer -- defaults to "void *". 204 If you are an very old compiler you may need to define 205 this to be "char *". 206LA_TYPE The type of load average your kernel supports. These 207 can be LA_SUBR (4) if you have the getloadavg(3) routine, 208 LA_MACH (5) to use MACH-style load averages (calls 209 processor_set_info()), LA_FLOAT (3) if you read kmem and 210 interpret the value as a floating point number, LA_INT (2) 211 to interpret as a long integer, or LA_SHORT (6) to 212 interpret as a short integer. These last three have 213 several other parameters that they try to divine: the 214 name of your kernel, the name of the variable in the 215 kernel to examine, the number of bits of precision in 216 a fixed point load average, and so forth. In desparation, 217 use LA_ZERO (1) -- it always returns the load average as 218 "zero" (and does so on all architectures). The actual 219 code is in conf.c -- it can be tweaked if you are brave. 220ERRLIST_PREDEFINED 221 If set, assumes that some header file defines sys_errlist. 222 This may be needed if you get type conflicts on this 223 variable -- otherwise don't worry about it. 224WAITUNION The wait(2) routine takes a "union wait" argument instead 225 of an integer argument. This is for compatibility with 226 old versions of BSD. 227SCANF You can set this to extend the F command to accept a 228 scanf string -- this gives you a primitive parser for 229 class definitions -- BUT it can make you vulnerable to 230 core dumps if the target file is poorly formed. 231SYSLOG_BUFSIZE You can define this to be the size of the buffer that 232 syslog accepts. If it is not defined, it assumes a 233 1024-byte buffer. If the buffer is very small (under 234 256 bytes) the log message format changes -- each 235 e-mail message will log many more messages, since it 236 will log each piece of information as a separate line 237 in syslog. 238 239 240+-----------------------+ 241| COMPILE-TIME FEATURES | 242+-----------------------+ 243 244There are a bunch of features that you can decide to compile in, such 245as selecting various database packages and special protocol support. 246Several are assumed based on other compilation flags -- if you want to 247"un-assume" something, you probably need to edit conf.h. Compilation 248flags that add support for special features include: 249 250NDBM Include support for "new" DBM library for aliases and maps. 251 Normally defined in the Makefile. 252NEWDB Include support for Berkeley "db" package (hash & btree) 253 for aliases and maps. Normally defined in the Makefile. 254NIS Define this to get NIS (YP) support for aliases and maps. 255 Normally defined in the Makefile. 256USERDB Include support for the User Information Database. Implied 257 by NEWDB in conf.h. 258IDENTPROTO Define this as 1 to get IDENT (RFC 1413) protocol support. 259 This is assumed unless you are running on Ultrix or 260 HP-UX, both of which have a problem in the UDP 261 implementation. You can define it to be 0 to explicitly 262 turn off IDENT protocol support. 263MIME Include support for MIME-encapsulated error messages. 264LOG Set this to get syslog(3) support. Defined by default 265 in conf.h. You want this if at all possible. 266NETINET Set this to get TCP/IP support. Defined by default 267 in conf.h. You probably want this. 268NETISO Define this to get ISO networking support. 269SMTP Define this to get the SMTP code. Implied by NETINET 270 or NETISO. 271NAMED_BIND Define this to get DNS (name daemon) support, including 272 MX support. The specs you must use this if you run 273 SMTP. Defined by default in conf.h. 274QUEUE Define this to get queueing code. Implied by NETINET 275 or NETISO; required by SMTP. This gives you other good 276 stuff -- it should be on. 277DAEMON Define this to get general network support. Implied by 278 NETINET or NETISO. Defined by default in conf.h. You 279 almost certainly want it on. 280MATCHGECOS Permit fuzzy matching of user names against the full 281 name (GECOS) field in the /etc/passwd file. This should 282 probably be on, since you can disable it from the config 283 file if you want to. Defined by default in conf.h. 284SETPROCTITLE Try to set the string printed by "ps" to something 285 informative about what sendmail is doing. Defined by 286 default in conf.h. 287 288 289+---------------------+ 290| DNS/RESOLVER ISSUES | 291+---------------------+ 292 293Many systems have old versions of the resolver library. At a minimum, 294you should be running BIND 4.8.3; older versions may compile, but they 295have known bugs that should give you pause. 296 297Common problems in old versions include "undefined" errors for 298dn_skipname. 299 300Some people have had a problem with BIND 4.9; it uses some routines 301that it expects to be externally defined such as strerror(). It may 302help to link with "-l44bsd" to solve this problem. 303 304!PLEASE! be sure to link with the same version of the resolver as 305the header files you used -- some people have used the 4.9 headers 306and linked with BIND 4.8 or vice versa, and it doesn't work. 307Unfortunately, it doesn't fail in an obvious way -- things just 308subtlely don't work. 309 310 311+-------------------------------------+ 312| OPERATING SYSTEM AND COMPILE QUIRKS | 313+-------------------------------------+ 314 315GCC 2.5.x problems *** IMPORTANT *** 316 Date: Mon, 29 Nov 93 19:08:44 PST 317 From: wilson@cygnus.com (Jim Wilson) 318 Message-Id: <9311300308.AA04608@cygnus.com> 319 To: kenner@vlsi1.ultra.nyu.edu 320 Subject: [cattelan@thebarn.com: gcc 2.5.4-2.5.5 -O bug] 321 Cc: cattelan@thebarn.com, rms@gnu.ai.mit.edu, sendmail@cs.berkeley.edu 322 323 This fixes a problem that occurs when gcc 2.5.5 is used to compile 324 sendmail 8.6.4 with optimization on a sparc. 325 326 Mon Nov 29 19:00:14 1993 Jim Wilson (wilson@sphagnum.cygnus.com) 327 328 * reload.c (find_reloads_toplev): Replace obsolete reference to 329 BYTE_LOADS_*_EXTEND with LOAD_EXTEND_OP. 330 331 *** clean-ss-931128/reload.c Sun Nov 14 16:20:01 1993 332 --- ss-931128/reload.c Mon Nov 29 18:52:55 1993 333 *************** find_reloads_toplev (x, opnum, type, ind 334 *** 3888,3894 **** 335 force a reload in that case. So we should not do anything here. */ 336 337 else if (regno >= FIRST_PSEUDO_REGISTER 338 ! #if defined(BYTE_LOADS_ZERO_EXTEND) || defined(BYTE_LOADS_SIGN_EXTEND) 339 && (GET_MODE_SIZE (GET_MODE (x)) 340 <= GET_MODE_SIZE (GET_MODE (SUBREG_REG (x)))) 341 #endif 342 --- 3888,3894 ---- 343 force a reload in that case. So we should not do anything here. */ 344 345 else if (regno >= FIRST_PSEUDO_REGISTER 346 ! #ifdef LOAD_EXTEND_OP 347 && (GET_MODE_SIZE (GET_MODE (x)) 348 <= GET_MODE_SIZE (GET_MODE (SUBREG_REG (x)))) 349 #endif 350 351 352SunOS 4.x (Solaris 1.x) 353 You may have to use -lresolv on SunOS. However, beware that 354 this links in a new version of gethostbyname that does not 355 understand NIS, so you must have all of your hosts in DNS. 356 357 Some people have reported problems with the SunOS version of 358 -lresolv and/or in.named, and suggest that you get a newer 359 version. The symptoms are delays when you connect to the 360 SMTP server on a SunOS machine or having your domain added to 361 addresses inappropriately. There is a version of BIND 362 version 4.9 on gatekeeper.DEC.COM in pub/BSD/bind/4.9. 363 364 There is substantial disagreement about whether you can make 365 this work with resolv+, which allows you to specify a search-path 366 of services. Some people report that it works fine, others 367 claim it doesn't work at all (including causing sendmail to 368 drop core when it tries to do multiple resolv+ lookups for a 369 single job). I haven't tried resolv+, as we use DNS exclusively. 370 371 Should you want to try resolv+, it is on ftp.uu.net in 372 /networking/ip/dns. 373 374Solaris 2.x (SunOS 5.x) 375 To compile for Solaris, be sure you use -DSOLARIS. 376 377 From a correspondent: 378 379 For solaris 2.2, I have 380 381 hosts: files dns 382 383 in /etc/nsswitch.conf and /etc/hosts has to have the fully 384 qualified host name. I think "files" has to be before "dns" 385 in /etc/nsswitch.conf during bootup. 386 387 To the best of my knowledge, Solaris does not have the 388 gethostbyname problem described above. 389 390 The Solaris "syslog" function is apparently limited to something 391 about 90 characters because of a kernel limitation. If you have 392 source code, you can probably up this number. The syslogd patch 393 is included in kernel jumbo patch for Solaris 2.2 as of revision 394 -39 or so. At least one person is running with patch 100999-45 395 and their long lost sendmail logging is finally showing up. At 396 least one other person is running with patch 101318 installed 397 under Solaris 2.3 with success. 398 399OSF/1 400 If you are compiling on OSF/1 (DEC Alpha), you must use 401 -non_shared (otherwise it core dumps on startup). You may also 402 need -mld to get the nlist() function, although some versions 403 apparently don't need this. 404 405 Also, the enclosed makefile removed /usr/sbin/smtpd; if you need 406 it, just create the link to the sendmail binary. 407 408NeXT 409 If you are compiling on NeXT, you will have to create an empty 410 file "unistd.h" and create a file "dirent.h" containing: 411 412 #include <sys/dir.h> 413 #define dirent direct 414 415 (The Makefile.NeXT should try to do both of these for you.) 416 417 Apparently, there is a bug in getservbyname on Nextstep 3.0 418 that causes it to fail under some circumstances with the 419 message "SYSERR: service "smtp" unknown" logged. You should 420 be able to work around this by including the line: 421 422 OOPort=25 423 424 in your .cf file. 425 426 You may have to use -DNeXT. 427 428BSDI (BSD/386) 1.0, NetBSD 0.9, FreeBSD 1.0 429 The "m4" from BSDI won't handle the config files properly. 430 I haven't had a chance to test this myself. 431 432 The M4 shipped in FreeBSD and NetBSD 0.9 don't handle the config 433 files properly. One must use either GNU m4 1.1 or the PD-M4 434 recently posted in comp.os.386bsd.bugs (and maybe others). 435 NetBSD-current includes the PD-M4 (as stated in the NetBSD file 436 CHANGES). 437 438 FreeBSD 1.0 RELEASE has uname(2) now. Use -DUSEUNAME in order to 439 use it (look into Makefile.FreeBSD). NetBSD-current may have 440 it too but it has not been verified. 441 442 You cannot port the latest version of the Berkeley db library 443 and use it with sendmail without recompiling the world. This 444 is because C library routines use the older version which have 445 incompatible header files -- the result is that it can't read 446 other system files, such as /etc/passwd, unless you use the 447 new db format throughout your system. You should normally just 448 use the version of db supplied in your release. You may need 449 to use -DOLD_NEWDB to make this work -- this turns off some 450 new interface calls (for file locking) that are not in older 451 versions of db. You'll get compile errors if you need this 452 flag and don't have it set. 453 4544.3BSD 455 If you are running a "virgin" version of 4.3BSD, you'll have 456 a very old resolver and be missing some header files. The 457 header files are simple -- create empty versions and everything 458 will work fine. For the resolver you should really port a new 459 version (4.8.3 or later) of the resolver; 4.9 is available on 460 gatekeeper.DEC.COM in pub/BSD/bind/4.9. If you are really 461 determined to continue to use your old, buggy version (or as 462 a shortcut to get sendmail working -- I'm sure you have the 463 best intentions to port a modern version of BIND), you can 464 copy ../contrib/oldbind.compat.c into src and add 465 oldbind.compat.o to OBJADD in the Makefile. 466 467A/UX 468 Date: Tue, 12 Oct 1993 18:28:28 -0400 (EDT) 469 From: "Eric C. Hagberg" <hagberg@med.cornell.edu> 470 Subject: Fix for A/UX ndbm 471 472 I guess this isn't really a sendmail bug, however, it is something 473 that A/UX users should be aware of when compiling sendmail 8.6. 474 475 Apparently, the calls that sendmail is using to the ndbm routines 476 in A/UX 3.0.x contain calls to "broken" routines, in that the 477 aliases database will break when it gets "just a little big" 478 (sorry I don't have exact numbers here, but it broke somewhere 479 around 20-25 aliases for me.), making all aliases non-functional 480 after exceeding this point. 481 482 What I did was to get the gnu-dbm-1.6 package, compile it, and 483 then re-compile sendmail with "-lgdbm", "-DNDBM", and using the 484 ndbm.h header file that comes with the gnu-package. This makes 485 things behave properly. 486 487 I suppose porting the New Berkeley db package is another route, 488 however, I made a quick attempt at it, and found it difficult 489 (not easy at least); the gnu-dbm package "configured" and 490 compiled easily. 491 492DG/UX 493 Apparently, /bin/mail doesn't work properly for delivery on 494 DG/UX -- the person who has this working, Douglas Anderson 495 <dlander@afterlife.ncsc.mil>, used procmail instead. 496 497System V Release 4 Based Systems 498 There is a single Makefile that is intended for all SVR4-based 499 systems (called Makefile.SVR4). It defines __svr4__, which is 500 predefined by some compilers. If your compiler already defines 501 this compile variable, you can delete the definition from the 502 Makefile. 503 504 It's been tested on Dell Issue 2.2. 505 506DELL SVR4 507 Date: Mon, 06 Dec 1993 10:42:29 EST 508 From: "Kimmo Suominen" <kim@grendel.lut.fi> 509 Message-ID: <2d0352f9.lento29@lento29.UUCP> 510 To: eric@cs.berkeley.edu 511 Cc: sendmail@cs.berkeley.edu 512 Subject: Notes for DELL SVR4 513 514 Eric, 515 516 Here are some notes for compiling Sendmail 8.6.4 on DELL SVR4. I ran 517 across these things when helping out some people who contacted me by 518 e-mail. 519 520 1) Use gcc 2.4.5 (or later?). Dell distributes gcc 2.1 with their 521 Issue 2.2 Unix. It is too old, and gives you problems with 522 clock.c, because sigset_t won't get defined in <sys/signal.h>. 523 This is due to a problematic protection rule in there, and is 524 fixed with gcc 2.4.5. 525 526 2) If you don't use the new Berkeley DB (-DNEWDB), then you need 527 to add "-lc -lucb" to the libraries to link with. This is because 528 the -ldbm distributed by Dell needs the bcopy, bcmp and bzero 529 functions. It is important that you specify both libraries in 530 the given order to be sure you only get the BSTRING functions 531 from the UCB library (and not the signal routines etc.). 532 533 3) Don't leave out "-lelf" even if compiling with "-lc -lucb". 534 The UCB library also has another copy of the nlist routines, 535 but we do want the ones from "-lelf". 536 537 If anyone needs a compiled gcc 2.4.5 and/or a ported DB library, they 538 can use anonymous ftp to fetch them from lut.fi in the /kim directory. 539 They are copies of what I use on grendel.lut.fi, and offering them 540 does not imply that I would also support them. I have sent the DB 541 port for SVR4 back to Keith Bostic for inclusion in the official 542 distribution, but I haven't heard anything from him as of today. 543 544 - gcc-2.4.5-svr4.tar.gz (gcc 2.4.5 and the corresponding libg++) 545 - db-1.72.tar.gz (with source, objects and a installed copy) 546 547 Cheers 548 + Kim 549 -- 550 * Kimmo.Suominen@lut.fi * SysVr4 enthusiast at GRENDEL.LUT.FI * 551 * KIM@FINFILES.BITNET * Postmaster and Hostmaster at LUT.FI * 552 * + 358 200 865 718 * Unix area moderator at NIC.FUNET.FI * 553 554 555Non-DNS based sites 556 This version of sendmail always tries to connect to the Domain 557 Name System (DNS) to resolve names, regardless of the setting 558 of the `I' option. On most systems that are not running DNS, 559 this will fail quickly and sendmail will continue, but on some 560 systems it has a long timeout. If you have this problem, you 561 will have to recompile without NAMED_BIND. Some people have 562 claimed that they have successfully used "OI+USEVC" to force 563 sendmail to use a virtual circuit -- this will always time out 564 quickly, but also tells sendmail that a failed connection 565 should requeue the message (probably not what you intended). 566 A future release of sendmail will correct this problem. 567 568Both NEWDB and NDBM 569 If you use both -DNDBM and -DNEWDB, you must delete the module 570 ndbm.o from libdb.a and delete the file "ndbm.h" from the files 571 that get installed (that is, use the OLD ndbm.h, not the new 572 ndbm.h). This compatibility module maps ndbm calls into DB 573 calls, and breaks things rather badly. 574 575GNU getopt 576 I'm told that GNU getopt has a problem in that it gets confused 577 by the double call. Use the version in conf.c instead. 578 579 580+--------------+ 581| MANUAL PAGES | 582+--------------+ 583 584The manual pages have been written against the -mandoc macros 585instead of the -man macros. The latest version of groff has them 586included. You can also get a copy from FTP.UU.NET in directory 587/systems/unix/bsd-sources/share/tmac. 588 589 590+-----------------+ 591| DEBUGGING HOOKS | 592+-----------------+ 593 594As of 8.6.5, sendmail daemons will catch a SIGUSR1 signal and log 595some debugging output (logged at LOG_DEBUG severity). The 596information dumped is: 597 598 * The value of the $j macro. 599 * A warning if $j is not in the set $=w. 600 * A list of the open file descriptors. 601 * The contents of the connection cache. 602 * If ruleset 89 is defined, it is evaluated and the results printed. 603 604This allows you to get information regarding the runtime state of the 605daemon on the fly. This should not be done too frequently, since 606the process of rewriting may lose memory which will not be recovered. 607Also, ruleset 89 may call non-reentrant routines, so there is a small 608non-zero probability that this will cause other problems. It is 609really only for debugging serious problems. 610 611A typical formulation of ruleset 89 would be: 612 613 R$* $@ $>0 some test address 614 615 616+-----------------------------+ 617| DESCRIPTION OF SOURCE FILES | 618+-----------------------------+ 619 620The following list describes the files in this directory: 621 622Makefile The makefile used here; this version only works with 623 the new Berkeley make. 624Makefile.dist A trimmed down version of the makefile that works with 625 the old make. 626READ_ME This file. 627TRACEFLAGS My own personal list of the trace flags -- not guaranteed 628 to be particularly up to date. 629alias.c Does name aliasing in all forms. 630arpadate.c A subroutine which creates ARPANET standard dates. 631clock.c Routines to implement real-time oriented functions 632 in sendmail -- e.g., timeouts. 633collect.c The routine that actually reads the mail into a temp 634 file. It also does a certain amount of parsing of 635 the header, etc. 636conf.c The configuration file. This contains information 637 that is presumed to be quite static and non- 638 controversial, or code compiled in for efficiency 639 reasons. Most of the configuration is in sendmail.cf. 640conf.h Configuration that must be known everywhere. 641convtime.c A routine to sanely process times. 642daemon.c Routines to implement daemon mode. This version is 643 specifically for Berkeley 4.1 IPC. 644deliver.c Routines to deliver mail. 645domain.c Routines that interface with DNS (the Domain Name 646 System). 647err.c Routines to print error messages. 648envelope.c Routines to manipulate the envelope structure. 649headers.c Routines to process message headers. 650macro.c The macro expander. This is used internally to 651 insert information from the configuration file. 652main.c The main routine to sendmail. This file also 653 contains some miscellaneous routines. 654map.c Support for database maps. 655mci.c Routines that handle mail connection information caching. 656parseaddr.c The routines which do address parsing. 657queue.c Routines to implement message queueing. 658readcf.c The routine that reads the configuration file and 659 translates it to internal form. 660recipient.c Routines that manipulate the recipient list. 661savemail.c Routines which save the letter on processing errors. 662sendmail.h Main header file for sendmail. 663srvrsmtp.c Routines to implement server SMTP. 664stab.c Routines to manage the symbol table. 665stats.c Routines to collect and post the statistics. 666sysexits.c List of error messages associated with error codes 667 in sysexits.h. 668trace.c The trace package. These routines allow setting and 669 testing of trace flags with a high granularity. 670udb.c The user database interface module. 671usersmtp.c Routines to implement user SMTP. 672util.c Some general purpose routines used by sendmail. 673version.c The version number and information about this 674 version of sendmail. Theoretically, this gets 675 modified on every change. 676 677Eric Allman 678 679(Version 8.40, last update 12/24/93 06:29:20) 680