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 8.58 (Berkeley) 03/13/94 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 and uses syntax 17that is not recognized by older makes. It also has assumptions 18about the 4.4 file system layout built in. See below for details 19about other Makefiles. 20 21There is also a Makefile.dist which is much less clever, but works on 22the old traditional make. You can use this using: 23 24 make -f Makefile.dist 25 26************************************************** 27** Read below for more details of Makefiles. ** 28************************************************** 29 30There is also a shell script (makesendmail) that tries to be clever 31about using object subdirectories. It's pretty straightforward, and 32may help if you share a source tree among different architectures. 33 34************************************************************************** 35** IMPORTANT: DO NOT USE OPTIMIZATION (``-O'') IF YOU ARE RUNNING ** 36** GCC 2.4.x or 2.5.x. THERE IS A BUG IN THE GCC OPTIMIZER THAT ** 37** CAUSES SENDMAIL COMPILES TO FAIL MISERABLY. ** 38************************************************************************** 39 40Jim Wilson of Cygnus believes he has found the problem -- it will 41probably be fixed in GCC 2.5.6 -- but until this is verified, be 42very suspicious of gcc -O. 43 44************************************************************************** 45** IMPORTANT: Read the appropriate paragraphs in the section on ** 46** ``Operating System and Compile Quirks''. ** 47************************************************************************** 48 49 50+-----------+ 51| MAKEFILES | 52+-----------+ 53 54The "Makefile"s in these directories are from 4.4 BSD, and hence 55really only work properly if you are on a 4.4 system. In particular, 56they use new syntax that will not be recognized on old make programs, 57and some of them do things like ``.include ../../Makefile.inc'' to 58pick up some system defines. If you are getting sendmail separately, 59these files won't be included in the distribution, as they are 60outside of the sendmail tree. 61 62Instead, you should use one of the other Makefiles, such as 63Makefile.SunOS for a SunOS system, and so forth. These should 64work with the version of make that is appropriate for that 65system. 66 67There are a bunch of other Makefiles for other systems with names 68like Makefile.HPUX for an HP-UX system. They use the version of 69make that is native for that system. These are the Makefiles that 70I use, and they have "Berkeley quirks" in them. I can't guarantee 71that they will work unmodified in your environment. Many of them 72include -I/usr/sww/include/db and -L/usr/sww/lib -- this is Berkeley's 73location (the ``Software Warehouse'') for the new database libraries, 74described below. You don't have to remove these definitions if you 75don't have these directories. 76 77Please look for an appropriate Makefile before you start trying to 78compile with Makefile or Makefile.dist. 79 80If you want to port the new Berkeley make, you can get it from 81ftp.uu.net in the directory /systems/unix/bsd-sources/usr.bin/make. 82Diffs and instructions for building this version of make under 83SunOS 4.1.x are available on ftp.css.itd.umich.edu in 84/pub/systems/sun/Net2-make.sun4.diff.Z. Diffs and instructions 85for building this version of make under IBM AIX 3.2.4 are available 86on ftp.uni-stuttgart.de in /sw/src/patches/bsd-make-rus-patches. 87Paul Southworth <pauls@umich.edu> published a description of porting 88this make in comp.unix.bsd. 89 90The complete text of the Makefile.inc that is in the parent of the 91sendmail directory is: 92 93 # @(#)Makefile.inc 8.1 (Berkeley) 6/6/93 94 95 BINDIR?= /usr/sbin 96 97 98+----------------------+ 99| DATABASE DEFINITIONS | 100+----------------------+ 101 102There are several database formats that can be used for the alias files 103and for general maps. When used for alias files they interact in an 104attempt to be back compatible. 105 106The three options are NEWDB (the new Berkeley DB package), NDBM (the 107older DBM implementation -- the very old V7 implementation is no 108longer supported), and NIS (Network Information Services). Used alone 109these just include the support they indicate. [If you are using NEWDB, 110get the latest version from FTP.CS.Berkeley.EDU in /ucb/4bsd. DO NOT 111use the version from the Net2 distribution! However, if you are on 112BSD/386 or 386BSD-based systems, use the one that already exists 113on your system. You may need to define OLD_NEWDB to do this.] 114 115[NOTE WELL: it is CRITICAL that you remove ndbm.o from libdb.a and 116ndbm.h from the appropriate include directories if you want to get 117ndbm support. These files OVERRIDE calls to ndbm routines -- in 118particular, if you leave ndbm.h in, you can find yourself using 119the new db package even if you don't define NEWDB.] 120 121If NEWDB and NDBM are defined (but not NIS), then sendmail will read 122NDBM format alias files, but the next time a newaliases is run the 123format will be converted to NEWDB; that format will be used forever 124more. This is intended as a transition feature. [Note however that 125the NEWDB library also catches and maps NDBM calls; you will have to 126back out this feature to get this to work. See ``Quirks'' section 127below for details.] 128 129If all three are defined, sendmail operates as described above, and also 130looks for the file /var/yp/Makefile. If it exists, newaliases will 131build BOTH the NEWDB and NDBM format alias files. However, it will 132only use the NEWDB file; the NDBM format file is used only by the 133NIS subsystem. 134 135If NDBM and NIS are defined (regardless of the definition of NEWDB 136or the existance of /var/yp/Makefile), sendmail adds the special 137tokens "YP_LAST_MODIFIED" and "YP_MASTER_NAME", both of which are 138required if the NDBM file is to be used as an NIS map. 139 140All of -DNEWDB, -DNDBM, and -DNIS are normally defined in the DBMDEF 141line in the Makefile. 142 143 144+---------------+ 145| COMPILE FLAGS | 146+---------------+ 147 148Whereever possible, I try to make sendmail pull in the correct 149compilation options needed to compile on various environments based on 150automatically defined symbols. Some machines don't seem to have useful 151symbols availble, requiring the following compilation flags in the 152Makefile: 153 154SOLARIS Define this if you are running Solaris 2.0 or higher. 155SOLARIS_2_3 Define this if you are running Solaris 2.3 or higher. 156SUNOS403 Define this if you are running SunOS 4.0.3. 157NeXT Define this if you are on a NeXT box. (This one may 158 be pre-defined for you.) There are other hacks you 159 have to make -- see below. 160_AIX3 Define this if you are IBM AIX 3.x. 161RISCOS Define this if you are running RISC/os from MIPS. 162IRIX Define this if you are running IRIX from SGI. 163_SCO_unix_ Define this if you are on SCO UNIX. 164_SCO_unix_4_2 Define this if you are on SCO Open Server 3.2v4. 165 166If you are a system that sendmail has already been ported to, you 167probably won't have to touch these. But if you are porting, you may 168have to tweak the following compilation flags in conf.h in order to 169get it to compile and link properly: 170 171SYSTEM5 Adjust for System V (not necessarily Release 4). 172SYS5SIGNALS Use System V signal semantics -- the signal handler 173 is automatically dropped when the signal is caught. 174 If this is not set, use POSIX/BSD semantics, where the 175 signal handler stays in force until an exec or an 176 explicit delete. Implied by SYSTEM5. 177SYS5SETPGRP Use System V setpgrp() semantics. Implied by SYSTEM5. 178HASFLOCK Set this if you prefer to use the flock(2) system call 179 rather than using fcntl-based locking. Fcntl locking 180 has some semantic gotchas, but many vendor systems 181 also interface it to lockd(8) to do NFS-style locking. 182 For this reason, this should not be set unless you 183 don't have an alternative. 184HASUNAME Set if you have the "uname" system call. Implied by 185 SYSTEM5. 186HASUNSETENV Define this if your system library has the "unsetenv" 187 subroutine. 188HASSETSID Define this if you have the setsid(2) system call. This 189 is implied if your system appears to be POSIX compliant. 190HASINITGROUPS Define this if you have the initgroups(3) routine. 191HASSETVBUF Define this if you have the setvbuf(3) library call. 192 If you don't, setlinebuf will be used instead. This 193 defaults on if your compiler defines __STDC__. 194HASSETREUID Define this if you have setreuid(2) ***AND*** root can 195 use setreuid to change to an arbitrary user. This second 196 condition is not satisfied on AIX 3.x. You may find that 197 your system has setresuid(2), (for example, on HP-UX) in 198 which case you will also have to #define setreuid(r, e) 199 to be the appropriate call. Some systems (such as Solaris) 200 have a compatibility routine that doesn't work properly, 201 but may have "saved user ids" properly implemented so you 202 can ``#define setreuid(r, e) seteuid(e)'' and have it work. 203 The important thing is that you have a call that will set 204 the effective uid independently of the real or saved uid 205 and be able to set the effective uid back again when done. 206 There's a test program in ../test/t_setreuid.c that will 207 try things on your system. Setting this improves the 208 security, since sendmail doesn't have to read .forward 209 and :include: files as root. There are certain attacks 210 that may be unpreventable without this call. 211HASLSTAT Define this if you have symbolic links (and thus the 212 lstat(2) system call). This improves security. Unlike 213 most other options, this one is on by default, so you 214 need to #undef it in conf.h if you don't have symbolic 215 links (these days everyone does). 216NEEDGETOPT Define this if you need a reimplementation of getopt(3). 217 On some systems, getopt does very odd things if called 218 to scan the arguments twice. This flag will ask sendmail 219 to compile in a local version of getopt that works 220 properly. 221NEEDSTRTOL Define this if your standard C library does not define 222 strtol(3). This will compile in a local version. 223NEEDVPRINTF Define this if your standard C library does not define 224 vprintf(3). Note that the resulting fake implementation 225 is not very elegant and may not even work on some 226 architectures. 227HASGETUSERSHELL Define this to 1 if you have getusershell(3) in your 228 standard C library. If this is not defined, or is defined 229 to be 0, sendmail will scan the /etc/shells file (no 230 NIS-style support, defaults to /bin/sh and /bin/csh if 231 that file does not exist) to get a list of unrestricted 232 user shells. This is used to determine whether users 233 are allowed to forward their mail to a program or a file. 234GIDSET_T The type of entries in a gidset passed as the second 235 argument to getgroups(2). Historically this has been an 236 int, so this is the default, but some systems (such as 237 IRIX) pass it as a gid_t, which is an unsigned short. 238 This will make a difference, so it is important to get 239 this right! However, it is only an issue if you have 240 group sets. 241SLEEP_T The type returned by the system sleep() function. 242 Defaults to "unsigned int". Don't worry about this 243 if you don't have compilation problems. 244ARBPTR_T The type of an arbitrary pointer -- defaults to "void *". 245 If you are an very old compiler you may need to define 246 this to be "char *". 247LA_TYPE The type of load average your kernel supports. These 248 can be one of: 249 LA_ZERO (1) -- it always returns the load average as 250 "zero" (and does so on all architectures). 251 LA_SUBR (4) if you have the getloadavg(3) routine, 252 LA_MACH (5) to use MACH-style load averages (calls 253 processor_set_info()), 254 LA_PROCSTR (7) to read /proc/loadavg and interpret it 255 as a string representing a floating-point 256 number (Linux-style), 257 LA_FLOAT (3) if you read kmem and interpret the value 258 as a floating point number, 259 LA_INT (2) to interpret as a long integer, 260 LA_SHORT (6) to interpret as a short integer. 261 These last three have several other parameters that they 262 try to divine: the name of your kernel, the name of the 263 variable in the kernel to examine, the number of bits of 264 precision in a fixed point load average, and so forth. 265 In desperation, use LA_ZERO. The actual code is in 266 conf.c -- it can be tweaked if you are brave. 267SFS_TYPE Encodes how your kernel can locate the amount of free 268 space on a disk partition. This can be set to SFS_NONE 269 (0) if you have no way of getting this information, 270 SFS_USTAT (1) if you have the ustat(2) system call, 271 SFS_4ARGS (2) if you have a four-argument statfs(2) 272 system call (and the include file is <sys/statfs.h>), 273 and SFS_VFS (3), SFS_MOUNT (4), or SFS_STATFS (5) if 274 you have the two-argument statfs(2) system call, with 275 includes in <sys/vfs.h>, <sys/mount.h>, or <sys/statfs.h> 276 respectively. The default if nothing is defined is 277 SFS_NONE. 278ERRLIST_PREDEFINED 279 If set, assumes that some header file defines sys_errlist. 280 This may be needed if you get type conflicts on this 281 variable -- otherwise don't worry about it. 282WAITUNION The wait(2) routine takes a "union wait" argument instead 283 of an integer argument. This is for compatibility with 284 old versions of BSD. 285SCANF You can set this to extend the F command to accept a 286 scanf string -- this gives you a primitive parser for 287 class definitions -- BUT it can make you vulnerable to 288 core dumps if the target file is poorly formed. 289SYSLOG_BUFSIZE You can define this to be the size of the buffer that 290 syslog accepts. If it is not defined, it assumes a 291 1024-byte buffer. If the buffer is very small (under 292 256 bytes) the log message format changes -- each 293 e-mail message will log many more messages, since it 294 will log each piece of information as a separate line 295 in syslog. 296BROKEN_RES_SEARCH 297 On Ultrix (and maybe other systems?) if you use the 298 res_search routine with an unknown host name, it returns 299 -1 but sets h_errno to 0 instead of HOST_NOT_FOUND. If 300 you set this, sendmail considers 0 to be the same as 301 HOST_NOT_FOUND. 302 303 304+-----------------------+ 305| COMPILE-TIME FEATURES | 306+-----------------------+ 307 308There are a bunch of features that you can decide to compile in, such 309as selecting various database packages and special protocol support. 310Several are assumed based on other compilation flags -- if you want to 311"un-assume" something, you probably need to edit conf.h. Compilation 312flags that add support for special features include: 313 314NDBM Include support for "new" DBM library for aliases and maps. 315 Normally defined in the Makefile. 316NEWDB Include support for Berkeley "db" package (hash & btree) 317 for aliases and maps. Normally defined in the Makefile. 318NIS Define this to get NIS (YP) support for aliases and maps. 319 Normally defined in the Makefile. 320USERDB Include support for the User Information Database. Implied 321 by NEWDB in conf.h. 322IDENTPROTO Define this as 1 to get IDENT (RFC 1413) protocol support. 323 This is assumed unless you are running on Ultrix or 324 HP-UX, both of which have a problem in the UDP 325 implementation. You can define it to be 0 to explicitly 326 turn off IDENT protocol support. 327MIME Include support for MIME-encapsulated error messages. 328LOG Set this to get syslog(3) support. Defined by default 329 in conf.h. You want this if at all possible. 330NETINET Set this to get TCP/IP support. Defined by default 331 in conf.h. You probably want this. 332NETISO Define this to get ISO networking support. 333SMTP Define this to get the SMTP code. Implied by NETINET 334 or NETISO. 335NAMED_BIND Define this to get DNS (name daemon) support, including 336 MX support. The specs you must use this if you run 337 SMTP. Defined by default in conf.h. 338QUEUE Define this to get queueing code. Implied by NETINET 339 or NETISO; required by SMTP. This gives you other good 340 stuff -- it should be on. 341DAEMON Define this to get general network support. Implied by 342 NETINET or NETISO. Defined by default in conf.h. You 343 almost certainly want it on. 344MATCHGECOS Permit fuzzy matching of user names against the full 345 name (GECOS) field in the /etc/passwd file. This should 346 probably be on, since you can disable it from the config 347 file if you want to. Defined by default in conf.h. 348SETPROCTITLE Try to set the string printed by "ps" to something 349 informative about what sendmail is doing. Defined by 350 default in conf.h. 351 352 353+---------------------+ 354| DNS/RESOLVER ISSUES | 355+---------------------+ 356 357Many systems have old versions of the resolver library. At a minimum, 358you should be running BIND 4.8.3; older versions may compile, but they 359have known bugs that should give you pause. 360 361Common problems in old versions include "undefined" errors for 362dn_skipname. 363 364Some people have had a problem with BIND 4.9; it uses some routines 365that it expects to be externally defined such as strerror(). It may 366help to link with "-l44bsd" to solve this problem. 367 368!PLEASE! be sure to link with the same version of the resolver as 369the header files you used -- some people have used the 4.9 headers 370and linked with BIND 4.8 or vice versa, and it doesn't work. 371Unfortunately, it doesn't fail in an obvious way -- things just 372subtly don't work. 373 374 375+-------------------------------------+ 376| OPERATING SYSTEM AND COMPILE QUIRKS | 377+-------------------------------------+ 378 379GCC 2.5.x problems *** IMPORTANT *** 380 Date: Mon, 29 Nov 93 19:08:44 PST 381 From: wilson@cygnus.com (Jim Wilson) 382 Message-Id: <9311300308.AA04608@cygnus.com> 383 To: kenner@vlsi1.ultra.nyu.edu 384 Subject: [cattelan@thebarn.com: gcc 2.5.4-2.5.5 -O bug] 385 Cc: cattelan@thebarn.com, rms@gnu.ai.mit.edu, sendmail@cs.berkeley.edu 386 387 This fixes a problem that occurs when gcc 2.5.5 is used to compile 388 sendmail 8.6.4 with optimization on a sparc. 389 390 Mon Nov 29 19:00:14 1993 Jim Wilson (wilson@sphagnum.cygnus.com) 391 392 * reload.c (find_reloads_toplev): Replace obsolete reference to 393 BYTE_LOADS_*_EXTEND with LOAD_EXTEND_OP. 394 395 *** clean-ss-931128/reload.c Sun Nov 14 16:20:01 1993 396 --- ss-931128/reload.c Mon Nov 29 18:52:55 1993 397 *************** find_reloads_toplev (x, opnum, type, ind 398 *** 3888,3894 **** 399 force a reload in that case. So we should not do anything here. */ 400 401 else if (regno >= FIRST_PSEUDO_REGISTER 402 ! #if defined(BYTE_LOADS_ZERO_EXTEND) || defined(BYTE_LOADS_SIGN_EXTEND) 403 && (GET_MODE_SIZE (GET_MODE (x)) 404 <= GET_MODE_SIZE (GET_MODE (SUBREG_REG (x)))) 405 #endif 406 --- 3888,3894 ---- 407 force a reload in that case. So we should not do anything here. */ 408 409 else if (regno >= FIRST_PSEUDO_REGISTER 410 ! #ifdef LOAD_EXTEND_OP 411 && (GET_MODE_SIZE (GET_MODE (x)) 412 <= GET_MODE_SIZE (GET_MODE (SUBREG_REG (x)))) 413 #endif 414 415 416SunOS 4.x (Solaris 1.x) 417 You may have to use -lresolv on SunOS. However, beware that 418 this links in a new version of gethostbyname that does not 419 understand NIS, so you must have all of your hosts in DNS. 420 421 Some people have reported problems with the SunOS version of 422 -lresolv and/or in.named, and suggest that you get a newer 423 version. The symptoms are delays when you connect to the 424 SMTP server on a SunOS machine or having your domain added to 425 addresses inappropriately. There is a version of BIND 426 version 4.9 on gatekeeper.DEC.COM in pub/BSD/bind/4.9. 427 428 There is substantial disagreement about whether you can make 429 this work with resolv+, which allows you to specify a search-path 430 of services. Some people report that it works fine, others 431 claim it doesn't work at all (including causing sendmail to 432 drop core when it tries to do multiple resolv+ lookups for a 433 single job). I haven't tried resolv+, as we use DNS exclusively. 434 435 Should you want to try resolv+, it is on ftp.uu.net in 436 /networking/ip/dns. 437 438Solaris 2.x (SunOS 5.x) 439 To compile for Solaris, be sure you use -DSOLARIS. 440 441 To the best of my knowledge, Solaris does not have the 442 gethostbyname problem described above. However, it does 443 have another one: 444 445 From a correspondent: 446 447 For solaris 2.2, I have 448 449 hosts: files dns 450 451 in /etc/nsswitch.conf and /etc/hosts has to have the fully 452 qualified host name. I think "files" has to be before "dns" 453 in /etc/nsswitch.conf during bootup. 454 455 From another correspondent: 456 457 When running sendmail under Solaris, the gethostbyname() 458 hack in conf.c which should perform proper canonicalization 459 of host names could fail. Result: the host name is not 460 canonicalized despite the hack, and you'll have to define $j 461 and $m in sendmail.cf somewhere. 462 463 The reason could be that /etc/nsswitch.conf is improperly 464 configured (at least from sendmail's point of view). For 465 example, the line 466 467 hosts: files nisplus dns 468 469 will make gethostbyname() look in /etc/hosts first, then ask 470 nisplus, then dns. However, if /etc/hosts does not contain 471 the full canonicalized hostname, then no amount of 472 gethostbyname()s will work. 473 474 Solution (or rather, a workaround): Ask nisplus first, then 475 dns, then local files: 476 477 hosts: nisplus dns [NOTFOUND=return] files 478 479 The Solaris "syslog" function is apparently limited to something 480 about 90 characters because of a kernel limitation. If you have 481 source code, you can probably up this number. You can get patches 482 that fix this problem: the patch ids are: 483 484 Solaris 2.1 100834 485 Solaris 2.2 100999 486 Solaris 2.3 101318 487 488 Be sure you have the appropriate patch installed or you won't 489 see system logging. 490 491OSF/1 492 If you are compiling on OSF/1 (DEC Alpha), you must use 493 -L/usr/shlib (otherwise it core dumps on startup). You may also 494 need -mld to get the nlist() function, although some versions 495 apparently don't need this. 496 497 Also, the enclosed makefile removed /usr/sbin/smtpd; if you need 498 it, just create the link to the sendmail binary. 499 500IRIX 501 The header files on SGI IRIX are completely prototyped, and as 502 a result you can sometimes get some warning messages during 503 compilation. These can be ignored. There are two errors in 504 deliver only if you are using gcc, both of the form ``warning: 505 passing arg N of `execve' from incompatible pointer type''. 506 Also, if you compile with -DNIS, you will get a complaint 507 about a declaration of struct dom_binding in a prototype 508 when compiling map.c; this is not important because the 509 function being prototyped is not used in that file. 510 511NeXT 512 If you are compiling on NeXT, you will have to create an empty 513 file "unistd.h" and create a file "dirent.h" containing: 514 515 #include <sys/dir.h> 516 #define dirent direct 517 518 (The Makefile.NeXT should try to do both of these for you.) 519 520 Apparently, there is a bug in getservbyname on Nextstep 3.0 521 that causes it to fail under some circumstances with the 522 message "SYSERR: service "smtp" unknown" logged. You should 523 be able to work around this by including the line: 524 525 OOPort=25 526 527 in your .cf file. 528 529 You may have to use -DNeXT. 530 531BSDI (BSD/386) 1.0, NetBSD 0.9, FreeBSD 1.0 532 The "m4" from BSDI won't handle the config files properly. 533 I haven't had a chance to test this myself. 534 535 The M4 shipped in FreeBSD and NetBSD 0.9 don't handle the config 536 files properly. One must use either GNU m4 1.1 or the PD-M4 537 recently posted in comp.os.386bsd.bugs (and maybe others). 538 NetBSD-current includes the PD-M4 (as stated in the NetBSD file 539 CHANGES). 540 541 FreeBSD 1.0 RELEASE has uname(2) now. Use -DUSEUNAME in order to 542 use it (look into Makefile.FreeBSD). NetBSD-current may have 543 it too but it has not been verified. 544 545 You cannot port the latest version of the Berkeley db library 546 and use it with sendmail without recompiling the world. This 547 is because C library routines use the older version which have 548 incompatible header files -- the result is that it can't read 549 other system files, such as /etc/passwd, unless you use the 550 new db format throughout your system. You should normally just 551 use the version of db supplied in your release. You may need 552 to use -DOLD_NEWDB to make this work -- this turns off some 553 new interface calls (for file locking) that are not in older 554 versions of db. You'll get compile errors if you need this 555 flag and don't have it set. 556 5574.3BSD 558 If you are running a "virgin" version of 4.3BSD, you'll have 559 a very old resolver and be missing some header files. The 560 header files are simple -- create empty versions and everything 561 will work fine. For the resolver you should really port a new 562 version (4.8.3 or later) of the resolver; 4.9 is available on 563 gatekeeper.DEC.COM in pub/BSD/bind/4.9. If you are really 564 determined to continue to use your old, buggy version (or as 565 a shortcut to get sendmail working -- I'm sure you have the 566 best intentions to port a modern version of BIND), you can 567 copy ../contrib/oldbind.compat.c into src and add 568 oldbind.compat.o to OBJADD in the Makefile. 569 570A/UX 571 Date: Tue, 12 Oct 1993 18:28:28 -0400 (EDT) 572 From: "Eric C. Hagberg" <hagberg@med.cornell.edu> 573 Subject: Fix for A/UX ndbm 574 575 I guess this isn't really a sendmail bug, however, it is something 576 that A/UX users should be aware of when compiling sendmail 8.6. 577 578 Apparently, the calls that sendmail is using to the ndbm routines 579 in A/UX 3.0.x contain calls to "broken" routines, in that the 580 aliases database will break when it gets "just a little big" 581 (sorry I don't have exact numbers here, but it broke somewhere 582 around 20-25 aliases for me.), making all aliases non-functional 583 after exceeding this point. 584 585 What I did was to get the gnu-dbm-1.6 package, compile it, and 586 then re-compile sendmail with "-lgdbm", "-DNDBM", and using the 587 ndbm.h header file that comes with the gnu-package. This makes 588 things behave properly. 589 590 I suppose porting the New Berkeley db package is another route, 591 however, I made a quick attempt at it, and found it difficult 592 (not easy at least); the gnu-dbm package "configured" and 593 compiled easily. 594 595DG/UX 596 Apparently, /bin/mail doesn't work properly for delivery on 597 DG/UX -- the person who has this working, Douglas Anderson 598 <dlander@afterlife.ncsc.mil>, used procmail instead. 599 600Apollo DomainOS 601 If you are compiling on Apollo, you will have to create an empty 602 file "unistd.h" and create a file "dirent.h" containing: 603 604 #include <sys/dir.h> 605 #define dirent direct 606 607 (The Makefile.DomainOS will attempt to do both of these for you.) 608 609HP-UX 8.00 610 Date: Mon, 24 Jan 1994 13:25:45 +0200 611 From: Kimmo Suominen <Kimmo.Suominen@lut.fi> 612 Subject: 8.6.5 w/ HP-UX 8.00 on s300 613 614 Just compiled and fought with sendmail 8.6.5 on a HP9000/360 (ie. a 615 series 300 machine) running HP-UX 8.00. 616 617 I was getting segmentation fault when delivering to a local user. 618 With debugging I saw it was faulting when doing _free@libc... *sigh* 619 It seems the new implementation of malloc on s300 is buggy as of 8.0, 620 so I tried out the one in -lmalloc (malloc(3X)). With that it seems 621 to work just dandy. 622 623 When linking, you will get the following error: 624 625 ld: multiply defined symbol _freespace in file /usr/lib/libmalloc.a 626 627 but you can just ignore it. You might want to add this info to the 628 README file for the future... 629 630Linux 631 Something broke between versions 0.99.13 and 0.99.14 of Linux: 632 the flock() system call gives errors. If you are running .14, 633 you must not use flock. You can do this with -DHASFLOCK=0. 634 635AIX 636 This version of sendmail does not support MB, MG, and MR resource 637 records, which are supported by AIX sendmail. 638 639RISC/os 640 RISC/os from MIPS is a merged AT&T/Berkeley system. When you 641 compile on that platform you will get duplicate definitions 642 on many files. You can ignore these. 643 644System V Release 4 Based Systems 645 There is a single Makefile that is intended for all SVR4-based 646 systems (called Makefile.SVR4). It defines __svr4__, which is 647 predefined by some compilers. If your compiler already defines 648 this compile variable, you can delete the definition from the 649 Makefile. 650 651 It's been tested on Dell Issue 2.2. 652 653DELL SVR4 654 Date: Mon, 06 Dec 1993 10:42:29 EST 655 From: "Kimmo Suominen" <kim@grendel.lut.fi> 656 Message-ID: <2d0352f9.lento29@lento29.UUCP> 657 To: eric@cs.berkeley.edu 658 Cc: sendmail@cs.berkeley.edu 659 Subject: Notes for DELL SVR4 660 661 Eric, 662 663 Here are some notes for compiling Sendmail 8.6.4 on DELL SVR4. I ran 664 across these things when helping out some people who contacted me by 665 e-mail. 666 667 1) Use gcc 2.4.5 (or later?). Dell distributes gcc 2.1 with their 668 Issue 2.2 Unix. It is too old, and gives you problems with 669 clock.c, because sigset_t won't get defined in <sys/signal.h>. 670 This is due to a problematic protection rule in there, and is 671 fixed with gcc 2.4.5. 672 673 2) If you don't use the new Berkeley DB (-DNEWDB), then you need 674 to add "-lc -lucb" to the libraries to link with. This is because 675 the -ldbm distributed by Dell needs the bcopy, bcmp and bzero 676 functions. It is important that you specify both libraries in 677 the given order to be sure you only get the BSTRING functions 678 from the UCB library (and not the signal routines etc.). 679 680 3) Don't leave out "-lelf" even if compiling with "-lc -lucb". 681 The UCB library also has another copy of the nlist routines, 682 but we do want the ones from "-lelf". 683 684 If anyone needs a compiled gcc 2.4.5 and/or a ported DB library, they 685 can use anonymous ftp to fetch them from lut.fi in the /kim directory. 686 They are copies of what I use on grendel.lut.fi, and offering them 687 does not imply that I would also support them. I have sent the DB 688 port for SVR4 back to Keith Bostic for inclusion in the official 689 distribution, but I haven't heard anything from him as of today. 690 691 - gcc-2.4.5-svr4.tar.gz (gcc 2.4.5 and the corresponding libg++) 692 - db-1.72.tar.gz (with source, objects and a installed copy) 693 694 Cheers 695 + Kim 696 -- 697 * Kimmo.Suominen@lut.fi * SysVr4 enthusiast at GRENDEL.LUT.FI * 698 * KIM@FINFILES.BITNET * Postmaster and Hostmaster at LUT.FI * 699 * + 358 200 865 718 * Unix area moderator at NIC.FUNET.FI * 700 701 702Non-DNS based sites 703 This version of sendmail always tries to connect to the Domain 704 Name System (DNS) to resolve names, regardless of the setting 705 of the `I' option. On most systems that are not running DNS, 706 this will fail quickly and sendmail will continue, but on some 707 systems it has a long timeout. If you have this problem, you 708 will have to recompile without NAMED_BIND. Some people have 709 claimed that they have successfully used "OI+USEVC" to force 710 sendmail to use a virtual circuit -- this will always time out 711 quickly, but also tells sendmail that a failed connection 712 should requeue the message (probably not what you intended). 713 A future release of sendmail will correct this problem. 714 715Both NEWDB and NDBM 716 If you use both -DNDBM and -DNEWDB, you must delete the module 717 ndbm.o from libdb.a and delete the file "ndbm.h" from the files 718 that get installed (that is, use the OLD ndbm.h, not the new 719 ndbm.h). This compatibility module maps ndbm calls into DB 720 calls, and breaks things rather badly. 721 722GNU getopt 723 I'm told that GNU getopt has a problem in that it gets confused 724 by the double call. Use the version in conf.c instead. 725 726BIND 4.9.2 and Ultrix 727 If you are running on Ultrix, be sure you read the conf/Info.Ultrix 728 carefully -- there is information in there that you need to know 729 in order to avoid errors of the form: 730 731 /lib/libc.a(gethostent.o): sethostent: multiply defined 732 /lib/libc.a(gethostent.o): endhostent: multiply defined 733 /lib/libc.a(gethostent.o): gethostbyname: multiply defined 734 /lib/libc.a(gethostent.o): gethostbyaddr: multiply defined 735 736 during the link stage. 737 738 739+--------------+ 740| MANUAL PAGES | 741+--------------+ 742 743The manual pages have been written against the -mandoc macros 744instead of the -man macros. The latest version of groff has them 745included. You can also get a copy from FTP.UU.NET in directory 746/systems/unix/bsd-sources/share/tmac. 747 748 749+-----------------+ 750| DEBUGGING HOOKS | 751+-----------------+ 752 753As of 8.6.5, sendmail daemons will catch a SIGUSR1 signal and log 754some debugging output (logged at LOG_DEBUG severity). The 755information dumped is: 756 757 * The value of the $j macro. 758 * A warning if $j is not in the set $=w. 759 * A list of the open file descriptors. 760 * The contents of the connection cache. 761 * If ruleset 89 is defined, it is evaluated and the results printed. 762 763This allows you to get information regarding the runtime state of the 764daemon on the fly. This should not be done too frequently, since 765the process of rewriting may lose memory which will not be recovered. 766Also, ruleset 89 may call non-reentrant routines, so there is a small 767non-zero probability that this will cause other problems. It is 768really only for debugging serious problems. 769 770A typical formulation of ruleset 89 would be: 771 772 R$* $@ $>0 some test address 773 774 775+-----------------------------+ 776| DESCRIPTION OF SOURCE FILES | 777+-----------------------------+ 778 779The following list describes the files in this directory: 780 781Makefile The makefile used here; this version only works with 782 the new Berkeley make. 783Makefile.dist A trimmed down version of the makefile that works with 784 the old make. 785READ_ME This file. 786TRACEFLAGS My own personal list of the trace flags -- not guaranteed 787 to be particularly up to date. 788alias.c Does name aliasing in all forms. 789arpadate.c A subroutine which creates ARPANET standard dates. 790clock.c Routines to implement real-time oriented functions 791 in sendmail -- e.g., timeouts. 792collect.c The routine that actually reads the mail into a temp 793 file. It also does a certain amount of parsing of 794 the header, etc. 795conf.c The configuration file. This contains information 796 that is presumed to be quite static and non- 797 controversial, or code compiled in for efficiency 798 reasons. Most of the configuration is in sendmail.cf. 799conf.h Configuration that must be known everywhere. 800convtime.c A routine to sanely process times. 801daemon.c Routines to implement daemon mode. This version is 802 specifically for Berkeley 4.1 IPC. 803deliver.c Routines to deliver mail. 804domain.c Routines that interface with DNS (the Domain Name 805 System). 806err.c Routines to print error messages. 807envelope.c Routines to manipulate the envelope structure. 808headers.c Routines to process message headers. 809macro.c The macro expander. This is used internally to 810 insert information from the configuration file. 811main.c The main routine to sendmail. This file also 812 contains some miscellaneous routines. 813map.c Support for database maps. 814mci.c Routines that handle mail connection information caching. 815parseaddr.c The routines which do address parsing. 816queue.c Routines to implement message queueing. 817readcf.c The routine that reads the configuration file and 818 translates it to internal form. 819recipient.c Routines that manipulate the recipient list. 820savemail.c Routines which save the letter on processing errors. 821sendmail.h Main header file for sendmail. 822srvrsmtp.c Routines to implement server SMTP. 823stab.c Routines to manage the symbol table. 824stats.c Routines to collect and post the statistics. 825sysexits.c List of error messages associated with error codes 826 in sysexits.h. 827trace.c The trace package. These routines allow setting and 828 testing of trace flags with a high granularity. 829udb.c The user database interface module. 830usersmtp.c Routines to implement user SMTP. 831util.c Some general purpose routines used by sendmail. 832version.c The version number and information about this 833 version of sendmail. Theoretically, this gets 834 modified on every change. 835 836Eric Allman 837 838(Version 8.58, last update 03/13/94 09:38:23) 839