changes.ps

早期freebsd实现

(hort history of)173.36 352.2 R F2(sendmail)3.08 E F1 .58(through v)3.08 F .58
(ersion 5 and the moti)-.15 F -.25(va)-.25 G .58(tion behind w).25 F .58
(orking on)-.1 F -.15(ve)72 364.2 S .126(rsion 8.).15 F .126
(Section 2 has a rather detailed description of what has changed between v)
5.126 F .125(ersion 5 and v)-.15 F .125(ersion 8.)-.15 F
(The paper \214nishes of)72 376.2 Q 2.5(fw)-.25 G
(ith some thoughts about what still needs to be done.)168.95 376.2 Q/F3 10
/Times-Bold@0 SF 2.5(1. HIST)72 400.2 R(OR)-.18 E(Y)-.35 E F1 .151
(As discussed else)112 416.4 R .151
(where, [Allman83a, Allman83b, Allman&Amos85] sendmail has e)-.25 F .151
(xisted in v)-.15 F(ar)-.25 E(-)-.2 E .405(ious forms since 1980.)87 428.4 R
.405(It w)5.405 F .405(as released under the name)-.1 F F2(delivermail)2.905 E
F1 .404(in 4BSD and 4.1BSD, and as)2.905 F F2(send-)2.904 E(mail)87 440.4 Q F1
(in 4.2BSD.)2.5 E(It quickly became the dominant mail system for netw)5 E(ork)
-.1 E(ed UNIX systems.)-.1 E 1.569(Prior the release of 4.3BSD in No)112 456.6
R -.15(ve)-.15 G 1.569(mber 1986, the author had left the Uni).15 F -.15(ve)
-.25 G 1.57(rsity for pri).15 F -.25(va)-.25 G(te).25 E(industry)87 468.6 Q
3.347(,b)-.65 G .847(ut continued to do some w)129.777 468.6 R .847(ork on)-.1
F F2(sendmail)3.347 E F1 .847(with acti)3.347 F .846(vity slo)-.25 F .846
(wly trailing of)-.25 F 3.346(fu)-.25 G .846(ntil ef)445.204 468.6 R(fecti)-.25
E -.15(ve)-.25 G(ly).15 E .255(stopping after February 1987.)87 480.6 R .255
(There w)5.255 F .255(as minimal support done by man)-.1 F 2.756(yp)-.15 G .256
(eople for se)389.796 480.6 R -.15(ve)-.25 G .256(ral years, until).15 F
(July of 1991 when the original author)87 492.6 Q 2.5(,w)-.4 G
(ho had returned the Uni)249.36 492.6 Q -.15(ve)-.25 G(rsity).15 E 2.5(,s)-.65
G(tarted acti)379.4 492.6 Q .3 -.15(ve w)-.25 H(ork on it ag).05 E(ain.)-.05 E
1.271(There were se)112 508.8 R -.15(ve)-.25 G 1.271(ral reasons for rene).15 F
1.271(wed w)-.25 F 1.271(ork on)-.1 F F2(sendmail)3.771 E F1 6.271(.T)C 1.271
(here w)369.549 508.8 R 1.27(as a desire at Berk)-.1 F(ele)-.1 E 3.77(yt)-.15 G
(o)499 508.8 Q(con)87 520.8 Q -.15(ve)-.4 G .097
(rt to a subdomained structure so that indi).15 F .098
(viduals were identi\214ed by their subdomain rather than by)-.25 F 1.758
(their indi)87 532.8 R 1.758(vidual w)-.25 F 1.758(orkstation; although possib\
le in the old code, there were some problems, and the)-.1 F .66(author w)87
544.8 R .66(as the ob)-.1 F .66(vious person to address them.)-.15 F .66
(The Computer Systems Research Group \(CSRG\), the)5.66 F 1.89
(group that produced the Berk)87 556.8 R(ele)-.1 E 4.39(yS)-.15 G(oftw)238.12
556.8 Q 1.89(are Distrib)-.1 F 1.89(utions, w)-.2 F 1.89(as w)-.1 F 1.89
(orking on 4.4BSD, and w)-.1 F 1.89(anted an)-.1 F .053
(update to the mail system.)87 568.8 R .053(Bryan Costales w)5.053 F .053(as w)
-.1 F .053(orking on a book on)-.1 F F2(sendmail)2.553 E F1 .053(that w)2.553 F
.053(as being re)-.1 F(vie)-.25 E(wed)-.25 E .923(by the author)87 580.8 R
3.423(,w)-.4 G .923(hich encouraged him to mak)154.359 580.8 R 3.422(es)-.1 G
.922(ome re)283.572 580.8 R 3.422(visions. And)-.25 F .922(the author w)3.422 F
.922(anted to try to unify)-.1 F(some of the disparate v)87 592.8 Q(ersions of)
-.15 E F2(sendmail)2.5 E F1(that had been permitted to proliferate.)2.5 E .023
(During the 1987\25591 f)112 609 R(allo)-.1 E 2.523(wp)-.25 G .023(eriod, man)
228.482 609 R 2.523(yv)-.15 G .023(endors and outside v)283.498 609 R .023
(olunteers had produced v)-.2 F .024(ariants of)-.25 F F2(sendmail)87 621 Q F1
5.518(.P)C .517(erhaps the best kno)136.688 621 R .517(wn is the ID)-.25 F
3.017(Av)-.4 G .517(ersion [ID)280.317 621 R 3.017(A87]. Originally)-.4 F .517
(intended to be a ne)3.017 F 3.017(ws)-.25 G .517(et of)485.433 621 R .268
(con\214guration \214les, ID)87 633 R 2.768(Ae)-.4 G .269(xpanded into a f)
189.464 633 R .269(airly lar)-.1 F .269(ge set of patches for the code.)-.18 F
.269(Originally produced in)5.269 F .471(Sweden, ID)87 645 R 2.971(Ad)-.4 G
-2.15 -.25(ev e)149.472 645 T .471(lopment passed to the Uni).25 F -.15(ve)-.25
G .471(rsity of Illinois, and w).15 F .47(as widely used by the f)-.1 F .47
(airly lar)-.1 F(ge)-.18 E .077
(set of people who prefer to get and compile their o)87 657 R .077
(wn source code rather than use v)-.25 F(endor)-.15 E .078(-supplied bina-)-.2
F(ries.)87 669 Q .32 LW 76 678.6 72 678.6 DL 80 678.6 76 678.6 DL 84 678.6 80
678.6 DL 88 678.6 84 678.6 DL 92 678.6 88 678.6 DL 96 678.6 92 678.6 DL 100
678.6 96 678.6 DL 104 678.6 100 678.6 DL 108 678.6 104 678.6 DL 112 678.6 108
678.6 DL 116 678.6 112 678.6 DL 120 678.6 116 678.6 DL 124 678.6 120 678.6 DL
128 678.6 124 678.6 DL 132 678.6 128 678.6 DL 136 678.6 132 678.6 DL 140 678.6
136 678.6 DL 144 678.6 140 678.6 DL 148 678.6 144 678.6 DL 152 678.6 148 678.6
DL 156 678.6 152 678.6 DL 160 678.6 156 678.6 DL 164 678.6 160 678.6 DL 168
678.6 164 678.6 DL 172 678.6 168 678.6 DL 176 678.6 172 678.6 DL 180 678.6 176
678.6 DL 184 678.6 180 678.6 DL 188 678.6 184 678.6 DL 192 678.6 188 678.6 DL
196 678.6 192 678.6 DL 200 678.6 196 678.6 DL 204 678.6 200 678.6 DL 208 678.6
204 678.6 DL 212 678.6 208 678.6 DL 216 678.6 212 678.6 DL/F4 8/Times-Roman@0
SF .045(*An earlier v)93.6 690.6 R .045(ersion of this paper w)-.12 F .044
(as printed in the Proceedings of the 1994 A)-.08 F .044
(UUG Queensland Summer T)-.44 F .044(echnical Conference,)-.56 F(Gate)72 700.2
Q -.08(wa)-.2 G 2(yH).08 G(otel, Brisbane, March 1994.)107.928 700.2 Q F3
(Changes in Sendmail V)72 756 Q(ersion 8)-1 E(1)499 756 Q EP
%%Page: 2 2
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Bold@0 SF 294.65(2C)72 60 S(hanges in Sendmail V)378.87 60 Q
(ersion 8)-1 E/F1 10/Times-Roman@0 SF .151
(In about the same time frame, attempts were made to clean up and e)112 96 R
.151(xtend the Simple Mail T)-.15 F(rans-)-.35 E .468
(port Protocol \(SMTP\) [RFC821].)87 108 R .468(This in)5.468 F -.2(vo)-.4 G
(lv).2 E .469(ed clari\214cations of some ambiguities in the protocol, and)-.15
F .085(correction of some problem areas [RFC1123], as well as e)87 120 R .084
(xtensions for additional functionality \(dubbed)-.15 F 1.052
(Extended Simple Mail T)87 132 R 1.053
(ransport Protocol, or ESMTP\) [RFC1425, RFC1426, RFC1427] and a richer)-.35 F
1.376(set of semantics in the body of messages \(the Multipurpose Internet Mai\
l Extensions, a.k.a. MIME\))87 144 R .497([RFC1521, RFC1344].)87 156 R .497
(Neither the ID)5.497 F 2.998(Ag)-.4 G .498(roup nor most v)258.526 156 R .498
(endors were modifying)-.15 F/F2 10/Times-Italic@0 SF(sendmail)2.998 E F1 .498
(to conform)2.998 F 1.7(to these ne)87 168 R 4.2(ws)-.25 G 4.2(tandards. It)
148.23 168 R 1.699(seemed clear that these were `)4.2 F 1.699(`good things')
-.74 F 4.199('t)-.74 G 1.699(hat should be encouraged.)394.483 168 R(Ho)87 180
Q(we)-.25 E -.15(ve)-.25 G 1.635 -.4(r, s).15 H .835(ince no one w).4 F .835
(as w)-.1 F .835(orking on a publically a)-.1 F -.25(va)-.2 G .835(ilable v).25
F .836(ersion of)-.15 F F2(sendmail)3.336 E F1 .836(with these updates,)3.336 F
(the)87 192 Q 2.5(yw)-.15 G(ere unlik)113.79 192 Q(ely to be widely deplo)-.1 E
(yed an)-.1 E 2.5(yt)-.15 G(ime in the near future.)274.25 192 Q .466
(There are, of course, other mail transport agents a)112 208.2 R -.25(va)-.2 G
.465(ilable, such as).25 F F2 .465(MMDF zmailer smail)2.965 F F1(and)2.965 E F2
(PP)2.965 E F1(Ho)87 220.2 Q(we)-.25 E -.15(ve)-.25 G .842 -.4(r, n).15 H .042
(one of these seemed to be g).4 F .043(aining the prominence of)-.05 F F2
(sendmail)2.543 E F1 2.543(;i)C 2.543(ta)390.518 220.2 S .043
(ppeared that most compa-)400.281 220.2 R .238(nies w)87 232.2 R .238
(ould not con)-.1 F -.15(ve)-.4 G .238(rt to another mail transport agent an)
.15 F 2.737(yt)-.15 G .237(ime in the forseeable future.)327.438 232.2 R(Ho)
5.237 E(we)-.25 E -.15(ve)-.25 G 1.037 -.4(r, t).15 H(he).4 E(y)-.15 E
(might be persuaded to con)87 244.2 Q -.15(ve)-.4 G(rt to a ne).15 E(wer v)-.25
E(ersion of)-.15 E F2(sendmail)2.5 E F1(.)A .841(All of these con)112 260.4 R
.841(vinced the author to w)-.4 F .841(ork on a updated v)-.1 F .841(ersion of)
-.15 F F2(sendmail)3.342 E F1 .842(for public distrib)3.342 F(u-)-.2 E(tion.)87
272.4 Q 1.024(The ne)112 288.6 R 3.524(wv)-.25 G 1.023(ersion of)155.858 288.6
R F2(sendmail)3.523 E F1 1.023(is referred to as v)3.523 F 1.023
(ersion eight \(V8\).)-.15 F -1.11(Ve)6.023 G 1.023(rsions six and se)1.11 F
-.15(ve)-.25 G 3.523(nw).15 G(ere)491.79 288.6 Q 1.281
(skipped because of an agreement that all \214les in 4.4BSD w)87 300.6 R 1.281
(ould be numbered as \2318.1\232.)-.1 F 1.282(Rather than)6.282 F(ha)87 312.6 Q
2.05 -.15(ve a)-.2 H 4.25(ne).15 G 1.75(xternal v)127.76 312.6 R 1.75
(ersion number that dif)-.15 F 1.75(fered from the \214le v)-.25 F 1.75
(ersion numbers,)-.15 F F2(sendmail)4.25 E F1 1.75(just jumped)4.25 F
(directly to V8.)87 324.6 Q F0 2.5(2. CHANGES)72 348.6 R(IN VERSION EIGHT)2.5 E
F1 .138(The follo)112 364.8 R .139
(wing is a summary of the changes between the last commonly a)-.25 F -.25(va)
-.2 G .139(ilable v).25 F .139(ersion of send-)-.15 F(mail from Berk)87 376.8 Q
(ele)-.1 E 2.5(y\()-.15 G(5.67\) and the latest v)170.9 376.8 Q
(ersion \(8.6.6\).)-.15 E(Man)112 393 Q 2.5(yo)-.15 G 2.5(ft)142.68 393 S
(hese are ideas that had been tried in ID)151.29 393 Q(A, b)-.4 E(ut man)-.2 E
2.5(yo)-.15 G 2.5(ft)363.27 393 S(hem were generalized in V8.)371.88 393 Q F0
2.5(2.1. P)87 417 R(erf)-.2 E(ormance Enhancements)-.25 E F1 .549
(Instead of closing SMTP connections immediately)127 433.2 R 3.049(,o)-.65 G
.549(pen connections are cached for possible)342.135 433.2 R .029(future use.)
102 445.2 R .029(There is a limit to the number of simultaneous open connectio\
ns and the idle time of an)5.029 F(y)-.15 E(indi)102 457.2 Q
(vidual connection.)-.25 E 1.219(This is of best help during queue processing \
\(since there is the potential of man)127 473.4 R 3.719(yd)-.15 G(if)474.82
473.4 Q(ferent)-.25 E 1.113(messages going to one site\), although it can also\
 help when processing MX records which aren')102 485.4 R(t)-.18 E
(handled by MX Piggybacking.)102 497.4 Q 1.258(If tw)127 513.6 R 3.757(oh)-.1 G
1.257(osts with dif)161.075 513.6 R 1.257
(ferent names in a single message happen to ha)-.25 F 1.557 -.15(ve t)-.2 H
1.257(he same set of MX).15 F .94(hosts, the)102 525.6 R 3.44(yc)-.15 G .94
(an be sent in the same transaction.)153.45 525.6 R -1.11(Ve)5.94 G .94
(rsion 8 notices this and tries to batch the mes-)1.11 F(sages.)102 537.6 Q
-.15(Fo)127 553.8 S 3.638(re).15 G 1.138(xample, if tw)148.668 553.8 R 3.637
(os)-.1 G 1.137(ites `)216.42 553.8 R(`foo.com')-.74 E 3.637('a)-.74 G 1.137
(nd `)286.914 553.8 R(`bar)-.74 E(.com')-.55 E 3.637('a)-.74 G 1.137
(re both serv)352.408 553.8 R 1.137(ed by UUNET)-.15 F 3.637(,t)-.74 G(he)
470.513 553.8 Q 3.637(yw)-.15 G(ill)495.66 553.8 Q(ha)102 565.8 Q .557 -.15
(ve t)-.2 H .257(he same set of MX hosts and will be sent in one transaction.)
.15 F .258(UUNET will then split the mes-)5.258 F(sage and send it to the tw)
102 577.8 Q 2.5(oi)-.1 G(ndi)213.28 577.8 Q(vidual hosts.)-.25 E F0 2.5
(2.2. RFC)87 601.8 R(1123 Changes)2.5 E F1 2.607(An)127 618 S .107
(umber of changes ha)141.827 618 R .407 -.15(ve b)-.2 H .106(een made to mak)
.15 F 2.606(es)-.1 G .106(endmail `)321.07 618 R .106
(`conditionally compliant')-.74 F 2.606('\()-.74 G .106(that is, it)469.058 618
R(satis\214es all of the MUST clauses and most b)102 630 Q
(ut not all of the SHOULD clauses in RFC 1123\).)-.2 E
(The major areas of change are \(numbers are RFC 1123 section numbers\):)127
646.2 Q 26.5(\2475.2.7 Response)102 662.4 R .565(to RCPT command is f)3.065 F
3.065(ast. Pre)-.1 F(viously)-.25 E 3.065(,s)-.65 G .565(endmail e)362.295
662.4 R .565(xpanded all aliases as f)-.15 F(ar)-.1 E .686
(as it could \212 this could tak)156 674.4 R 3.186(eav)-.1 G .685
(ery long time, particularly if there were name serv)290.118 674.4 R(er)-.15 E
3.891(delays. V)156 686.4 R 1.391(ersion 8 only checks for the e)-1.11 F 1.392
(xistence of an alias and does the e)-.15 F(xpansion)-.15 E(later)156 698.4 Q
5.176(.I)-.55 G 2.676(td)184.226 698.4 S .176
(oes still do a DNS lookup if there is an e)194.682 698.4 R .175
(xplicit host name in the RCPT com-)-.15 F(mand, b)156 710.4 Q
(ut this time is bounded.)-.2 E EP
%%Page: 3 3
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Bold@0 SF(Changes in Sendmail V)72 60 Q(ersion 8)-1 E(3)499 60 Q
/F1 10/Times-Roman@0 SF 26.5(\2475.2.8 Numeric)102 96 R .612
(IP addresses are logged in Recei)3.112 F -.15(ve)-.25 G .613(d: lines.).15 F
.613(This helps tracing spoofed mes-)5.613 F(sages.)156 108 Q 21.5
(\2475.2.17 Self)102 124.2 R .127(domain literal is properly handled.)2.627 F
(Pre)5.126 E(viously)-.25 E 2.626(,i)-.65 G 2.626(fs)368.196 124.2 S .126
(omeone sent to user@[1.2.3.4],)378.042 124.2 R .12
(where 1.2.3.4 is your IP address, the mail w)156 136.2 R .12
(ould probably be rejected with a `)-.1 F(`con\214gu-)-.74 E(ration error')156
148.2 Q 2.5('. V)-.74 F(ersion 8 can handle these addresses.)-1.11 E 26.5
(\2475.3.2 Better)102 164.4 R 1.189(control o)3.69 F -.15(ve)-.15 G 3.689(ri)
.15 G(ndi)240.088 164.4 Q 1.189(vidual timeouts.)-.25 F 1.189
(RFC 821 speci\214ed no timeouts.)6.189 F 1.189(Older v)6.189 F(er)-.15 E(-)-.2
E .002(sions of sendmail had a single timeout, typically set to tw)156 176.4 R
2.502(oh)-.1 G 2.502(ours. V)398.142 176.4 R .002(ersion 8 allo)-1.11 F .002
(ws the)-.25 F(con\214guration \214le to set timeouts for v)156 188.4 Q
(arious SMTP commands indi)-.25 E(vidually)-.25 E(.)-.65 E 26.5
(\2475.3.3 Error)102 204.6 R 1.06(messages are sent as From:<>.)3.56 F 1.059
(This w)6.059 F 1.059(as ur)-.1 F 1.059(ged by RFC 821 and reiterated by)-.18 F
.237(RFC 1123, b)156 216.6 R .237(ut older v)-.2 F .237(ersions of sendmail ne)
-.15 F -.15(ve)-.25 G 2.737(rr).15 G .237(eally did it properly)355.186 216.6 R
5.237(.V)-.65 G .238(ersion 8 does.)448.254 216.6 R(Ho)156 228.6 Q(we)-.25 E
-.15(ve)-.25 G 1.934 -.4(r, s).15 H 1.134
(ome systems cannot handle this perfectly le).4 F -.05(ga)-.15 G 3.633(la).05 G
1.133(ddress; if necessary)402.941 228.6 R 3.633(,y)-.65 G(ou)494 228.6 Q
(can create a special mailer that uses the `g' \215ag to disable this.)156
240.6 Q 26.5(\2475.3.3 Error)102 256.8 R 3.212(messages are ne)5.712 F -.15(ve)
-.25 G 5.712(rs).15 G 3.212(ent to <>.)275.628 256.8 R(Pre)8.213 E(viously)-.25
E 5.713(,s)-.65 G 3.213(endmail w)383.028 256.8 R 3.213(as happ)-.1 F 5.713(yt)
-.1 G 5.713(os)474.957 256.8 S(end)489.56 256.8 Q 6
(responses-to-responses which sometimes resulted in responses-to-responses-to-)
156 268.8 R(responses which resulted in ....)156 280.8 Q(you get the idea.)5 E
26.5(\2475.3.3 Route-addrs)102 297 R .111(\(the ugly `)2.611 F
(`<@hosta,@hostb:user@hostc>')-.74 E 2.611('s)-.74 G .111(yntax\) are pruned.)
389.124 297 R .112(RFC 821)5.112 F(ur)156 309 Q 1.001
(ged the use of this bletcherous syntax.)-.18 F 1
(RFC 1123 has seen the light and of)6.001 F(\214cially)-.25 E 1.124
(deprecates them, further ur)156 321 R 1.125(ging that you eliminate all b)-.18
F 1.125(ut `)-.2 F(`user@hostc')-.74 E 3.625('s)-.74 G 1.125(hould you)462.595