1 2 3 K N O W N B U G S I N S E N D M A I L 4 (for 8.6.5) 5 6 7The following are bugs or deficiencies in sendmail that I am aware of 8but which have not been fixed in the current release. You probably 9want to get the most up to date version of this from FTP.CS.Berkeley.EDU 10in /ucb/sendmail/KNOWNBUGS. 11 12This list is not guaranteed to be complete, especially for fixed bugs. 13Many bugs are reported and fixed without ever making it as far as this 14file. See the file RELEASE_NOTES (in the root directory of the sendmail 15distribution) for more details. 16 17 18+----------------------------------------------+ 19| THE FOLLOWING PROBLEMS ARE STILL OUTSTANDING | 20+----------------------------------------------+ 21 22 23* Null bytes are not handled properly. 24 25 Sendmail should handle full binary data. As it stands, it handles 26 any value from 0x01-0xFF in the body and 0x01-0x80 and 0xA0-0xFF in 27 the header. Notably missing is 0x00, which would require a major 28 restructuring of the code -- for example, almost no C library support 29 could be used to handle strings. 30 31* Duplicate error messages. 32 33 Sometimes identical, duplicate error messages can be generated. As 34 near as I can tell, this is rare and relatively innocuous. 35 36* No "exposed users" in "nullrelay" configuration. 37 38 The "nullrelay" configuration hides all addresses behind the mail 39 hub name. Some sites might prefer to expose some names such as 40 root. This information is always available in Received: lines. 41 42* $c (hop count) macro improperly set. 43 44 The $c macro is supposed to contain the current hop count, for use 45 when calling a mailer. This macro is initialized too early, and 46 is always zero (or the value of the -c command line flag, if any). 47 This macro will probably be removed entirely in a future release; 48 I don't believe there are any mailers left that require it. 49 50* If you EXPN a list or user that has a program mailer, the output of 51 EXPN will include ``@local.host.name''. You can't actually mail to 52 this address. It's not clear what the right behaviour is in this 53 circumstance. 54 55* REDIRECT aliases don't work with `n' option. 56 57 If you have option `n' set when you use newaliases and have 58 REDIRECT addresses in your aliases file, you'll get the error 59 messages during the newaliases instead of when email is sent to 60 the address in question. The workaround is to turn off the `n' 61 option. 62 63* MX records that point at non-existent hosts work strangly. 64 65 Consider the DNS records: 66 67 hostH MX 1 hostA 68 MX 2 hostB 69 hostA A 128.32.8.9 70 71 (note that there is no A record for hostB). If hostA is down, 72 an attempt to send to hostH gives "host unknown" -- that is, it 73 reflects out the status on the last host it tries, which in this 74 case is hostB, which is unknown. It probably ought to eliminate 75 hostB early in processing. 76 77* NAME environment variables with commas break. 78 79 If you define your NAME environment variable to have a comma 80 (e.g., ``Lastname, Firstname''), and you are using the $q definition 81 that uses ``name <address>'' format, sendmail treats the first and 82 last names as two addresses, thus producing a bogus From line. You 83 can work around this by changing the $q definition to use 84 ``address (name)''. 85 86* \231 considered harmful. 87 88 Header addresses that have the \231 character (and possibly others 89 in the range \201 - \237) behave in odd and usually unexpected ways. 90 91* DEC Alphas (OSF/1 1.3) sometimes time out on sending mail. 92 93 I have one report that DEC Alphas acting as SMTP clients sometimes 94 will apparently not see the "250 OK" message in response to the 95 dot that indicates the end of the message. This only happens if 96 the message is run from the queue -- if it gets through on first 97 try, everything is fine. I have been unable to reproduce this 98 problem at Berkeley. 99 100* accept() problem on SVR4. 101 102 Apparently, the sendmail daemon loop (doing accept()s on the network) 103 can get into a wierd state on SVR4; it starts logging ``SYSERR: 104 getrequests: accept: Protocol Error''. The workaround is to kill 105 and restart the sendmail daemon. We don't have an SVR4 system at 106 Berkeley that carries more than token mail load, so I can't validate 107 this. It is likely to be a glitch in the sockets emulation, since 108 "Protocol Error" is not possible error code with Berkeley TCP/IP. 109 110 I've also had someone report the message ``sendmail: accept: 111 SIOCGPGRP failed errno 22'' on an SVR4 system. This message is 112 not in the sendmail source code, so I assume it is also a bug 113 in the sockets emulation. (Errno 22 is EINVAL "Invalid Argument" 114 on all the systems I have available, including Solaris 2.x.) 115 116* Sending user deletion not done properly in :include: lists. 117 118 If you don't have the "m" (me too) option set, then a person 119 sending to a list that contains themselves should not get a copy 120 of the message. However, if that list points to a :include: file 121 that has one address per line, this will break, and the sender 122 will always get a copy of their own message, just as though the 123 "m" option were set. 124 125 You can eliminate this by adding commas at the end of each line 126 of the :include: file. 127 128* Excessive mailing list nesting can run out of file descriptors. 129 130 If you have a mailing list that includes lots of other mailing 131 lists, each of which has a separate owner, you can run out of 132 file descriptors. Each mailing list with a separate owner uses 133 one open file descriptor (prior to 8.6.6 it was three open 134 file descriptors per list). This is particularly egregious if 135 you have your connection cache set to be large. 136 137 138+-------------------------------------------+ 139| THE FOLLOWING PROBLEMS ARE FIXED IN 8.6.5 | 140+-------------------------------------------+ 141 142* Route-addrs missing angle brackets. 143 144 There are cases where route-addrs do not get angle brackets around them, 145 such as in the "-r" flag on mailers or in the From_ line created when 146 mailing to files. 147 148* No "exposed users" in "nullrelay" configuration. 149 150 The "nullrelay" configuration hides all addresses behind the mail 151 hub name. Some sites might prefer to expose some names such as 152 root. This information is always available in Received: lines. 153 154* owner-* alias that uses :include: broken. 155 156 If you have aliases set up as: 157 158 owner-listname: :include:/some/file 159 160 sendmail will break because it considers this a "sender address", 161 which is not permitted to use the :include: syntax. The easiest 162 workaround is to change this to: 163 164 owner-listname: :include:/some/file, 165 166 (note the trailing comma); a somewhat cleaner solution is to use: 167 168 owner-listname: listname-request 169 listname-request: :include:/some/file 170 171* "SYSERR: openmailer(local): fd 1 not open" message 172 173 File descriptor 1 (standard output) should not be closed during normal 174 processing. This is checked periodically, and sometimes this condition 175 is found and this message is produced. Sendmail repairs the problem, 176 and the mail is still delivered, but I still don't know why it happens. 177 (There was a bug that was fixed in 8.6.beta.13 that might be related, 178 but I think this bug still exists.) 179 180(Version 8.16, last updated 02/06/94) 181