X-NEWS: ids alt.sys.amiga.uucp: 909 Path: paperboy.ids.net!uunet!tymix.Tymnet.COM!tardis!olivea!spool.mu.edu!howland.reston.ans.net!xlink.net!fauern!imp.ch!alphanet.ch!uucpfaq From: uucpfaq@satcom.alphanet.ch (UUCP Faq Handler) Newsgroups: alt.sys.amiga.uucp,alt.answers,news.answers Subject: alt.sys.amiga.uucp Frequently Asked Questions (FAQ 1/2) - AmigaUUCP general information Summary: AmigaUUCP installation, utilities and common problems. Message-ID: Date: 28 Nov 93 15:00:09 GMT Expires: Wed, 5 Jan 1994 08:00:00 GMT Sender: UUCP-Faq@alphanet.ch (UUCP-Faq handler) Reply-To: UUCP-Faq@alphanet.ch Followup-To: alt.sys.amiga.uucp Organization: ALPHANET NF, Colombier (NE) - Switzerland Lines: 644 Approved: news-answers-request@MIT.Edu Comment: This is an automated monthly posting, part 1 of 2. Xref: paperboy.ids.net alt.sys.amiga.uucp:909 alt.answers:657 news.answers:4235 Archive-name: amiga/AmigaUUCP-FAQ/part1 AMIGA-UUCP-FAQ version 2.A.2 [Posting 18] MONTHLY POSTING, last update 02 August 1993 This FAQ is posted monthly (28th of month) author: Marc SCHAEFER, Bugs, typos, ideas to ("ch" stands for Switzerland) NOTE: The primary goal for this FAQ is to prevent questions from looping over and over. If you have new and interesting material, post it to alt.sys.amiga.uucp with "Addition to FAQ" somewhere in the subject. I will add it for the next "release". You may also send any ideas, changes, flames, typos to the address UUCP-Faq@alphanet.ch. They will be incorporated in the next release with your name in the CHANGES section as a reward :-) NOTE TO UUCP-BEGINNERS: Please take some of your time and READ the UUCP documentation. Most of the questions posted on a.s.a.u are related to manual pages. This FAQ contains also some information on common problems and utilities. Don't forget to get the FAQS from news.announce.newusers. You may also read UUMAN:Standards (for UUCP internals) and UUMAN:how2usenet. CHANGES FROM ORIGINAL MATT DILLON'S FAQ ARE NOTED WITH A (*). To skip to a topic, search for the roman numeral surrounded by parenthesis. For example, (I). FAQ.1 (this file) (*) 0. Changes from last posting (*) I. Introduction to alt.sys.amiga.uucp[.patches] II. Introduction to AmigaUUCP (*) III. Principal utilities IV. Constructing mail addresses V. Using DCRON VI. US domain clarification FAQ.2 (a different post) (*) VII. Common problems (new, please submit things to go in here). VIII. Using SENDMAIL directly. (*) IX. Other UUCP utilities. (*) X. How to get UUCP stuff ? (*) XI. BBS software supporting UUCP. (*) XII. Other UUCP implementations for AmigaOS. (*) XIII. Unresolved topics. (0) RECENT CHANGES TO THIS FILE Changes since 28-October-93 are listed below [Last posting was posting 17] None (I) INTRODUCTION TO ALT.SYS.AMIGA.UUCP[.PATCHES] (1) Configuration ALT.SYS.AMIGA.UUCP and ALT.SYS.AMIGA.UUCP.PATCHES are two newsgroups dedicated to the UUCP system for the Amiga microcomputer, AmigaUUCP. Both news groups are gatewayed to two mailing lists containing additional recipients who would otherwise not have access to the ALT groups. That is, posting to an alt group will automatically relay to the appropriate mailing list, and mailing to the mailing will automatically relay to the alt group. If you do not have ALT group access and are not on the mailing list, and would like to be on the mailing list, send your request to: amiga-uucp-request@idiom.berkeley.ca.us and/or amiga-uucp-patches-request@idiom.berkeley.ca.us To get off the mailing list, you can send your request to either address. Matt Dillon manually reads this alias. Note that you must provide a proper return address as part of your signature if you are a UUCP node so he can properly format your return address. If you are on the internet (i.e. have a fully domained address), it isn't a problem. TO POST ARTICLES VIA THE MAILING LIST, send email containing your posting to either of the following two addresses: amiga-uucp@idiom.berkeley.ca.us amiga-uucp-patches@idiom.berkeley.ca.us Sending email to either address causes it to be automatically posted to the alt.sys.amiga.uucp[.patches] newsgroup. You do not have to be on the mailing list to be able to post via the list. Report any problems to: amiga-uucp-owner@idiom.berkeley.ca.us amiga-uucp-patches-owner@idiom.berkeley.ca.us (2) Usage Of [Note: Original author is Matt Dillon. See next comment] The purpose of alt.sys.amiga.uucp is to convey the bulk of any discussion relating to AmigaUUCP. Discussion, bug reports, questions, etc... The purpose of alt.sys.amiga.uucp.patches is for the posting of any source code, scripts, or binaries relating to AmigaUUCP. Full distributions will NEVER be sent over alt.sys.amiga.uucp.patches. Anybody may post to alt.sys.amiga.uucp.patches and, in fact, it is best that any code you wish to submit to be merged into the master distribution that Matt Dillon keep be submitted to this newsgroup instead of to me personally. This will allow anybody to pick off the code and immediately implement it on their own system without waiting for the next master distribution. Matt Dillon will also use alt.sys.amiga.uucp.patches to post updates to the current master distribution, generally small to medium sized SHAR or uuencoded LHARC files. Matt Dillon personally would like to get a system together so multiple-source postings can be archived in a text form instead of a uuencoded form because all netnews is compressed anyway, and compressing uuencoded lharc files generally makes the result larger than the original instead of smaller. (3) BUG / ENHANCEMENT REPORTS [Note that the following text author is considered to be the current UUCP source maintainer which seems to be Michael B. Smith, mbs@adastra.cvl.va.us] The alt.sys.amiga.uucp and alt.sys.amiga.uucp.patches groups are fed through a filter when they reach my machine, and any bug or enhancement reports of a specific format will be automatically extracted and appended to my TODO file. To issue a bug report or enhancement request, use the following format: ##B unique-id ## Note that there are TWO '#'s. ##B stands for a bug report, ##E stands for an enhancement request. WARNING! The ##'s must begin a line, you CANNOT PRECEDE ## WITH WHITESPACE. Doing so will result in the filter passing the report by. For example, the ##B/## lines in the example above, not being flush with the left margin, will be ignored by my filter program. The unique-id should be a unique identifier for your bug report, for example, I might use '##B dillon.23'. Do NOT encode the date in the unique ID because my filter program will automatically extract the Date: and From: fields from the news message header. Matt Dillon will use the ID when refering to previous bug reports rather than posting the whole bug report. (4) This FAQ sheet If you have information you think would be useful on this FAQ sheet, please submit it to UUCP-Faq@alphanet.ch. (II) INTRODUCTION TO AmigaUUCP This section consists of a brief introduction to AmigaUUCP. It is not meant to describe installation of the distribution. Installation of the distribution is more involved and best served by the instructions that come with the distribution. AmigaUUCP was originally derived from GNU-UUCP and UUPC (was UUPC derived from GNU? I dunno). This was several years ago. It eventually fell into William Loftus's hands who molded it into a workable system for the Amiga. From there, about a year later, it fell into my hands and has since remained. What little GNU/UUPC code remains is in uucico, and even that is rapidly disappearing. AmigaUUCP is now almost entirely made up of code written after the original port to the Amiga. At this point, there is no comparison at all between the older GNU/UUPC stuff and the state of the art AmigaUUCP distribution. AmigaUUCP is a public domain project, though not properly in the public domain because all authors involved have maintained copyrights on the code. legally, this may not mean much, but it does give us a sense of security and more control over what is done with the code. Be that as it may, the entire distribution, source and all, is available to anybody who wants it. There are about a dozen principal authors and a few dozen contributors, not to mention the hundreds of people who have sent in helpful suggestions and bug reportrs. What is AmigaUUCP? Well, if you are reading this article then you have some idea how EMAIL and NETNEWS works ... AmigaUUCP is a set of utilities and documentation to implement an EMAIL/NETNEWS link directly on your amiga. All you need to do is find what is known as a 'feed' site who is willing to give you a UUCP connection, and, of course, a modem with which to communicate with that feed. (III) PRINCIPAL UTILITIES AmigaUUCP is made up of a plethora of utilities. Many of the utilities mimic their UNIX counterparts but it should be noted that none are really based on actual UNIX C code except for those sections still existing from the original GNU/UUPC port. Only the major utilities are listed below: UUCico UUCico is the workhorse of the system. It calls your feed site via the modem and transfers both outgoing and incoming mail and news. This mail and news will have been previously stored by you or your feed site. It has been updated a lot, mainly for reliability reasons. Last version is uucico_sd3.lha. Getty Getty handles incoming calls. It allows remote login (interactive and uucico logins). Sendmail/RMail Sendmail/RMail is the workhorse of the MAIL subsystem. The two utilities are actually the same executable just renamed and I will refer to them collectively as 'sendmail' from now on. Sendmail handles incoming mail, breaking it apart and sending it to the appropriate mailbox, or re-queueing it if it is simply passing through your system to another system down the line. Sendmail deals with any aliases you might have defined and also with any domains you have defined for routing email. Sendmail also handles, under the aegis of 'rmail', all incoming mail. RNews RNews handles all incoming news, including local news you send out. It breaks apart compressed batches and creates an individual file for each article in the UUNEWS: directory. It also creates a directory for each newsgroup. A lot of patches have been made to increase reliability, and speed. BatchNews Batchnews compresses and batches any news you have sent posted into a single batch file, making its transfer to your feed that much more efficient. Read the Newssetup.doc in the distribution for more information on how to set up news. DMail DMail is the amiga's mail shell. It scans your mail box and presents mail in an orderly fashion, allowing you to reply to the mail and do other operations. DNews DNews is the amiga's news reader. It is not quite as sophisticated as RN but is getting there. It sports an intuition windowing system to make it easy to scan through news. UUcp UUcp (the command) can be used to copy files from your local system to some of your neighbours. Note that the way it is implemented on the AmigaUUCP system is a little different than in Unix. In Unix, as soon as the uucp command has been executed, a copy of the implied file is done in a data file in the spool directory. Then uucico copies it to the other unix system that extracts the file from the data file. In AmigaUUCP, if sending the file is only read while UUCico is online, and that explains why if you UUCP a file which path is NOT authorized in the UULIB:Security file, there will be an error while online. This prevents the ability to forward the file to another host. However most of the time in UNIX, uucp is very restricted. AmigaUUCP does not allow directory-deep file send. For sending to a far site, BMS is more convenient. (IV) CONSTRUCTING MAIL ADDRESSES (1) GENERAL Unfortunately, the internet mail system is made up of a huge number of nearly incompatible networks. Mail addresses are constructed with various types of punctuation that mean various things .. indeed, some punctuation means one thing in one domain and another in another domain. I have found that the absolute best way to construct a mail address is either with the '@' format or with a '!' path. If your feed is a 'smart' host, any fully domained mail address can be replied to with simply: user@fubar.subdomain.subdomain....domain dillon@overload.Berkeley.CA.US Any address with dots in it is called a fully domained address. Unfortunately, there are a few exceptions... any address ending with .UUCP is *NOT* I repeat, *NOT* a domained address... it's a hack that some sendmails will add to properly route the mail internally. This hack generally extends to the From: field of an email message, and AmigaUUCP will do this, but not being a domain, you cannot SPECIFY a .UUCP trailer in the To: address. For example, my UUCP address is: uunet.uu.net!overload Note that there is NO .UUCP specification tacked on to overload. Note also that when you specify your UUCP address in your signature you should start with a fully domained machine name, *not* one ending with .UUCP. On other fronts, some unexperienced administrators will give their machines a full domain name without properly registering it. If you have not registered your domain with the proper authorities, DO NOT GIVE YOUR MACHINE A FULL DOMAIN. For example, when I first connected to my feed, which is uunet, I did not have a .US domain and so my machine name was simply 'overload'. After I registered in the .US domain I changed my machine name to its registered equivalent, 'overload.Berkeley.CA.US'. (2) BANG PATHS Nearly all the systems on the internet accept what are known as bang paths. There are only a few exceptions. One of the design decisions for AmigaUUCP was to convert all addresses into bang paths before sending them out. There have been one or two sites (so far) that have been unable to run AmigaUUCP because the feed they picked was running news software so old it did not recognize bang paths. To those sites I say: find a different feed, AmigaUUCP would become extremely messy were I to implement UNIX sendmail style address parsing. A bang path work by specifying the exact path your mail is to go along, in the following format: first_machine!machine!machine!users_machine!user Any machine name in the path may be a fully domained name. If you have a smart feed it will be able to optimize the path accordingly. For example, the bang path to me would normally be: uunet.uu.net!overload!dillon If your feed has a STUPID mailer, it may be necessary to use a bang path to get *past* your feed to a nearby site that has a SMART mailer. For example, lets say your feed is named 'fubar' and has a dumb mailer. Let us also say that the feed has a UUCP connection to 'harvard' which just happens to have a smart mailer. To get your message to me you might use: fubar!harvard!uunet.uu.net!overload!dillon your feed may or may not accept harvard's fully domained name, which is harvard.harvard.edu, it depends on how stupid your feed's mail system is. If it does, it makes more sense to use: fubar!harvard.harvard.edu!uunet.uu.net!overload!dillon (3) INTERNET DOMAINS VERSES UUCP MAP ENTRIES The internet domain system is based on domain servers, real time servers residing on known machines that know all the machines in a particular domain and how to get to them. When you send mail through an internet machine, like this (assuming you have a UUCP connection to UUNET): uunet!caps.ibm.com!user uunet (actually uunet.uu.net) will talk to the domain server for the .COM domain to find caps.ibm.com (a name I made up). UUCP works differently. While the internet is a real time network, UUCP is a batch network. UUCP has what is known as a MAP entry for every UUCP site that submits one. If you are a new UUCP site just connected to your feed, you should send a MAP entry to the appropriate administrator. A MAP entry is *NOT* a domain entry. The UUCP MAPS are used by machines on the USENET to find other machines on the USENET without the aid of domains. Not all machines on the USENET use MAPS to find some destination. uunet.uu.net does, so here is an example. I can send email from overload to (again, a made up name): uunet.uu.net!fubar!user Even if uunet does not talk directly to fubar.. assuming fubar has a MAP entry. uunet will search its maps to find the best path to reach fubar, and then route the mail accordingly. The actual route that uunet constructs might be: mcsun!gab!fubar!user If your feed is a machine that does NOT use maps, then you must specify an explicit bang path to get past your feed to a site that does. For example, lets say your feed is named 'char00' and has a dumb mailer, but connects to harvard.harvard.edu via UUCP. You want to email me. you can do it in two ways: char00!harvard!uunet.uu.net!overload!dillon. or char00!harvard!overload.Berkeley.CA.US!dillon But, since your mailer is dumb, you would not be able to use: char00!overload.Berkeley.CA.US!dillon If, on the otherhand, char00 is a SMART USENET mailer that uses the USENET MAPS (but still isn't on the internet itself), you can use: char00!overload!dillon Finally, if char00 is on the INTERNET, you can use: char00!overload.Berkeley.CA.US!dillon (4) WHEN ALL ELSE FAILS - BOUNCED EMAIL email will bounce for a variety of reasons. The fact that the global email system is made up of so many different types of mail systems causes lots of havoc... in many cases a system will munge the path you attempt to send email through by misinterpreting it or by attempting to 'optimize' it. When all else fails, and your attempt to reply to a piece of email bounces, you may have to construct the return address by hand. Several possibilities come to mind. You want to use the 'h' command from dmail to look at the actual mail headers (use dmail's help command to get full info on the header command). You want to look at both the original message that was sent to you, and the headers of your BOUNCED reply. -------- SAMPLE OF ORIGINAL MESSAGE ------- From uunet!SASK.USask.CA!telepro!oliphant Fri, 28 Dec 90 13:04:57 PST Received: by overload.Berkeley.CA.US (V1.07/Amiga) id AA00000; Fri, 28 Dec 90 13:04:57 PST Received: from sask.usask.ca by uunet.UU.NET (5.61/1.14) with SMTP id AA22874; Fri, 28 Dec 90 01:30:48 -0500 Received: from herald.USask.Ca by SASK.USask.CA with PMDF#10255; Fri, 28 Dec 1990 00:30 CST Received: by herald.USask.Ca (5.57/GLH-1.0); Fri, 28 Dec 90 00:30:06 -0600 id AA01058 for amiga-uucp-patches-request@overload.berkeley.ca.us Received: by telepro.UUCP (1.05D/Amiga) id AA04612; Thu, 27 Dec 90 21:25:00 CST Date: Thu, 27 Dec 90 21:25:00 CST Message-Id: <9012280325.AA04612@telepro.UUCP> X-Envelope-To: amiga-uucp-patches-request@overload.berkeley.ca.us From: uunet!SASK.USask.CA!telepro!oliphant (Mike Oliphant) To: amyuucp@sask.usask.ca Subject: Mailing list Please add me to amiga-uucp-patches. Thanks. -- Mike Oliphant UUCP: alberta!herald!telepro!oliphant Internet: oliphant@telepro.uucp -------- ADDRESS I SENT MY RESPONSE TO ------ uunet!SASK.USask.CA!telepro!oliphant -------- SAMPLE OF BOUNCE THAT CAME BACK TO ME ------- From uunet!sask.usask.ca!postmaster Mon, 31 Dec 90 01:02:30 PST Received: by overload.Berkeley.CA.US (V1.07/Amiga) id AA00000; Mon, 31 Dec 90 01:02:30 PST Received: from sask.usask.ca by uunet.UU.NET (5.61/1.14) with SMTP id AA13985; Sat, 29 Dec 90 17:18:48 -0500 Date: Sat, 29 Dec 1990 16:18 CST Message-Id: X-Envelope-To: overload!dillon@uunet.UU.NET From: PMDF Mail Server To: overload!dillon Subject: Undeliverable mail: local delivery failure The message could not be delivered to: Addressee: telepro!oliphant Reason: %MAIL-E-LOGLINK, error creating network link to node TELEPRO -SYSTEM-F-NOSUCHNODE, remote node is unknown -------- END OF SAMPLE HEADERS -------------------- So, why did my response fail? First, I have to tell you something about mail headers: Except for Received: headers, intervening systems can and will turn the standard headers into mush. That is, the 'From ' encapsulation, the From: header, the To: header, even the Reply-To: header might be modified by an intervening system. There are only two things that are not mushed. They are the Received: headers and the mail message itself - which might contain the sender's signature at the end. This is a good reason to always put your email address in your signature, and always base it at a known internet node so anybody can figure out how to get back to you. A Received: header is PREPENDED by *EVERY* site a piece of email goes through, and is NEVER modified by any other site. These headers tell you *exactly* how the mail was routed. If you look at the original message, you will note that one of the machines, probably SASK.USask.CA, modified the From: line in an attempt to optimize it: From: uunet!SASK.USask.CA!telepro!oliphant (Mike Oliphant) Note that, by the From: line, SASK.USask.CA talks directly to telepro. The 'From ' encapsulation was also modified, and there is no Reply-To: header. When I sent my reply to SASK using From:, the mail bounced because SASK was unable to find telepro ... if you look at the Received: lines you can see why ... because telepro talked to Herald before getting to Sask. It is amusing because SASK is probably the node that ripped out Herald's name in the From: and 'From ' lines in the first place. Also, take a look at Mike's signature line: Mike Oliphant UUCP: alberta!herald!telepro!oliphant Internet: oliphant@telepro.uucp Interesting, eh? The Internet: address is actually wrong (sorry Mike!) using .UUCP is not legal because it is not a proper domain. However, if you forward through an internet host that also uses the UUCP MAPS, and assuming mike is in the maps, the address *will* work. It's the first address that confirms our fears... mike shows telepro talking to herald. This combined with the knowledge we gained from the Received: lines tells us that the path: SASK.USask.CA!herald!telepro!oliphant Will work as a return address. When in doubt, trace the Received: headers to determine the return path. Sometimes a UUCP MAP entry will be incorrect, in which case using the Received: headers will be the ONLY way to reply to a message. There are some situations which are impossible to reply to ... if a message goes through a broken node that allows it to be propogated one way but not the other, even using the headers will not work. Also, some sites will attempt to optimize the path you specified. If SASK.USask.CA were to optimize the path: SASK.USask.CA!herald!telepro!oliphant To SASK.USask.CA!telepro!oliphant Before processing, the mail could fail due to SASK.USask.CA breaking itself. There are many nodes, especially gateways between networks, that are broken in this manner and there will be times when you will not be able to reply at all. (V) USING DCRON Many AmigaUUCP users leave their machines on 24 hours a day. With the advent of 2.0, and assuming the serial.device gets fixed, you can conceivably run your Amiga 24 hours a day under a heavy load for weeks without a crash. DCron is a program that runs in the background and executes other programs at intervals defined in S:CRONTAB. It is quite flexible.. you can run a program or script at specific times of day, every X minutes, only on certain days of the week, or even only in certain months! I will not discuss the actual format, that can be looked up in UUMAN:DCron. There are two reasons to run DCron: (1) Maintenance. (2) Automatic polling. If you call a system on a regular basis and want to automate the process, you can run UUCico from DCron at specific times of the day. First maintenance. Programs like UUCico, Getty, DCron itself, and sendmail generate log files which, if left alone, would eventually fill up your disk. Also, if you are receiving NEWS, you need to delete expired articles. Due to the volume of news, not deleting old articles can fill up your HD very quickly. The TRIMFILES utility trims log files to a specified number of lines, default 100. I normally run TRIMFILES on the various log files once a day early in the morning. The S:CRONTAB entry I use is: # trim log files at 3:01 A.M. 1 3 * * * uucp:c/trimfile tmp:dcron.log uu:spool/logfile getty:logfile Note that the file paths will be somewhat different for your system. Second, keeping your UUNEWS: directory reasonable. The TRIMNEWS utility will handle this. TRIMNEWS scans your UULIB:Newsgroups file for the list of newsgroups, then scans each news group deleting articles over N days old, where N is specified in the Newsgroups file. A sample NewsGroups file might be: comp.sys.amiga 7 comp.sys.amiga.tech 7 comp.sys.amiga.programmer 7 comp.sys.amiga.announce 7 alt.sys.amiga.uucp 14 alt.sys.amiga.uucp.patches 30 Which essentially tells TRIMNEWS to delete all articles in comp.sys.amiga.* over 7 days old (7 days from reception), to delete all articles in alt.sys.amiga.uucp over 14 days old, and to delete all articles in alt.sys.amiga.uucp.patches over 30 days old. I normally run TRIMNEWS in the morning too, my S:CRONTAB file has: # run TRIMNEWS at 3:06 A.M. 6 3 * * * uucp:c/trimnews --- DCRON is also useful to control the modem configuration. You can run the Getty utility from DCron to turn off the modem speaker while you are asleep. I use DCRON for other things as well, such as to automatically revise UUNET's amiga-uucp[-patchces] mailing list whenever I make a local change, and to backup my hard disk. I also use it to post this sheet once a month. (VI) .US DOMAIN CLARIFICATION This is a clarification to the information on registering in a .US domain. It turns out that you can register in the .US domain even if your 'feed' node is NOT on the internet. What you need to do is find some node that IS on the internet that is willing to be an MX FORWARDER to your machine (via a path). This might prove difficult, but it is possible. END OF FAQ PART 1.