SpamFryer is a utility which deletes unsolicited emails from an ISP's server, without you having to download them. Over the last five years it has reliably got rid of 85 to 95 per cent of the spam directed to the author's email accounts, without deleting any messages he actually wanted to read. That said, your mileage may vary. SpamFryer is free and you're free to use it or not, at your own risk.
A plain text version of this hypertext document is in the file SpamFryer.text. Apart from the links and formatting, the content of both files is the same. This HTML version is for people who prefer to read things in a web browser. The plain text suits those who prefer to refer to a printout, or be sure they've read everything. An AmigaGuide version is also available, SpamFryer.guide, for those who prefer to read this in MultiView.
Contents | Index | < Browse | Browse >
AmiTCP or compatible TCP/IP stack (e.g. Genesis, Miami or Roadshow).
An account with an Internet Service Provider that offers POP3 email (virtually all do).
ARexx (add-on or as bundled by Commodore with Kickstart 2 and 3).
A network-capable Amiga (A500 to A4000, or authentic emulation e.g. Amithlon, or UAE running a genuine Amiga TCP/IP stack), or an AmigaOne PowerPC system.
Contents | Index | < Browse | Browse >
Before you try to use SpamFryer you need to configure it to access your internet service provider and email account. If you are unable to edit a text file, please give up at this stage :-(. Presuming you know how to edit a text file, and can get hold of your own account details (I can't help you with this!) there are two ways to do this - by adding your details to the SpamFryer.accounts file (accountFile), or embedding them in the SpamFryer.rexx script. It's simpler to put them in the SpamFryer.accounts file as long as you don't mind keeping several SpamFryer files together in one directory.
Three bits of information are needed before SpamFryer can log in to an email account and check for spam on your behalf. One of these, the password, is optional - if you choose not put it in the text files (e.g. for security reasons, if you are running a server and allow remote users access to all drives on your machine) SpamFryer will pop up a requester, requiring you to type in the password each time it needs access to the relevant account.
This information appears in the YAM2 mailer's Settings/Configuration/First Steps dialogue, and in the yam:.config file (under POP00.Server, POP00.User and POP00.Password - though the latter is deliberately obscured in the YAM files, for security reasons). If all else fails, you can get the password, or a replacement one, from your Internet Service Provider (ISP).
Typically you'll include three lines in SpamFryer.accounts for each POP3 mail account you want SpamFryer to check, e.g.
Hostname: mail.isp.org Username: ISP_Account_Name Password: guessthis
or
Hostname: 194.5.6.7 Username: JoeBlogs Password: tribbles
HostName identifies the mail server (the machine that stores emails for you to collect) on the internet. This can be an IP number e.g. 194.5.6.7 or a domain name like mail.demon.net
UserName is the name your computer uses to identify itself, and you, to the mail server. Typically this is your ISP account name.
You can give details of as many accounts as you like. SpamFryer will check them all, in the order you specify them. The Hostname must be specified before the corresponding UserName (this is a change since earlier versions which removes a risk that a badly- formatted line might invalidate the remainder of the account information). If you miss out the password SpamFryer will ask for the password interactively. This feature no longer requires Workbench 2, reqtools.library and rexxreqtools.library, though it will use those if available.
You can put comments in any of the SpamFryer configuration files (SpamFryer.accounts, SpamFryer.keepList and SpamFryer.loseList) by starting the line with two slashes and a space, like this:
// this is a comment
(Important: Read also the section about the sortMethod option.)
Blank lines are allowed, and ignored.
Alternatively, if you'd prefer to keep all your changes in one file and be able to run SpamFryer without needing other files, add three lines near the start of the file SpamFryer.rexx, after the comment /*** CONFIGURATION BLOCK **/ and before the line starting keepList, as follows:
Account.1.1 = "mail.isp.net"
This can be an IP number e.g. 194.5.6.7 or a domain name like mail.demon.net
Account.1.2 = "yourname"
This needs to be your user name - the name your computer uses to identify itself, and you, to the mail server. Typically this is your ISP account name.
Account.1.3 = "password"
This should be the confidential login password you use for email. If you run any sort of server on your Amiga, you should read this from somewhere that people logging on to your machine can't access, or they might extract your password from the SpamFryer script (or any other software you use for email access). If you make a new version of SpamFryer, remember to remove this before you send the modified code to anyone else!
SpamFryer can check mails from more than one server, and make a log of what it finds and does on each one. To do this, add additional lines for other accounts, called Account.2.1, Account.2.2 and Account.2.3 for the same three items of information listed above for your second server, and more (e.g. Account.3.1, etc) for as many other accounts as you like. Alternatively you can put these details in the SpamFryer.accounts file, or keep some in the file and some in the script - which may be handy if you want to change part of the configuration (e.g. by renaming several copies of the accounts file as required) while always checking servers listed in the script.
You don't need to provide any more information to use SpamFryer. But you might like to adjust the reports it generates to make it easier to understand and tune what it's doing. If you'd like to know about this now, click for details of these options.
Otherwise, follow this link to get using SpamFryer right away.
Contents | Index | < Browse | Browse >
You don't need to provide any information other than the server and user names and a password to use SpamFryer. However there are a few other lines in the configuration block which you may wish to add or alter to tune SpamFryer for your system. Of course, you're free to change the whole script, but the block at the start contains the main configurable lines.
Four lines specify files that SpamFryer will attempt to read or write. You can edit these lines to change the name or location of the files. The first three are optional data files, presumed to be in the same directory as the SpamFryer.rexx script unless you alter the assignments to include an alternative path (e.g. WORK:Secure/SpamFryer.accounts):
Other options are:
Contents | Index | < Browse | Browse >
If the 'verbose' variable is set to 1, in the script or by passing an eponymous argument when the script is called, for instance
rx SpamFryer.rexx VERBOSE
the log will contain lots of extra information, including lists of account details and rules, header information from each mail checked, and a note on what SpamFryer has done about it. By default the logFile just reports the subject and number of each mail it fries, and information about errors in the configuration files, with simple headings. The same information is written to a console window as the script runs, so you can see what it's doing as it goes along.
SpamFryer6 adds an option to store only the log for the latest session, rather than to append each new log to any existing file. The variable 'lastLogOnly' controls this. By default it is 0, for compatibility with previous releases, but if you set it to 1 the old log will be deleted and replaced with a log for each latest SpamFryer session.
<- Configuring SpamFryer to suit yourself
Contents | Index | < Browse | Browse >
screenName gives the name of the public screen on which you want to see a temporary copy of the log as SpamFryer runs . Close this when you've finished with it. The default is '*', which opens a window on the current screen, but you can redirect SpamFryer to use the same screen as your mailer or any other application by editing the 'screenName' configuration.
If you set the screenName to '' no window will be opened unless you run SpamFryer 6 in interactive mode by giving the BROWSE argument, in which case both windows will be opened on the current screen.
If you want to use a particular font in the windows of the interactive (BROWSE) mode, edit the SpamBrowser text file and add a 'setfont' line specifying the name and size of the font you prefer. The script contains the following line, commented out for now, to select the classic Amiga font in 8 point size:
setfont Topaz 8
Edit this to set any non-proportional font and size you prefer. Proportionally-spaced fonts like Helvetica and Times will not work well as the browse view uses fixed-width columns.
<- Configuring SpamFryer to suit yourself
Contents | Index | < Browse | Browse >
lineMax sets the maximum number of characters of subject lines reported to the screen. Some mails - usually spams - have very long subjects, so this limits the maximum amount of screen space such a line can take up.
maxLineSize sets the maximum length a line in a mail can be before it is considered abnormal or dangerous (potentially causing buffer overflow). Setting this variable to 0 or anything below 80 causes the limit to be ignored.
Note: RFC2822 (which supercedes RFC822) specifies that:
"Each line of characters MUST be no more than 998 characters, and SHOULD be no more than 78 characters, excluding the CR/LF."
maxMailSize sets the maximum total size in bytes for an acceptable mail. Setting this variable to 0 or anything below 1024 causes this limit to be ignored.
Note that these mail limits are respectful of the keep rules (which override all lose rules) - so if a mail is identified via your keepList it will not be deleted regardless of line lengths or its total size.
<- Configuring SpamFryer to suit yourself
Contents | Index | < Browse | Browse >
SpamFryer 6:18 adds an option to sort rules so that the most effective are tried first. This is not switched on by default, as it removes comments from the rule files (not knowing how to sort them), and its effectiveness depends very much upon how many rules you have defined.
When sortMethod is set to 1 or 2 SpamFryer will sort your keepList and loseList. Leave it set to 0 if you don't want your hand-made keep/lose lists to be over-written (and to lose your comments and own order of rules). But note that a backup copy of your lists will be made (with a ".old" filename extension) during the first run of SpamFryer with the sorting option enabled.
When set to 1, all the rules are sorted in descending order on their "Trigger Hits" score.
When set to 2, the rules are grouped by field types ("From:", "Subject:", ...), then sorted in descending order upon each group "Trigger Hits" score. This sorting method is faster than the first as it saves the rules grouped as they are used internally. However the lists can end up longer because there are more "Trigger Hits" lines added, and they are a bit less convenient for humans to read.
When sorting is activated, you will see in your keep/lose lists that the rules have been sorted, and that new comment lines like this
// Trigger Hits: 342
have been added here and there. Those lines tell you how often the rules following them have matched when checking mails.
The keepList or the loseList are re-sorted if at least one mail has been automatically deleted from the server by SpamFryer. If, in interactive mode, you have unselected some of the mails previously marked as 'to delete' by SpamFryer, the rules triggered for those mails will still score one more Hit, even if the same mails will re-appear in the next session of SpamFryer. Now you know, consider this a feature ;-) .
(See also: "How can I make SpamFryer faster?".)
<- Configuring SpamFryer to suit yourself
Contents | Index | < Browse | Browse >
'noDigits' sets or clears an flag which marks mails with any numeric digits in their To: address for zapping. Many spammers make up email addresses for known servers, and if you have your own domain this rule helps you sift out those non-existent user names which would otherwise be delivered to your Amiga. BEWARE: This flag may delete wanted mails (e.g. if you are cc'd on a mail to an Yahoo, Freeserve or AOL victim with a human-hostile user name) unless you apply other rules, e.g. in the keepList, to make sure mails that are specifically addressed to you don't get flushed with the spams.
<- Configuring SpamFryer to suit yourself
Contents | Index | < Browse | Browse >
This optional extension from SpamFryer 5 catches spams that use MIME or redundant accents or equivalent characters to hide their unwanted payload from brute-force searches.
Fuzzy matching is controlled by the fuzzyMatch variable in the configuration block near the start of the SpamFryer.rexx script. If this is 0, all matches are case-independent but otherwise exact. If spam includes a phrase in your loseList but adds redundant accents - so viagra appears as viägrä, for example - the modified word will NOT be recognised. If fuzzyMatch = 1 the accents will be ignored - accented characters will be translated to their Latin counterparts - so the modified word WILL match the normal one it closely resembles. :-)
This could be considered an extension of the upper and lower case equivalence that catches ViAgRa as well as Viagra and viagra. It is safest to use if you and your correspondents use an unaccented Latin-based language; if you use accents, you may prefer to set fuzzyMatch = 0 to reduce the risk of vital distinctions in your own language being misinterpreted, and to allow you to use acents in your own loseList and keepList rules. Since fuzzyMatch strips accents, you must likewise remove accents from your rules, or they'll never be matched!
Depending on the characters you use, the
header = TRANSLATE(header ...
lines in SpamFryer.rexx may merit customisation. Most of those specify a set of characters any one of which will be considered equivalent to a single one following, e.g. any of 'ÈÉÊËèéêë' are remapped to 'E'. The final TRANSLATE line is rather different in effect, as you can probably tell from the syntax. Rather than strip accents it replaces three characters among the ANSI Latin 1 extensions to the original ASCII set with the Latin ASCII characters they most closely resemble. £ is considered equivalent to E, the multiplication symbol × is equivalenced to X and x, and ß is treated like a Latin letter B or b.
It is not a good idea to include any ASCII characters (those with ANSI codes less than 128) among those remapped, as that would prevent any of those being matched if they appear in rules.
<- Configuring SpamFryer to suit yourself
Contents | Index | < Browse | Browse >
The emptyMIME option was implemented in SpamFryer 5 to meet the challenge of increasing numbers of spams that use MIME blocks to conceal spam indicators in HTML or images, later in the mail.
Since SpamFryer tries to identify unwanted mail without reading the entire message (which could be large and hard to decode in the case of HTML and image attachments) Simon Goodwin and Allan Rasmussen have compared hundreds of spams of this type with others containing wrappers in MIME (Multipurpose Internet Mail Extensions) format, and come up with a rule that sifted out almost all mails of this type (as of 2006) with only a small - but still problematic - risk of falsely identifying wanted mail as bulk-mailed spam.
The emptyMIME routine searches for lines indicating a complex MIME structure followed by two blank lines, at the top of the message. Those are flagged as having a 'suspicious MIME header' and will be deleted from your mail server unless they match rules in your SpamFryer.keepList.
We don't know why some, or which, programs sometimes generate two blank lines after the MIME header, but statistically it seems to be a rare phenomenon in mails that are not spam.
Out of more than 2000 "good" emails in Allan's test set, only one was zapped by this option. One out of 2000 is not bad, but still a little bit worrying, so we do not recommend you use this option without a clever and well-kept SpamFryer.keepList.
We welcome ideas to improve this code so that it catches more spam and fewer mails you're likely to want to read. The MIME format is explained at
http://mailformat.dan.info/headers/mime.html
To enable this rule, set emptyMIME = 1 in the configuration block of the SpamFryer ARexx script. By default it is set to emptyMIME = 0, which means the rule is NOT applied and spams of this type - along with similar messages you might actually want to read - are NOT sifted out by SpamFryer.
<- Configuring SpamFryer to suit yourself
Contents | Index | < Browse | Browse >
You can start SpamFryer by clicking on the supplied icon or with an AmigaOS command (rx SpamFryer.rexx when SpamFryer.rexx is in the current directory or on your path). The script has been successfully tested on AmigaOS 3.0 with Genesis, versions 3 and 4 of AmiTCP, Miami and Roadshow, the AmigaOS 4 stack. It will not work with the tiny TCP/IP implementation forwarded from the host by UAE unless that implements the AmigaOS TCP: device (see FAQ 4).
You can also configure it to be launched as an ARexx script by Miami, by integration into the Thor mailer (using the special version provided) or run it from the console. One optional parameter 'verbose' makes it list the details of mails it leaves, as well as those it zaps, in the console log. (See also the Log options section.)
You do need to be online and have the ARexx server running when you start SpamFryer. The easiest way to ensure the latter is to put RexxMast in your sys:wbstartup directory when you boot your Amiga - but if some other programs loaded early in your boot sequence (such as MiamiDX) need ARexx you may find you need to launch it from your S:user-startup script. ARexx was a standard component of Commodore's AmigaOS versions 2 and 3 and sold as an add-on for other versions.
Once launched, SpamFryer checks through emails on a POP3 host and purges the unwanted ones by applying simple rules that sift out nine-tenths of the mass-produced junk mail and attempts by dimwitted PC trojan and virus programs to mate with Amigas. While these viruses die out when they reach a real computer (indeed, anything not running Windoze) they are large, arrive many times over, and are of course useless.
SpamFryer shows a summary of waiting messages and deletes any with html right at the start, or subjects or a recipient address known to be bogus. It includes special sifting for Windoze .PIF worms, and mails flagged by SpamAssassin on the server. It's easy to extend for similar cases (e.g. new viruses or trojans and mails from particular idiots who unfortunately have your address) but I find it catches 80..90 per cent of spams without adding any extra code - if I get a rash of new spams that it misses I add an extra rule or two. New rules that I and other users value end up on the Wiki page - please post yours, too!
If SpamFryer gives an error message, click here for information about problems and solutions, Frequently Asked Questions and answers.
If that still doesn't resolve your problem, click here to find out how to contact the author for personal advice. You may have found a problem that no other user has run into. But make sure you've read the documentation first, or you should expect short shrift!
Contents | Index | < Browse | Browse >
SpamFryer now supports the standard "?" argument. When calling SpamFryer with that argument, a template with the available arguments will be shown, and you will be prompted to enter one or more (or none) from the following options:
The 'BROWSE' switch (without quotes) is the most important new feature. It allows you to start in interactive mode that shows you a preview of the waiting mails in your POP account, and allows you to select the mails you want to delete, as explained in the SpamBrowser usage section of this document.
The 'LANGUAGE' argument (without quotes, but keyword needed)
is only useful if you want to use another language than the
one configured as your preference in Workbench (or if your
OS does not support Locale programs). Otherwise SpamFryer
will run automatically in your favourite (or native) language
as long as a translation file is available for that language.
Tip: If you run SpamFryer transparently (without any output,
that is without the BROWSE and VERBOSE arguments, and with the
screenName set to ''), you may want to add LANGUAGE='english'
as an argument. This will force SpamFryer to run with the built-in
default English text, and skip any attempt to load a language file.
The 'SORT' argument (without quotes, keyword required) overrides the settings for the sortMethod option. Acceptable values are 0, 1 or 2.
The 'VERBOSE' switch affects the log text output. If VERBOSE is specified SpamFryer outputs more details about what is going on.
Contents | Index | < Browse | Browse >
When started with the 'BROWSE' (without quotes) switch, SpamFryer will enter an interactive mode known as SpamBrowser. The same script and configurations files are used - the only difference is the switch and the opportunity to review waiting mail after it has been checked by SpamFryer and before the server is updated.
If you have more than one POP account, a requester will ask you to select the account you want to browse. If you have fewer than 10 accounts configured, you can quickly select one by pressing the numeric key shown in front of the account name.
Then SpamFryer will collect details of the mails waiting in the selected account. If more than ten mails are waiting in the mailbox, a countdown will inform you of how many mails have yet to be spam-checked.
When done, another window will open. This is where you will see the waiting mails (and the ones already flagged as 'to delete' by SpamFryer), and be able to select (or unselect) those you want to be deleted (or not).
If you press the HELP key, the first window will tell you how to navigate through the selection (BROWSE) window. You need to make sure the selection window is selected when scrolling through it.
The selection window looks rather like this:
----------------------------------------------------------------- |SpamFryer 6:18 -- Previewing mails waiting in pop.myisp.com | |---------------------------------------------------------------| | | Honey | Wake up, lazy boy ! | |*| Free pills | How to get a one meter dick easily | ... ... |*| Casino | 300 $ Free | |*| =?koi8-r?B? | =?koi8-r?B?gkjhdffgd;b,nnbfg;fg,;nns +| -----------------------------------------------------------------
The first column marks with a '*' the messages flagged as 'to be deleted'. Press the SPACE bar or M key to toggle that flagging. Even if you mark a message for deletion, it will be left alone if it matches a rule in your keepList; if so, SpamFryer will tell you.
The second column shows the TO: field, and the third column shows the SUBJECT: field.
The [+] symbol which might appear at the bottom right corner of the window indicates that there are more messages to be seen. Press the down-arrow key (or SHIFT down-arrow to page down) to see the later messages.
The reverse-coloured bar shows you the currently selected message. Press ? or I to get more information about that message. The main window (above the browser window) will then shows you the From: field, the To: field, the Subject: field, and the status of the message (a short explanation about why the message is marked as 'to delete', or why it is to be kept).
When you have finished reviewing and marking mails, press ENTER or A to delete the selected mails (marked *) from the server.
If you press Esc or Q at any time the SpamFrying process for the selected account will be cancelled, with no changes on the server.
Contents | Index | < Browse | Browse >
SpamFryer comes with more than 100 pre-programmed rules to help it identify spam and sift out mails that you might really want to read. You don't need to alter these to use SpamFryer, though it's a good idea to tune them to suit your system once you understand what SpamFryer can do for you.
If you'd like to know more about customising SpamFryer for the sort of mail you're likely to receive, either by adding and removing rules or modifying the script, click here.
Contents | Index | < Browse | Browse >
To make SpamFryer run automatically when Miami goes online, use a MiamiOnline.yam script something like this:
miamipath = "WORK:COMMS/MIAMI/Miami" /* Change this if required */ Spampath = "work:comms/SpamFryer/SpamFryer.rexx" /* Add this line */ OPTIONS RESULTS IF ~SHOW('P', 'MIAMI.1') THEN DO ADDRESS COMMAND 'Run <>NIL:' miamipath 'WaitForPort MIAMI.1' END ADDRESS 'MIAMI.1' ISONLINE IF RC=0 THEN ONLINE ISONLINE IF RC=0 THEN EXIT 5 HIDE ADDRESS COMMAND 'rx' Spampath 'wait 10' 'endcli' EXIT 0
<- Customising SpamFryer for your environment
Contents | Index | < Browse | Browse >
Here is an example of commands to be used to launch SpamFryer with a DOpus5 button:
AmigaDOS CD "Work:SpamFryer" ARexx SpamFryer.rexx
The first line is important if the path to SpamFryer.rexx contains any space as in such a case rx will not be able to run the script. The second line is the script itself. Add any argument you want to that line.
If you wish something else to be performed after SpamFryer has done its job (like loading a mail program), add these lines :
Command FinishSection Workbench <name of the program to run>
<- Customising SpamFryer for your environment
Contents | Index | < Browse | Browse >
Marcel Beck's YAM (Yet Another Mailer), now open source, is the premier Amiga email application. Two scripts contributed by users of SpamFryer are in the subdirectory YAM Support, with plain-text documentation.
MiamiOnline.yam modifies the script that launches Miami so that it also fires up SpamFryer, so your new email is filtered automatically, as noted in the Miami section.
SpamFryer.yam uses the rules and configuration of SpamFryer.rexx to scan mails already downloaded and move the likely spams from one YAM folder to another. It is most useful for those with continuous and fast internet connections, who do not need to eliminate download time but still wish to sift spams from their other mail automatically.
<- Customising SpamFryer for your environment
Contents | Index | < Browse | Browse >
SpamFryer.thor is an ARexx script for the shareware email and news client Thor, which uses SpamFryer rules to identify spam in POP accounts. It also has an interactive (browse) mode where you can preview waiting mails and select those you want to delete. SpamFryer.thor can be run with or without Thor loaded. For more details please see SpamFryer.Thor.readme, in the Thor Support drawer.
<- Customising SpamFryer for your environment
Contents | Index | < Browse | Browse >
The subdirectory Alternate Icons contains replacements for the default SpamFryer launch icon. These work the same way but vary in size and colouring to best suit old and new Amiga systems. The text file in the same drawer gives more details and there are ILBM images so you can see the variants on your own display, whatever screen mode and icon system you currently have installed.
<- Customising SpamFryer for your environment
Contents | Index | < Browse | Browse >
Since SpamFryer is a plain-text program you can be sure of what it does and doesn't do, and can easily extend it to do more - or less, e.g. just report likely spams without deleting them - by adding rules to the configuration files, or by editing the script. But if you can do it by editing the configuration, that's likely to be a lot easier and less risky than editing the script.
Contents | Index | < Browse | Browse >
At first SpamFryer automated the procedure I used to manually get rid of spams, by printing the commands automatically to save me typing them, and reading the message headers to find tell-tale signs of spam - these include HTML in the first few lines (i.e. a message with no plain text part); subjects known to be autogenerated by persistent spamming programs; mails marked as spam (but not actually deleted) by a server filter program like SpamAssassin, which was used by Amiga Inc; mails to any address you've abandoned but which is still redirected to you, and emails with .scr and .pif (etc) MIME attachments which are useless on Amigas and generally only a cause of infection to PCs... If anyone does really need to send you one of those, ask them to rename it or compress it into an archive with LHA, LZX, ZIP or whatever so that SpamFryer doesn't zap the mail.
At first SpamFryer only looked for reasons to delete a mail, and had to be conservative about such rules in case it accidentally deleted something important. Then Allan Rasmussen volunteered to add support for rules in separate files, and for rules that denote a mail that should be kept rather than lost (sometimes known as a whitelist). This means that tests can use shorter and more general patterns, as long as they are effective in identifying spams, as the risk of losing a wanted mail is reduced.
The snag of this is that the whole mail header and first part of every mail must be read by the script, in case a later line contains text that indicates the mail should be kept, despite spam matches earlier. This slows down the checks, as they can't skip a mail as soon as it's matched a spam test.
<- Customising SpamFryer for your messages
Contents | Index | < Browse | Browse >
SpamFryer now incorporates more than 100 rules in the two configuration files and in the script itself. Only the most complex yet effective rules are in the script, where they are hardest to lose. This means that SpamFryer has a basic but useful capability without any extra files (as long as you add your account details to the script).
SpamFryer.keepList contains rules that tell SpamFryer about mails you always want left on the server. These are mails you never want deleted, even if they contain text that is typically in junk mail. For instance, mails with SpamFryer in the checked text are always left on the server (unless you alter SpamFryer.keepList to remove the rule) so that people can quote messages that SpamFryer might delete (for better or worse) without having their discussion fried!
The tests look for references to 'Spam Fryer' (with a space) as well as SpamFryer (without) in case someone prefers to type the name as two words. The rule starts with the term 'Early-Ref:' so that the entire mail header and first few lines (up to the point where an HTML-only message can be identified) are checked for the pattern. This 'Early-Ref:' category is the slowest test; others look for certain lines in the mail header and only look further on the line, trying all the patterns in that category, providing the first part matches.
The first word in the list file tells SpamFryer what part of the mail to test - the quicker options test the From:, Subject:, or To: or Reply-To: lines in the header. The Colons and dashes are optional but recommended as they mean the category names correspond to those used in a mail header.
The remainder of each line in the keepList file indicates a pattern to look for. If that pattern appears anywhere in the part of the mail specified by the first word, the mail will be left on the server; it won't be fried even if it matches rules in the loseList.
Rules in the loseList work the same way, except that they detect patterns which may indicate junk mail, and can be over-ridden by any rule in the keepList. For this reason, and because spammers keep trying new scams, there are a lot more rules in the supplied loseList than there are in the keepList.
New rules for SpamFryer are updated continuously by users at http://www.amigawiki.com/cgi-bin/wiki.pl/SpamFryer
-> Safety first
<- Customising SpamFryer for your messages
Contents | Index | < Browse | Browse >
You should add your own name and any unpublished email addresses as 'To:' rules in your own SpamFryer.keepList to reduce the risk of mails explicitly addressed to you getting fried.
The default keepList contains certain user names which must be acknowledged by mail to any domain - RFC 2142 (one of the Internet standards, informally known as a 'Request For Comment') says that mails to postmaster, abuse, security and noc user names (for a server) must be handled by a human. If you don't own a domain, you can delete these rules. If you do, it's worth adding the rest of the address (e.g. abuse (at) vituperation (dot) org) to narrow down the default test.
A few more default rules let through mails associated with Yahoo group internet mailing lists, and those associated with Aminet and the YAM Amiga mailer. Add more rules to suit the lists you subscribe to, unless you'd prefer to have SpamFryer check those for spam references - some list servers filter spam well, but some act annoyingly as spam relays.
If you are concerned about the small risk of losing wanted mails due to the accidental use of terms common in junk mail, you should check the contents of SpamFryer.loseList and remove any rules that you consider might delete things you want to know of.
All the rules in that file are there because they have identified at least four (and often many more) bulk junk mails sent out between 2004 and 2008. But a few people might actually want unsolicited information about cable filters (even variously mis-spelt) penis enlargement or extended warranties, or might even send mails asking about such things - if so they should delete the rules containing those terms, or they may lose the replies.
<- Customising SpamFryer for your messages
Contents | Index | < Browse | Browse >
Comparisons ignore case, so THIS matches thIs and This, and any line with that word (even THIStle) will match. If you want to match a specific word, rather than part of one, put the pattern in 'single ' or " double" quotes with a space either side (as in the examples) - however if you have a space both sides you may not pick up such a reference at the start or end of a line.
Put // (remember the space) at the start of a line to temporarily disable a rule without removing it from the list. Blank lines are ignored, but can make sets clearer. But if you use the new option to sort rules so that the most effective are counted and moved up the list, blank lines and comments will be removed (as the sort cannot work out where to put them) so take a backup of your changes and comments before enabling the SORT option.
SpamFryer writes a report showing how many messages it has fried from each category, and how many it has left on the server. There might possibly still be a few spams there, but you'll have eliminated most of them by running SpamFryer before you collect mail in the usual way. As it's an ARexx script you can automate the process so the script runs first, whenever your mail client goes to get new mail - but I leave the details of this to you. :-)
<- Customising SpamFryer for your messages
Contents | Index | < Browse | Browse >
This section is for programmers who want do more than just add a rule which includes or excludes mails from SpamFryer's list of things to fry. You will need to understand a bit about ARexx programming, but not much. ARexx manuals come with Amiga OS 3.1, A4000s and some earlier Amiga systems. ARexx documentation is on-line (ARexxGuide2_0A is a good start) though there's rather less information about the ARexx interface to TCP/IP - I worked out how to write this by examining ChkMail (on Aminet) which was written in 1994 by Matt White <whitem (at) duke (dot) usask (dot) ca> and from basic knowledge of the POP3 protocol, which I previously used from an Amiga Shell via telnet to port 110 on my ISP's mail server (thanks to Peter Corlett for first showing me how to do this).
POP3 is a simple plain-text protocol - you send commands like "user <name>" and "pass <password>" and "stat" to log in and get a list of messages, then "top <message number> <lines>" to read back a given number of lines from the chosen message, and "dele <message number>" to delete it. The protocol is documented here: http://www.faqs.org/rfcs/rfc1939.html
The script can writes messages about the header of each mail and the reason it is considered spam. If you want to make it check but not delete anything you can comment out all the lines:
writeln(net,'DELE '||message)
by changing them to
/* writeln(net,'DELE '||message) */
The comment brackets in ARexx are /* */ just as in C.
You can then run the modified script and will see a list of the things that would have been deleted, but they'll all still be on the server after the test.
SpamFryer uses whatever rules it is programmed to apply by the configuration files. Be wary of checks that might find patterns in legitimate mails - the test for CIALIS looks for a space after that name so it doesn't mistakenly fry mails with the word 'specialist' in them, for example.
To be specific, if the mails you don't want all expect you to reply to an address dummy@nowhere.nul, you can sift them out like this:
if left(instring,10)='Reply-To: ' then do if pos('dummy@nowhere.nul',instring) >0 then spam=1 end
left(var,num) returns the first (leftmost) num characters of variable var, so in this case it checks if the line in the variable instring starts with the 10 characters 'Reply-to: ' and only performs the code between 'do' and 'end' if they match. If you didn't care about the case of the characters in this address you'd be far better off adding a REPLY-TO: rule to SpamFryer.loseList, but this example shows how to add a test that requires an ExAcT match.
The prewritten tests use the ARexx POS function rather than exact matches, so a check for .HTM would match .HTML - and POS("HTM"),inCaps) would match HTM, _Html, .html, hTm and so on - including Weightman which you might not want...
<- Customising SpamFryer for your messages
Contents | Index | < Browse | Browse >
Ad hoc code can be used to add more complex tests than the keep and lose lists will allow. For instance, messages with the text "created by women" in the first few lines are suspect, because a substantial proportion of junk mails that slipped through the net early in 2004 were from dating sites which use this phrase as a come-on. Since all the mails that triggered this also contained the word ' dating ' (with the spaces, so post-dating and carbon-dating would not match) I set the variable 'dating' to zero before starting to scan each mail header, and test as follows:
if pos(' DATING ',inCaps) >0 then dating = dating + 1; if pos('CREATED BY WOMEN',inCaps) >0 then dating = dating + 1;
Then after scanning each mail, I check for at least two such references, and only zap the mail if the above test triggered more than once, as follows:
if dating>1 then spam=1
You can add more variables for other cases, adjusting the threshold and the number added for each match, to create conditional rules that are triggered by a combination of indicators, rather than just one - although in many cases one indicator is sure enough.
<- Customising SpamFryer for your messages
Contents | Index | < Browse | Browse >
Check these things first before you email me for help:
(0) Remember to put your correct details into the configuration block at the start of the script.
See the Configuration page to find out how to do this. If you get it wrong you'll get a message like this:
Error from server: -ERR authorization failed Command returned 10or maybe
+++ Error 18 in line 197: Invalid argument to function Command returned 10/18: Invalid argument
That message suggests that your setting of Account.1.1 in SpamFryer.rexx, or HostName: in SpamFryer.accounts, is incorrect, as the first line that attempts to talk to the mail server is failing. You may be using the name of your ISP (e.g. demon.co.uk) rather than their POP3 mail server (e.g. mail.demon.net, or pop3.demon.co.uk).
The line number may vary if the script has been edited, so check point (2) in this list too if you get Error 18.
(1) Make sure you're running ARexx.
The shell command 'rexxmast' should ensure this.
If it gives an error message, fix that or no
ARexx scripts can work on your system.
(2) Make sure you're online - if you get:
+++ Error 18 in line 180: Invalid argument to function Command returned 10/18: Invalid argument to functionor maybe
Login [userName]: -ERR not found.
you're not connected to the Internet, or maybe not even running a TCP/IP stack! SpamFryer can't work under such conditions. Check you can connect to your mail server in the usual way, then try to run SpamFryer again while you are still online.
The line number may vary if the script has been edited, so also check point (0) in this list if you get an Error 18.
(3) Configuration error messages
SpamFryer attempts to catch all circumstances when errors in the configuration stop it working properly. Even assuming you have the basics needed to run the script you may still get messages if there are errors in the text files that customise SpamFryer for your system. By default these files are named SpamFryer.accounts, SpamFryer.keepList and SpamFryer.loseList.
This section lists the error messages and explains their meanings. Items in brackets vary depending on the error and other aspects of your configuration.
ERROR: Unrecognised item [word] in [accountFile]
The first word of a line in the SpamFryer.accounts file must be // (for a comment, ignored by the program) or 'HOSTNAME:', 'USERNAME:' or 'PASSWORD:' followed by the appropriate information for your mail server. If you get this message it indicates that the file (accountFile) (in the same directory as SpamFryer.rexx unless another path is specified) contains a line which starts with (word) rather than any of the expected terms.
ERROR: No user name configured for host [hostName] in [AccountFile]
The file (accountFile) specifies a host (hostName) with no associated user name for logging in to that host. To fix this, add a USERNAME: line with the required name.
ERROR: No host configured for host [userName] in [AccountFile]
The file (accountFile) specifies an account (userName) but contains no line indicating the server for that account. To fix this, add a HOSTNAME: line with the required address.
ERROR: Duplicate passwords [word1] and [word2] in [AccountFile]
Two passwords were specified in the file without an intervening host and user name. SpamFryer will attempt to use the second one. To fix this, remove the redundant line starting with PASSWORD:
ERROR: Password [word] in [AccountFile] needs a host and user name
You have specified a password but no corresponding host or user name. To fix this, add a HOSTNAME: followed by the matching server address and a USERNAME: line to indicate the relevant account name.
ERROR: Rule [rule] in [SpamFryer config file] skipped
The indicated rule was not understood. Search the config file shown for the rule indicated in the message and make sure that it starts with one of the correct terms ('TO:', 'FROM:', 'REPLY-TO:', 'SUBJECT:' or 'EARLY-REF:') to identify a section of the mail to be checked, followed by a space and some pattern for SpamFryer to look for in that context, on the same line.
If the message refers to 'embedded rules' rather than one of the keepList or loseList configuration files, the problem is in the indicated rule which is embedded in the SpamFryer.rexx script (or whatever you may have renamed it to) rather than in a list file. Search for the rule text (which will be in inverted commas, prefixed by CALL keep_rule or CALL lose_rule depending on the rule category) and ensure that the quoted text follows the syntax indicated above.
(4) UAE and Amiga Forever
I use real Amigas, unless I'm writing for people without access to the genuine article. But the monolithic and unsafe Internet software on most computers has led some people to run AmigaOS via the excellent UAE emulator. SpamFryer works with UAE, but requires some extra software to make the host's TCP/IP stack accessible to scripts.
Joe Cosby posted the following question on an Amiga message board:
SpamFryer tries to read from a device TCP:, which I guess must be created by the most common Amiga TCP/IP programs. This doesn't work under UAE though, there is no TCP: device. I am not really sure how UAE handles TCP/IP at all, actually [but] it is working for me, I browse the web and download email using AWeb and YAM. Does anybody know what, if anything, the equivalent in UAE is, or if there is a workaround for this?
Thomas Rapp helpfully replied:
You need the TCP-Handler program and the TCP mountlist. Get either AmiTCP or Genesis from Aminet, they should both contain the needed files. Don't install the archives, only copy the needed files to your System partition.
The UAE-Support directory contains files which should be added to the corresponding directories on the SYS: drive of a UAE setup, to enable the TCP: device which is used by SpamFryer and many other fine AmigaOS networking programs.
Both files come from the freely-distributable AmiTCP stack, release 3.0.2b. Please direct all questions about using UAE to emulator support sites, rather than to the authors of SpamFryer. If all else fails, get a real Amiga. ;-)
(5) Non-Latin character matching
If you find rules containing accented characters seem to be ignored, or accented characters are matching when you would not expect them to, check the setting of fuzzyMatch at the start of the ARexx script. If you want accented characters to be treated as distinct from unaccented ones, fuzzyMatch must be set to 0.
(6) No spam for users called 'amiga' detected
If the word 'amiga' appears in your name or email address you must remove the rule Early-Ref: Amiga from SpamFryer.keepList or the presence of that word in the mail header will prevent any spams from being fried. In general it's a good way to sift out and protect mails from friends - spammers don't use Amiga software or refer to Amigas, but friends often might!
(7) What does 'Server closed the connection' mean?
This message may appear in interactive ('BROWSE') mode. A POP server shuts down the connection if it detects no activity from the user's client after some time. If you start SpamFryer with the BROWSE argument, then go away for a cup of coffee, it's normal to find the connection closed upon your return.
The timeout delay varies from one server to another, but if you are actively browsing your mailbox, SpamFryer keeps talking to the server until you accept or cancel the mail selection. If you think you are getting this message too often, try lowering the 'pollRate' value to 10 seconds.
(8) How can I make SpamFryer faster?
Hey, remember this is an ARexx script. Don't expect the same performance as with a true machine-code program. That said, SpamFryer may be limited by the connection, not the ARexx.
However, it's true that the more rules you have in the keep or lose lists (typically the loseList is much bigger than the keepList), and the more messages you have waiting in your mailbox, the longer SpamFryer will take to do its job.
You can improve the performance a bit by enabling the sort option (see sortMethod in the "CONFIGURATION BLOCK" section of the script) for the keepList and the loseList (set 'sortMethod' to 1 or 2).
By doing this, SpamFryer will try to keep the rules in the order they are most efficient, by putting the most often triggered rules first in the lists.
If after some time you see that your loseList contains rules which were never triggered (eg.: the "// Trigger Hits:" line before those rules has 0 as value), you can decide to remove those rules. SpamFryer will then have fewer rules to compare mails with, and thus will run a bit faster.
Also, don't neglect the keepList. If you know how to recognise the mails you want to keep, add a rule to the keepList. A mail matching a keep rule will not be tested against all the loseList rules.
As technology improves to make our lives more comfortable, so do spammers' techniques in attempt to bypass spam-filters. So you will have to keep an eye on your keepList and loseList to keep them up-to-date, adding rules for new types of spam. Please put them on the Wiki page so can all benefit!
The 'sortMethod' option can help you here, by showing which rules are the most "efficient". But even with the sorting option enabled, SpamFryer has no way to self-learn from the newer spams. Let's take an idiotic example:
Today, you receive about 12 mails per day. Say Rule #1 in your loseList catches 10 spam from them, and Rule #63 catches another one. In a year's time, the spams caught by Rule #1 are not as pervasive as today, but the spams caught by Rule #63 is now more aggressive. Because Rule #1 has reached a "Trigger Hits" count of 3520 and Rule #63 has only 352 "Hits", it will take a lot of time before SpamFryer learns that Rule #63 is the most efficient at the moment.
Fortunately, SpamFryer 6 includes a technique that can help in this case:
When a rule has been triggered at least 250 times, all the rules "Trigger" counters are scaled down by 2. The orders of the rules on the keepList and loseList are preserved, but it will take less time for Rule #63 to replace Rule #1.
Of course, if every second counts, it's still up to you to re-organise the list to the make the best of each moment. ;-)
(9) I moved SpamFryer.rexx to another place on my hard disk, and since then, SpamFryer doesn't work anymore.
It's better to move the whole SpamFryer directory to ensure you have not left a required file behind in the course of the move.
Check in the CONFIGURATION BLOCK of the SpamFryer.rexx script to see if the path to the keepList, the loseList and the accountFile is still correct.
Check also if the path + filename to the script is not longer than 63 characters. If it is, try to find another location on your drive, with a shorter total path length.
Where you have to provide or configure a filename, check that you've used double-quotes around names with spaces in. Without the quotes some programs won't be able to find that file, if there is a space anywhere in the path or filename.
Contents | Index | < Browse | Browse >
Thanks to the following people for help and feedback since the first release of SpamFryer. Without them, there would be no updates of SpamFryer on Aminet:
Julian Aronowitz Keith Himmelreich Niels Bache Paul Juhusz Olaf Barthel Markus Lunk Fabrizio Bartoloni Mike McCool Adolf Brunner Vince Morra Scott Campbell Harry Patterson Peter Corlett Thomas Rapp Joe Cosby Allan Rasmussen Philippe Duchenne Brian Savage Herve Dupont Krister Skrtic Philippe Ferrucci John Smith John Giudice Stellan Simon N Goodwin Ivan Teskera Bernd Gollesch Thomas Wurgler Kelvin Graham Chris Young
If you'd like to see you name added to this list, send some feedback to the author Simon N Goodwin. The more interest is expressed in SpamFryer, the more likely it is that it will continue to be maintained and updated.
This version of the SpamFryer documentation was produced automatically from the AmigaGuide version using Guide2ML, and validated with the following programs from Aminet:
AmigaGuide Writer (AGW) by D.L McPaul (to get started)Contents | Index | < Browse | Browse >
This is the seventh release of SpamFryer and better than ever thanks to feedback from users - most notably Allan Rasmussen.
Version 2:1 rejected more spam, worked with more POP3 servers, generated better reports, ran faster, supported multiple POP3 accounts, and had pretty new icons.
Version 3:7 added the ability to specify rules via 'keep' and 'lose' lists set in the script or from configuration files. It can also read account details from a separate file, and accept passwords interactively if you prefer not to store them in plain-text on your system. It now works with public ARexx ports, comes with over 100 rules to sift out good and bad mails, provides even better reports, and has updated and improved documentation, as you can probably tell by now. :-)
Version 3:9 added a French translation of the script, two new rules, and further improvements to the documentation.
Version 4:5 typically ran three or four times faster as it
searches the entire early part of the mail for patterns,
rather than a line at a time, and avoids redundant checks
if a mail matches the keepList or any rule in the LoseList.
It also added fresh rules, with some redundant ones removed,
and has even better reporting and documentation.
Version 5:3 supported fuzzy matching and can identify common spams from their MIME header alone, and includes other minor bugfixes, extensions and optimisations, mostly associated with the exact way rules and email headers are interpreted. Details are in the ARexx source and the code history page. The MIME and fuzzy matching routines are explained in the Configuration section.
Version 6:18 of SpamFryer adds 'SpamBrowser' which gives you the option to run SpamFryer in an interactive mode, where you can preview the contents of your mailbox and select the mails you want to keep or delete. Passwords may now be encrypted according to the APOP standard. There's a new version for users of the Thor mailer. Localisation support is improved, as is the handling of control characters and long lines (to Internet standard RFC2822).
New rules and tips for SpamFryer users are updated between releases at http://www.amigawiki.com/cgi-bin/wiki.pl/SpamFryer
Contents | Index | < Browse | Browse >
Modified November 2003 to remove personal information First Aminet upload Modified December 2003 to explicitly QUIT and CLOSE (thanks to Chris Young) Second Aminet upload (Version 1:5) Modified January 2004 Verbose argument and multi-account system (by Elwood) Bug fix: multiply-identified spams only get fried once Reports errors and actions to log as well as console Configurable truncation of long console subject lines New check for 'Test, Yep.' Bagle/Beagle worm signature also purges mails 'To: undisclosed-recipients' (SNG) Combines reports for all accounts (Elwood's correction) Junks all mails with the word VIAGRA inside (Stellan) Junks all mails from excite.com (add more?) (Stellan) Skips quickly through headers once spam is identified Explicitly opens the console window - thanks Kelvin! Closes log only after last account (fixes Elwood's fix) POP3 commands end with explicit CR as well as LF (RFC) Junks references to V1agra, Xanax and 'Cialis ' drugs Combination tests sift out unwanted date invitations Third Aminet Upload (version 2:1) Modified February 2004 Major updates, most of them made by Allan Rasmussen. Added possibility to use a 'keep' and a 'lose' list. Fixed a minor bug which prevented the Reply-To: field from being scanned properly. I couldn't figure out what the variable 'both' was all about, so I got rid of it and invented a simpler way to produce the final output. Removed the 'bogus address' option, as it is redundant. Made it possible to keep a growing logfile. Made the script compatible with public ARexx ports. Fixed a bug which prevented the second account from being scanned if the first was empty. Modified March/April Initial keep/lose list update made by Allan Rasmussen. 2004 Now verbose also works on the log file. Made an offline version of the script for YAM. Verbose argument changed to quiet in the offline YAM version. Simon: Junks mails with the word 'searchingforasoulmate' inside. Allan: Added Early-Ref: argument to the lists. Renamed lists. Simon: removed checks for old worm subjects, added John Smith's patch for optional interactive password entry, added headings for Elwood's multiple account reporting. Symbolic names for constant list indices for readability. Case-independent list pattern handling. No categorisation. Modified May 2004 Call to learn rules. Allow keep and lose list information to be embedded in this program file. Allow comments in list files - such lines must start with "// " (including the space, which the ARexx parser treats as a delimiter). Many new rules. Modified June 2004 New code to check and report rule effectiveness. Code cleanup, consistent nesting and capitalisation. Allows account details to be read from a file. Modified July 2004 Option to kill all mail to addresses with digits in. Now reads other accounts even if first has no mail. Parsing of account file is more robust and error-trapped. Colons after first field in lists are now optional. Quoted pattern support allows leading or trailing spaces. Fourth Aminet Upload (Version 3:7) Modified August 2004 Always reports if a list file is not found if that list's name in the script is not null. Rejects mails with both no subject and no 'from' or 'reply-to' address (Elwood). New rule: rejects mails with 'prescription' in subject. French translation of ARexx added to package (Herve). Clarified integration in SpamFryer.yam.readme (Allan). Missing History file added to the archive. Clarified distribution and translation rules (Simon). Fifth Aminet Upload (Version 3:9) Data changes Sep..Nov New rules which get rid of Hotmail misdirected spam bounces and 75K Windows worms doing the rounds, and to catch other less common but still irritating spams. Code changes Dec 2004 Verbose report details the first rule matched by a mail. Checks for early-ref rules in whole mail, not each line. Skips further loseList rule tests after any one matches. Checks rather than scans server response to TOP command. Much faster scanning due to concatenation of the entire header into one line instead of checking in every line. Timer to report rate of mail scanning. Allan Rasmussen's optimisation tips and SpamFryer.yam updates integrated. Code changes Feb 2005 Fix for refactored no-subject and no-from-address case. Added SpamFrier to variant names for internal keepList and Allan's tweaks to improve reporting and rule parsing. French translation of the script updated - thanks Herve! Data changes Jan..Feb Several new rules, and some redundant old ones removed. Documentation checked and overhauled. Sixth Aminet upload (Version 4:5) February 2005 Data changes 2005/06 Some new rules, and some costly obsolete ones trimmed. Code changes 2005/06 fuzzyMatch configuration option Allows matching of patterns agains spam using accented characters to conceal literal matches, e.g. "VïÅgRà" for "VIAGRA". fuzzyMatch = 0 Recognise only exact matches. fuzzyMatch = 1 Consider accented characters (and "£×ß" symbols) equivalent to Latin letters in Early-Ref checks. Beware! this means that accented characters in Early-Ref patterns never match the corresponding accents in email (as those are translated into ASCII characters). However if you use patterns without accents those will match similar sequences with or without them in spams. MIME checks consistently report "suspicious MIME header". Description of emptyMIME in the "CONF BLOCK" changed. Timer code optimised. Removed redundant "datematch = 0". Tabs replaced with spaces. Window now headed SpamFryer5. Only responds to first From:, To: Subject or Reply-to: line encountered in top of message. Stores only the first Date: line seen. This copes with replies that cite the header of the mail being replied to, without indenting that citation. Otherwise spam checks are performed on those header lines as well as those of the new message, if the cited header lines appear in the top part of the message checked for give-away signs of spam. Optionally identifies mails with an extra blank line after an initial MIME multipart/alternative section as spam. Interactive password entry working again (>> expunged). Redundant template line removed: CALL addlib(LibName, Priority, Offset, Version) Allows quotes of either type - enclosed in the other type - in patterns, e.g. Early-Ref: '"ALLAN.TNT"' which could search for "ALLAN.TNT" inclusive of the quotation marks. Tests for initial parts (reply-to, subject etc) are now cASe inSenSItivE Allows Control C abort (leaving mails on the current server) Moved rule error messages to the 'learn_rule' subroutine, No longer writes passwords to the log file in verbose mode unless there is a password-related error in the accounts file. Seventh Aminet upload (Version 5:3), October 2006 Code changes 2006/08 Option to start SpamFryer in interactive mode: Preview the content of your mailbox, and select the mails you want to keep or delete. Custom screen and font allowed. Option to sort the rules to make checking more efficient: SpamFryer will re-build the keepList and loseList so that the most often triggered rules are the first ones used to check the mails. This also gives some stats about spam. Standard localisation support and new translation tool. Support for APOP (POP3 protocol) password encryption. Support for Thor mailer via customised version of SpamFryer. Folded mail headers are now fully supported to RFC 822. Option to fry mails above a specified size. No dependency upon external programs or libraries. Account selection shortcuts (in interactive mode). Option to over-write or append to previous logs. DOpus Magellan integration examples. Eighth Aminet upload (version 6:18), September 2008
The above list of changes made for version 6:18 is just a condensed summary, based on the contents of the WhatsNew.text file for recent upgraders. A blow-by-blow description of the changes made, mostly by AmigaPhil, in the course of development of SpamFryer 6 appears in the file ChangeLog6_18.text.
Most of the changes in versions 5 and 6 are due to the persistence of Allan Rasmussen and Philippe Duchenne, aka AmigaPhil, respectively - and Simon Goodwin's willingness to make time to test and integrate their suggestions.
Further updates depend upon the amount of feedback we get from users. If we don't hear from you, whether it's because you're entirely happy or have given up, you should not expect another version of this software. But Simon and Allan are always keen to hear from users - it makes our effort worthwhile.
For the latest rules, and support between releases, see: http://www.amigawiki.com/cgi-bin/wiki.pl/SpamFryer
Contents | Index | < Browse | Browse >
SpamFryer is copyright 2004-2008 Simon N Goodwin and individual contributors. It may be freely distributed, without change, as long as all of the original files in the current Aminet release are included in the distribution.
Nobody is allowed to distribute cut-down, edited, extended or otherwise modified packages based on SpamFryer, without the explicit permission of the author Simon N Goodwin.
This rule has been made explicit because otherwise there is a big risk that changes might be made or rules could be added which substantially increase the chance that mails users want to keep are inadvertently deleted. Whilst this risk cannot be avoided in any automatic program, a great deal of effort has gone into developing, testing and documenting SpamFryer with a view to minimising such risk.
Suggestions for changes and extensions are very welcome. In general these will be incorporated into the full release on Aminet.
Contents | Index | < Browse | Browse >
I welcome translations of the script and ideally all the documentation into any language, to make SpamFryer more useful to those who are not accustomed to English. But I am concerned that such changes, unless carefully done and tested, might make the script less safe for users. - Simon N Goodwin.
A translation may not be accepted unless the author is prepared to maintain it for future releases of SpamFryer.
The preferred format for the translation is in the form of a list of unique English and translated patterns which can be automatically searched for and replaced, using free Amiga software, so that fresh translated versions of the script can be produced from the original English version with minimal need for manual intervention. Translators should be prepared to help in the updating of SpamFryer so that their work is not made obsolete by developments.
Please study the contents of the Translation Tool drawer to find out how to add a new translation quite painlessly.
Contents | Index | < Browse | Browse >
I hope you like this package, and welcome emails about it as long as they are constructive and in English, and not entirely in HTML (or they will be fried as spam ;-).
The main updates needed in future are new rules to cope with the latest spams. As making an Aminet update is a big job (even if Aminet is working to distribute it) I encourage users to check the following web page for new rules between updates of the code and documentation on Aminet:
http://www.amigawiki.com/cgi-bin/wiki.pl/SpamFryer
Better still, if you find any effective new rules, please add them to the page, which is designed for anyone to edit. That way we all benefit from the customisations discovered by other keen users of SpamFryer.
My email address is simon (at) mooli (dot) org (dot) uk
Simon N Goodwin, Warwick UK, September 2008 Former Commodore Amiga hardware developer #2202 Former Commodore Amiga Certified Software developer Former Motorola 68060 processor Alpha/Beta tester Former columnist, Amiga Format, Amiga User International Former Amiga Inc Audio development team leader Continuing Amiga A4000/060 user and developer :-)
SPAMFRYER HTML CROSS-REFERENCE INDEX Keyword or phrase Theme of the related section A A500 What you must have in order to use SpamFryer abort Detailed history abuse Customisation - Safety first Account.1.1 FAQ, problems and solutions accountFile FAQ, problems and solutions accountFile Configuring SpamFryer to suit your server accounts Configuring SpamFryer to suit your server AGuideCheck The people who made SpamFryer possible AGW The people who made SpamFryer possible all-HTML Customisation - Background Allan Rasmussen The people who made SpamFryer possible Amiga What you must have in order to use SpamFryer Amiga Forever FAQ, problems and solutions AmigaGuide The people who made SpamFryer possible Aminet The people who made SpamFryer possible AmiTCP FAQ, problems and solutions AmiTCP How to use SpamFryer Amithlon What you must have in order to use SpamFryer AOL Configuration - noDigits ARexx Configuration - emptyMIME ARexx How to use SpamFryer ARexx What you must have in order to use SpamFryer argument Usage - Command line arguments attachments Configuration - emptyMIME authorization FAQ, problems and solutions Authors The people who made SpamFryer possible Automatic rule optimisation Configuration - sortMethod AWeb FAQ, problems and solutions B Background Customisation - Background Bagle/Beagle Detailed history bogus How to use SpamFryer BROWSE Usage - Command line arguments C capitalisation Detailed history Case-independent Detailed history category Customisation - keepList, loseList cc'd Configuration - noDigits Changes between released versions Changes between released versions ChkMail Customisation - Editing the source CIALIS Customisation - Editing the source Colons Customisation - keepList, loseList Combination Customising SpamFryer for your messages Combination tests Customisation - Combination tests command How to use SpamFryer Command line arguments Usage - Command line arguments comments Configuring SpamFryer to suit your server compatible What you must have in order to use SpamFryer conditional Customisation - Combination tests confidential Configuring SpamFryer to suit your server Configuration Configuring SpamFryer to suit your server Configuring to suit your environment Integration and support Configuring to suit your server Configuring SpamFryer to suit your server Configuring to suit yourself Configuring SpamFryer to suit yourself connect FAQ, problems and solutions console Option for the log file console How to use SpamFryer Contact How to contact the author Contents An overview of SpamFryer CR Detailed history Customisation How to use SpamFryer Customising for your environment Customising SpamFryer for your environment Customising for your messages Customising SpamFryer for your messages D date Changes between released versions dating Customisation - Combination tests dele Customisation - Editing the source Detailed history Detailed history Devpac The people who made SpamFryer possible digits Configuration - noDigits directories FAQ, problems and solutions Directory Opus Magellan Integration and support - DOpus Magellan integration disable Customisation - General rules about rules Distribution Copying and modification rights drives Configuring SpamFryer to suit your server drugs Detailed history Duplicate FAQ, problems and solutions E EARLY-REF FAQ, problems and solutions Early-Ref Customisation - keepList, loseList editing Customising SpamFryer for your messages Editing the source Customisation - Editing the source embedding Configuring SpamFryer to suit your server emptyMIME Configuration - emptyMIME emulator FAQ, problems and solutions ERR FAQ, problems and solutions excite.com Detailed history exclude Customisation - Editing the source F FAQ, problems and solutions FAQ, problems and solutions feedback The people who made SpamFryer possible filter Customisation - Background fixes Detailed history flag Configuration - noDigits flag Configuration - emptyMIME flag How to use SpamFryer flag Usage - The interactive (BROWSE) mode freely-distributable FAQ, problems and solutions Freeserve Configuration - noDigits Frequently asked questions FAQ, problems and solutions future How to contact the author fuzzyMatch Configuration - fuzzyMatch G General rules about rules Customisation - General rules about rules Genesis FAQ, problems and solutions Genesis How to use SpamFryer Genesis What you must have in order to use SpamFryer Goodwin, Simon How to contact the author GuideCheck The people who made SpamFryer possible Guido The people who made SpamFryer possible H headings Configuration - Option for the log file Hisoft The people who made SpamFryer possible history Changes between released versions HostName FAQ, problems and solutions Hostname Configuring SpamFryer to suit your server How can I make SpamFryer faster? FAQ, problems and solutions How to contact the author How to contact the author How to use SpamFryer How to use SpamFryer HTML-only Customisation - keepList, loseList I icon start How to use SpamFryer icon choices Integration and support - Alternative icons improvements Changes between released versions IndexAG The people who made SpamFryer possible infection Customisation - Background install FAQ, problems and solutions integration Integration and support interactive mode Usage - The interactive (BROWSE) mode interactive passwords Configuring SpamFryer to suit your server ISP_Account_Name Configuring SpamFryer to suit your server K keepList Customisation - keepList, loseList keepList Configuring SpamFryer to suit yourself keepList Configuration - sortMethod Kickstart What you must have in order to use SpamFryer L LANGUAGE Usage - Command line arguments lastLogOnly Option for the log file Latin-based Configuration - fuzzyMatch Launching How to use SpamFryer LF Detailed history LHA Customisation - Background limits Configuration - lineMax, maxLineSize, maxMailSize lineMax Configuration - lineMax, maxLineSize, maxMailSize lists FAQ, problems and solutions lists Customisation - keepList, loseList lists Configuration - sortMethod localisation Translations of SpamFryer to other languages localisation Detailed history localisation Usage - Command line arguments location Configuring SpamFryer to suit yourself location FAQ, problems and solutions Log options Option for the log file logFile Option for the log file logfile Detailed history logging Configuring SpamFryer to suit your server login Configuring SpamFryer to suit your server logs Option for the log file loseList Configuring SpamFryer to suit your yourself loseList Configuration - sortMethod loseList Customisation - keepList, loseList LZX Customisation - Background M Mail limits Configuration - lineMax, maxLineSize, maxMailSize matches Configuration - fuzzyMatch matches Customisation - keepList, loseList matching FAQ, problems and solutions maxLineSize Configuration - lineMax, maxLineSize, maxMailSize maxMailSize Configuration - lineMax, maxLineSize, maxMailSize message number Customisation - Editing the source Miami Integration and support - Miami integration Miami What you must have in order to use SpamFryer MiamiOnline.yam Integration and support - Yam support MIME Configuration - emptyMIME mountlist FAQ, problems and solutions multi-account Detailed history N name and password Configuring SpamFryer to suit your server New rules Customisation - keepList, loseList noDigits Configuration - noDigits Numeric addresses Configuration - noDigits O obscured Configuring SpamFryer to suit your server on-line Customisation - Editing the source optional Customisation - keepList, loseList optional How to use SpamFryer Optional Configuring SpamFryer to suit yourself optional Configuring SpamFryer to suit your server order Configuring SpamFryer to suit your server order Configuration - sortMethod over-ridden Customisation - keepList, loseList P parameter Usage - Command line arguments parameter Configuring SpamFryer to suit yourself password FAQ, problems and solutions password Customisation - Editing the source path FAQ, problems and solutions path How to use SpamFryer path Configuring SpamFryer to suit yourself path Integration and support - Miami integration path Integration and support - DOpus Magellan integration pattern FAQ, problems and solutions pattern Customisation - keepList, loseList patterns Customisation - Background penis Customisation - Safety first PIF How to use SpamFryer plain-text Customising SpamFryer for your messages pollRate FAQ, problems and solutions POP3 FAQ, problems and solutions POP3 Customisation - Editing the source POP3 How to use SpamFryer POP3 Configuring SpamFryer to suit yourself POP3 Configuring SpamFryer to suit your server POP3 What you must have in order to use SpamFryer postmaster Customisation - Safety first pre-programmed Integration and support printout An overview of SpamFryer problem FAQ, problems and solutions problem How to use SpamFryer protocol Customisation - Editing the source Q quicker Customisation - keepList, loseList quotes Customisation - General rules about rules R Rasmussen, Allan The people who made SpamFryer possible recipient How to use SpamFryer redirected Customisation - Background references Customisation - keepList, loseList references Customisation - Safety first references Customisation - Combination tests REPLY-TO FAQ, problems and solutions Reply-To Customisation - Editing the source report Customising SpamFryer for your messages report Customisation - General rules about rules reports Option for the log file reqtools.library Configuring SpamFryer to suit your server requester Configuring SpamFryer to suit your server requester Usage - The interactive (BROWSE) mode Requirements What you must have in order to use SpamFryer rexxmast FAQ, problems and solutions RexxMast How to use SpamFryer rexxreqtools.library Configuring SpamFryer to suit your server RFC Configuration - lineMax, maxLineSize, maxMailSize RFC Customisation - Safety first RFC Changes between released versions Rights Copying and modification rights Roadshow How to use SpamFryer Roadshow What you must have in order to use SpamFryer rules Configuration - sortMethod rules Customising SpamFryer for your messages rx How to use SpamFryer rx Integration and support - Miami integration rx Integration and support - DOpus Magellan integration rx Option for the log file S Screen selection Configuration - screenName screenName Configuration - screenName scripts Integration and support - Yam support scripts FAQ, problems and solutions Search FAQ, problems and solutions security Customisation - Safety first security Configuring SpamFryer to suit your server servers Customisation - Safety first servers Configuration - noDigits servers Configuring SpamFryer to suit your server Simon N Goodwin How to contact the author SORT Usage - Command line arguments sortMethod Configuration - sortMethod source Customisation - Editing the source Spam Browsing Usage - The interactive (BROWSE) mode SpamAssassin Customisation - Background SpamAssassin How to use SpamFryer SpamBrowser Usage - The interactive (BROWSE) mode SpamFryer lists Customisation - keepList, loseList SpamFryer.accounts FAQ, problems and solutions SpamFryer.accounts Configuring SpamFryer to suit yourself SpamFryer.accounts Configuring SpamFryer to suit your server SpamFryer.keepList FAQ, problems and solutions SpamFryer.keepList Customising SpamFryer for your messages SpamFryer.keepList Customisation - Safety first SpamFryer.keepList How to use SpamFryer SpamFryer.log Configuring SpamFryer to suit yourself SpamFryer.loseList FAQ, problems and solutions SpamFryer.loseList Customising SpamFryer for your messages SpamFryer.loseList Customisation - Safety first SpamFryer.loseList How to use SpamFryer SpamFryer.rexx FAQ, problems and solutions SpamFryer.rexx How to use SpamFryer SpamFryer.rexx Configuring SpamFryer to suit yourself SpamFryer.rexx Configuring SpamFryer to suit your server SpamFryer.thor Integration and support - Thor support SpamFryer.yam Integration and support - Yam support standards Customisation - Safety first subdirectory Integration and support - Alternative icons subject Configuration - lineMax, maxLineSize, maxMailSize Summary An overview of SpamFryer support Integration and support sys:wbstartup How to use SpamFryer T TCP-Handler FAQ, problems and solutions the author How to contact the author the basics What you must have in order to use SpamFryer Thor support Integration and support - Thor support threshold Customisation - Combination tests Translations Non-English versions of SpamFryer Trigger Hits Configuration - sortMethod trojan How to use SpamFryer U UAE FAQ, problems and solutions undisclosed-recipients Detailed history updates Changes between released versions updates The people who made SpamFryer possible Usage How to use SpamFryer userName FAQ, problems and solutions users Configuring SpamFryer to suit your server V variables Customisation - Combination tests VERBOSE Usage - Command line arguments verbose Option for the log file Version Changes between released versions virus How to use SpamFryer W What you must have What you must have in order to use SpamFryer whitelist Customisation - Background window Configuration - screenName window Usage - The interactive (BROWSE) mode Windoze How to use SpamFryer worms How to use SpamFryer Y YAM Integration and support - Yam support YAM Configuring SpamFryer to suit your server