Welcome to the "procmail" mailing list

On Sun, Jul 31, 2005 at 02:15:57PM +0530, Ligesh wrote:

[Please watch your line length. I wrapped your text for you (us)
here using par.]

Post by Ligesh
This question was asked here before, but it did not seem to
result in a solution. If I add a "filter through spamassassin"
rule, procmail stops working and gives a 'no match' on all the
subsequent rules. However, the message is alright, since if I
comment out the spamassassin rule and then pass this message -
The one that had earlier passed through spamassasin - through
procmail, it works fine.
I would like to know what could be the problem.

Well, what does this problem recipe look like? Why didn't you
post it? What do verbose logs say?

My guess based on your entirely scanty information is that you might
have a recipe with an H-flag on the first line, disabling subsequent
recipes that are written to look only at the body.

There is a known bug with regard to the H-flag, and it should be
employed with alternative syntax.

This has been described many times in the searchable list archives
at www.procmail.org (lower down on the page), so I won't repeate
it here, especially seeing as how I have no real idea if that is
your problem because you didn't give enough useful information
to know with more accuracy. The issue is also described well on
Nancy McGough's QuickStart procmail pages, which are in any case
a worthwhile read: http://www.ii.com/internet/robots/procmail/qs/

Dallman

Dallman Ross

2005-07-31 10:29:54 UTC

Post by Dallman Ross
[Please watch your line length. I wrapped your text for you (us)
here using par.]

I suspect you intended to write the list, and not me privately.
Please re-send to the list. Oh, never mind, I will send this to
the list. I have cut some of your text, however.

----------
| spamassassin -a
----------
The log file with verbose=on
--------------------------------
procmail: Assigning "INCLUDERC=.etc/mail/spam-config"
procmail: Couldn't determine implicit lockfile from "spamassassin"

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

While that's not your problem, it is another problem. Why
don't you remove the lockfile, since it's not necessary here
and is only confusing procmail?

procmail: No match on "^X-Spam-Status:.*Yes"

Well, apparently "spamassassin -a" is not creating the X-Spam-Status:
header, so you won't get a match.

procmail: Assigning "INCLUDERC=.etc/mail/proc-ligesh-forums"
procmail: Assigning "INCLUDERC=.etc/mail/spamfilter"
procmail: No match on "^From.*fabmart.com"
procmail: No match on "^From.*fabmall.com"
procmail: No match on "^From.*bazee.com"
procmail: No match on "^From.*(baazee.com|fabmart.com|abgymnics.com|ewrma.com)"
procmail: Assigning "INCLUDERC=.etc/mail/dump.prc"
procmail: No match on "^From.*redhat"
procmail: No match on "^X-Mailer: PhpCollab"
procmail: No match on "^To.*wordorigins"
procmail: No match on "^To.*wordoftheday.*"
procmail: No match on "^(To|Cc|Delivered-To).*lxhelp"
procmail: No match on "^From.*Mail Delivery .*system"
procmail: Locking "mail/mboxes/rest.lock"
procmail: Assigning "LASTFOLDER=mail/mboxes/rest"

What is it about that that you think isn't working?

--------------------------
i will check out the links that you have specified. Btw, which
is the MUA that you are using? Just curious. I am using vim,
which allows me treat a paragraph as a single line.

I am writing this in vim. That is not the problem. I'm not interested
in learning how to write with an entire paragraph as one line: I already
can do so. The point is, it's *rude* to do so. I don't *read* my email
in vim (unless I'm replying). I read my email in mutt (as my mail headers
could have told you).

Sorry, but I have to laugh out loud at the analogy your words here made
me think of. Suppose I had written, "please don't double-park your car
in the middle of the road!" and you had answered, "What kind of a car
do you drive? I drive a BMW, and it lets me steer it into the middle of
the street and park it."

Dallman

Ligesh

2005-07-31 11:33:34 UTC

Post by Dallman Ross
While that's not your problem, it is another problem. Why
don't you remove the lockfile, since it's not necessary here
and is only confusing procmail?

procmail: No match on "^X-Spam-Status:.*Yes"

header, so you won't get a match.

I did try without the '-a' too, but it doesn't seem to work, and no-match at x-spam-status is not the issue. It is giving no match at everywhere, and when the spamassassin rule is hashed out and the same message is sent, it does match 'ligesh.com'.

What is it about that that you think isn't working?

Since you complained, I thought you might be stuck with some tremendously outdated and absurdly ill-conceived softwares like emacs. Mutt word-wraps pretty much neatly, and I never though a mutt user would complain about a single line paragraph.

Post by Dallman Ross
Sorry, but I have to laugh out loud at the analogy your words here made
me think of. Suppose I had written, "please don't double-park your car
in the middle of the road!" and you had answered, "What kind of a car
do you drive? I drive a BMW, and it lets me steer it into the middle of
the street and park it."

What do you mean? A paragraph is ideally a single line. Why do you think it should be arbitrarily split at 75chars? From where does this 75chars come from anyway? Your message will get completely garbled in a pda, or if my terminal has less than 75 chars. On the other hand the single line paragraph will work no matter what's your terminal screen width. So what's your point?

It is the the prerogative of the final display software to wrap the paragraph as it sees fit, and the sender should not make any assumptions regarding this. Yeah, every rule should have some basic logic. If the government comes up with a law stating that no one should ever travel more than 75kmph anywhere, I would normally ask what kind of car does this particular government bureaucrat ride? Is he still stuck up with some 1914 Ford Model T, and have completely failed to realize that world has moved on since then? Normally I would start questioning the sanity of the person who asks me to put newlines at some arbitrary place (75 char is absurdly arbitrary since currently the screen widths range from 40chars to more than 150).

I have integrated mailing list with forum, and messages with hard newlines at 75 chars look really ugly when displayed in a forum, but single line pargaraphs will work everywhere (Except in some insanities like emacs, which no sane person should use anyway.).

Ruud H.G. van Tol

2005-07-31 12:42:50 UTC

Post by Ligesh
What do you mean? A paragraph is ideally a single line. Why do you
think it should be arbitrarily split at 75chars? From where does this
75chars come from anyway?

*ploink*

--
Grtz, Ruud

Michelle Konzack

2005-07-31 09:48:54 UTC

Post by Ligesh
This question was asked here before, but it did not seem to result in
a solution. If I add a "filter through spamassassin" rule, procmail
stops working and gives a 'no match' on all the subsequent rules.
However, the message is alright, since if I comment out the
spamassassin rule and then pass this message - The one that had
earlier passed through spamassasin - through procmail, it works fine.
I would like to know what could be the problem.

Wi do not know anything, because you have not posted
the relevant section of your ~/.procmailrc

How do you have included spamassassin in your procmailrc ?

Post by Ligesh
thanks in advance.

Greetings
Michelle

--
Linux-User #280138 with the Linux Counter, http://counter.li.org/
Michelle Konzack Apt. 917 ICQ #328449886
50, rue de Soultz MSM LinuxMichi
0033/3/88452356 67100 Strasbourg/France IRC #Debian (irc.icq.com)

Ligesh

2005-07-31 11:55:26 UTC

Post by Michelle Konzack

Post by Ligesh
I would like to know what could be the problem.

Wi do not know anything, because you have not posted
the relevant section of your ~/.procmailrc
How do you have included spamassassin in your procmailrc ?

This is the spamassassin rule.

---------
:0fw:
| spamassassin
------------

I had thought that the problem would be some standard, well known bug, since I did not attempt anything fancy with procmail. I already have a working procmail config, and I just added this line in the beggining, and lo, now procmail dumps everything to the 'rest' (the default) folder. I saw another guy had also asked the same question, but he seem to have left when someone from the list asked him for the details.

Thanks.

--
:: Ligesh :: http://ligesh.com

Ligesh

2005-07-31 12:16:43 UTC

Post by Ligesh
I had thought that the problem would be some standard, well known bug, since I did not attempt anything fancy with procmail. I already have a working procmail config, and I just added this line in the beggining, and lo, now procmail dumps everything to the 'rest' (the default) folder. I saw another guy had also asked the same question, but he seem to have left when someone from the list asked him for the details.

Actually I first sent a mail to this particular guy who had asked this question - most likely his no continuing with thread must be because he solved it - before coming to the list. But he hasn't answered me yet. If he replies I will post it here, so that it will turn up in google for the next person does the search.

Thanks.

--
:: Ligesh :: http://ligesh.com

Dallman Ross

2005-07-31 12:32:24 UTC

Post by Ligesh
If he replies I will post it here, so that it will turn up in
google for the next person does the search.

While I have nothing against Google, please note that this list has had
a searchable archive for about the last ten years.
http://www.xray.mpe.mpg.de/mailing-lists/procmail/ , which is
linked from the main page at www.procmail.org.

I have tonsillitis at the moment and don't have the energy to
continue the discussion about list netiquette and email formatting,
and it's rather OT anyway. But maybe someone else will pick up the
gauntlet. However -- and this is intended more as an observation
than a major flame, so please don't get too bent out of shape, but --
I note that you didn't include an attribution for the person you
quoted (Michelle Konzack, but I have elided her text here); and you
also emailed me instead of the list, previously. Not cardinal
sins, and mistakes happen, but all these things in concert lead
me to conclude that your list-posting skills may not be what you
think they are. Nearly endless line-lengths are not the way to
post to text lists. Here's your message I'm replying to as it
appears in the unexpiring "official" archive:

http://www.xray.mpe.mpg.de/mailing-lists/procmail/2005-07/msg00245.html

Do you think that looks all so readable?

Dallman, not really intending to sound angry, but it probably does
nonetheless

Ligesh

2005-07-31 13:35:26 UTC

Post by Ligesh
If he replies I will post it here, so that it will turn up in
google for the next person does the search.

I have tonsillitis at the moment and don't have the energy to
continue the discussion about list netiquette and email formatting,
and it's rather OT anyway. But maybe someone else will pick up the
gauntlet. However -- and this is intended more as an observation
than a major flame, so please don't get too bent out of shape, but --
I note that you didn't include an attribution for the person you
quoted (Michelle Konzack, but I have elided her text here); and you

I think I was quoting myself and not Michelle. Anyway, I just press 'r', and then remove some lines, so ideally mutt should have done the attribution part.

Post by Dallman Ross
also emailed me instead of the list, previously. Not cardinal
sins, and mistakes happen, but all these things in concert lead
me to conclude that your list-posting skills may not be what you
think they are. Nearly endless line-lengths are not the way to
post to text lists. Here's your message I'm replying to as it

I have configured all my mailing lists to munge the 'Reply-To' header and add the mailing list address there, so I always end up pressing just 'r'. I cannot understand the logic behind the other way, since most often the person would want to reply to the mailing list, rather than to the sender. I guess I have a problem with rules and practices which doesn't make logical sense to me.

I have 42 mailboxes (My mailer is actually vim, which calls mutt individually for each mailbox); I have configured both ezmlm, and mailman (In ezmlm, I just forcibly add a 'reply-to: <mailing list>'; mailman has a configuration option itself which will do it); Anyway, I have more than enough experience, though I do get into "replying" debacle in mailing lists that are configured in that fashion.

Post by Dallman Ross
http://www.xray.mpe.mpg.de/mailing-lists/procmail/2005-07/msg00245.html
Do you think that looks all so readable?

Of course, badly written softwares will screw up; but that can't be helped. As a matter of fact, I have a 'set wrapmargin=40' in my muttrc, which will soft wrap the lines at around 80 chars itself. The comfortable reading length is actually 65 chars, but I can't have such a setting because a lot of people arbitrarily put hard newlines at 75 chars. :-)

Anyway, the point is:

- A single line paragraph will work in all circumstances. (You can configure your display software to wrap the line at whatever you wish. And it will work very fine in PDAs, and also will look good in forums.

- 75 char word wrap will screw up entirely in a lot of situations.

To me it trivially makes a formidable case for the former. Couple this with the ease of editing at the sender end, I don't understand why people still get into some religious fervor about the word wrapping issue. 80 char is simply a historical relic from the days of the dumb terminals. It is not only difficult to edit at the sender end, it will screw up the display unless the screen width is more than 75 chars.

Post by Dallman Ross
Dallman, not really intending to sound angry, but it probably does
nonetheless

I am just pointing out what appears to be logical to me. If that is making people angry, I can't help that.

--
:: Ligesh :: http://ligesh.com

Patrick Shanahan

2005-07-31 14:47:56 UTC

Post by Ligesh
I am just pointing out what appears to be logical to me. If that is
making people angry, I can't help that.

Nor will you *receive* much help with your attitude.

--
Patrick Shanahan Registered Linux User #207535
http://wahoo.no-ip.org @ http://counter.li.org
HOG # US1244711 Photo Album: http://wahoo.no-ip.org/gallery

Ruud H.G. van Tol

2005-07-31 14:59:50 UTC

Post by Patrick Shanahan

[]

Nor will you *receive* much help with your attitude.

Right, plenty of other lists to discuss netiquette. Adopt, or be
ignored.

--
Grtz, Ruud

Ligesh

2005-07-31 18:48:16 UTC

Post by Ruud H.G. van Tol

Post by Patrick Shanahan

[]

Nor will you *receive* much help with your attitude.

Right, plenty of other lists to discuss netiquette. Adopt, or be
ignored.

You need to change your stupid mail archiving software that does not wrap lines properly. Your statement can be rephrased as: If you come to an insane asylum, you better start acting mad.

Damian:
I will try to figure out how to automatically configure flow text in vim, and implement it. I think that is the ideal solution.

Thanks.

--
:: Ligesh :: http://ligesh.com

Professional Software Engineering

2005-07-31 20:08:09 UTC

Post by Ligesh
You need to change your stupid mail archiving software that does not
wrap lines properly. Your statement can be rephrased as: If you come to
an insane asylum, you better start acting mad.

See my prior response to you.

What you're saying is that the standards by which everyone else manages to
communicate don't apply to you, and that everyone else should do as YOU want.

This is not a good attitude to present when seeking assistance amongst a
given group - a group quite capable of programmatically ignoring future
posts from you.

---
Sean B. Straw / Professional Software Engineering

Procmail disclaimer: <http://www.professional.org/procmail/disclaimer.html>
Please DO NOT carbon me on list replies. I'll get my copy from the list.

Jonesy

2005-07-31 20:46:11 UTC

Which I have now done.

Jonesy

Ligesh

2005-08-01 00:03:34 UTC

See my prior response to you.
What you're saying is that the standards by which everyone else manages to
communicate don't apply to you, and that everyone else should do as YOU want.

Every standard should have a reason. I am ready to defend my actions, and logically too. People have a habit of just following something because everyone else does it. So should I just blindly follow some idiotic procedure just because the whole world is doing so? And as I have said above, the 75 char lines screws up in a lot of places, especially in forums, where it ruins the whole appearance. I used to use vm (The mailer in Emacs), and I used to post messages to forums using the emacs itslef, and I really hate the way it is displayed there. And anyway, I am not the only one. Quite a lot of people - especially ones using Outlook - use a single line paragraph, and again, my posts used to look ugly when juxtapositioned with their replys. All in all, I find quite a lot of advantages to single line paragraphs, and quite a few disadvantages to the 75 char style. That is all.

--
:: Ligesh :: http://ligesh.com

Dallman Ross

2005-08-01 10:41:01 UTC

Post by Ligesh
And anyway, I am not the only one. Quite a lot of people -
especially ones using Outlook - use a single line paragraph, and

You owe me a new keyboard for that one, dude! :-) I'm on my
late-morning coffee and almost lost it.

Microsoft, the inventor of Outlook, _et al._, is also the
inventor, or at least promulgator, of that incarnate evil,
top-posting. Here you take a company that, at least in its
earlier years, ignored and abused practices and standards
in just about every tentacled reach of its behemoth
operations, and you uphold that breach of good sense as
the new defense for your self-proclaimed "I-don't-follow-the-
crowd" logic. Oh, the irony!

I actually use Outlook (but not Express), but I don't have it
configured to do any of that nonsense.

Dallman

Google Kreme

2005-08-01 11:56:19 UTC

Post by Ruud H.G. van Tol

Post by Patrick Shanahan
Nor will you *receive* much help with your attitude.

Right, plenty of other lists to discuss netiquette. Adopt, or be
ignored.

You need to change your stupid mail archiving software that does
not wrap lines properly.

No, it doesn't work that way. YOU conform to the list standards, or
YOU get ignored.

Post by Ligesh
Your statement can be rephrased as: If you come to an insane
asylum, you better start acting mad.

Strawman.

Post by Ligesh
I will try to figure out how to automatically configure flow
text in vim,

set wm=x

Where x is the width of your terminal -72 or so (pick a number
between 65 and 75).

So, if you have an 80 character wide terminal, wm=8

Or use a mailer that actually marks text as format flowed, as it should:

Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed

for example.

Peter Jones

2005-08-01 15:29:24 UTC

Post by Ligesh
You need to change your stupid mail archiving software that does not
wrap lines properly.

You do realise, of course, that there is a specific reason that the
archive presents mail as it does?

This is a mailing list in which, more often than not, formatted
programming code (ie, procmail recipes) is posted. The formatting of
such code *must* be preserved by the archive -- and the only sure way of
achieving that, when presenting it in HTML, is by wrapping the entire
message in <pre></pre>. This has the desired effect of preserving all
formatting -- including indents, etc, exactly as typed. Any other method
of archiving program-code posts (such as google's method of adding a <br>
in the place of each carriage return) should be considered broken.

Of course, it has the side effect of causing single-line paragraphs to
disappear off the edge of the page.

*shrug*

Pete.

Ligesh

2005-08-01 16:27:00 UTC

Post by Peter Jones

Post by Ligesh
You need to change your stupid mail archiving software that does not
wrap lines properly.

You do realise, of course, that there is a specific reason that the
archive presents mail as it does?
This is a mailing list in which, more often than not, formatted
programming code (ie, procmail recipes) is posted. The formatting of
such code *must* be preserved by the archive -- and the only sure way of
achieving that, when presenting it in HTML, is by wrapping the entire
message in <pre></pre>. This has the desired effect of preserving all
formatting -- including indents, etc, exactly as typed. Any other method
of archiving program-code posts (such as google's method of adding a <br>
in the place of each carriage return) should be considered broken.
Of course, it has the side effect of causing single-line paragraphs to
disappear off the edge of the page.

Ok, so that's the reason for lines not wrapping around. I think some solution is needed. A lot of people I know (professional writers actually) find it more comfortable working with single line paragraphs. In fact, I used to use emacs, and when arguing with them, their mails looked better (and more professional) when juxtapositioned with mine, and that is one of the reasons why I dumped emacs altogether. Of course, single line paragraphs are easier to manage too. Once I get 'par', I will try to configure mutt to trim the message just before sending.

--
:: Ligesh :: http://ligesh.com

Mike Peeler

2005-08-01 16:59:55 UTC

Post by Dallman Ross
Ligesh, you are claiming to apply careful logic; but (to the degree
your logic is not flawed, which point I won't go into right this sec)
you are doing so in a vaccuum, as if no other well-followed practice
had ever evolved and you are free simply to invent your own way,
irrespective of other users that have spent time, effort, and energy
designing their own interactive tools to comply with what's out there
already.

A google search for par is not giving me any meaningful result.
Can you just give me the package name or something.

Just typing the one word par into Google, the fourth hit turned up was
"Par 1.52 - paragraph reformatter", at this URL:

www.nicemice.net/par/

You do need a C compiler. It's written in portable C, so you wouldn't
need Cygwin. There is a Windows makefile on the "hints" page, in case
that might be helpful. (You did mention you use Outlook.)

Mike

Google Kreme

2005-08-05 21:20:25 UTC

Post by Ligesh
A lot of people I know (professional writers actually) find it more
comfortable working with single line paragraphs.

The issue is not single line paragraphs, it is your broken mailer. I
type single line paragraphs, but my mailer wraps them correctly and
tells the recipients that the content is flowed.

Professional Software Engineering

2005-08-05 21:37:27 UTC

Post by Google Kreme

Post by Ligesh
A lot of people I know (professional writers actually) find it more
comfortable working with single line paragraphs.

The issue is not single line paragraphs, it is your broken mailer. I
type single line paragraphs, but my mailer wraps them correctly and
tells the recipients that the content is flowed.

Well, and as I indicated, I use my mailer to submit to a procmail frontend
for my weblog, and I manage to re-flow the paragraphs just fine, even
though the original email message has newlines within paragraphs: thus, the
email as sent is compatible with both basic mailers, and anything that
wants to reformat it.

Beats the turd out of expecting the recipient to have to reformat a message
in order to view it.

I suspect this horse has been beaten to death anyway - we're all preaching
to the choir here on this list.

---
Sean B. Straw / Professional Software Engineering

Procmail disclaimer: <http://www.professional.org/procmail/disclaimer.html>
Please DO NOT carbon me on list replies. I'll get my copy from the list.

Dallman Ross

2005-07-31 15:29:55 UTC

Post by Ligesh
I guess I have a problem with rules and practices which doesn't
make logical sense to me.

What if I decided it didn't make logical sense to me to drive on
the right side of the road in the United States or Western Europe
(outside the UK, which many don't consider "Europe," anyway)? ;->

Sometimes you have to go with the flow.

Anyway, the purpose of the procmail list is not to look pretty in
forums. So formatting should be for the procmail list, and not to
look pretty in forums.

Dallman

Ligesh

2005-07-31 17:16:46 UTC

Post by Ligesh
I guess I have a problem with rules and practices which doesn't
make logical sense to me.

What if I decided it didn't make logical sense to me to drive on
the right side of the road in the United States or Western Europe
(outside the UK, which many don't consider "Europe," anyway)? ;->
Sometimes you have to go with the flow.
Anyway, the purpose of the procmail list is not to look pretty in
forums. So formatting should be for the procmail list, and not to
look pretty in forums.

That reasoning is undestandable. Now I will have to have separate vim configurations for separate mailing lists. I use vim + lynx for posting to forums, and so I will need another set for that. I mean, if you have a car that is custom built to drive on the left side... I think I will just do without driving altogether. :-)

Anyway I think it is a bug in procmail. I saw someone complain in usenet too, but I couldn't find a resolution. Also he claims the configuration works on one machine but not on the other. So there has to be some problem with procmail. I changed spamassassin to 'spamc' and it works fine.

Ok so I think I will just ask one question and go away. I need to sort the spam mails into different mailboxes according to their score..

Rule:

If 4 <= spam-score < 5
mail/mboxes/spam4

If 5 <= spam-score < 6
mail/mboxes/spam5

And so on. That way I can schedule checking these mailboxes accordingly. The larger the value, the less frequently you check it - say once a month, and remove everything. I think something like this is needed, since the spamassassin is not very perfect, and the frigging spammers are now using random statements in their mails to confuse the spam blockers.

Finally the ideal solution for spam is not through software, but through hardware. As in, get a gun a shoot those f*cking bastards.

Thanks.

--
:: Ligesh :: http://ligesh.com

Damian Menscher

2005-07-31 17:16:14 UTC

Post by Ligesh
That reasoning is undestandable. Now I will have to have separate vim configurations for separate mailing lists. I use vim + lynx for posting to forums, and so I will need another set for that. I mean, if you have a car that is custom built to drive on the left side... I think I will just do without driving altogether. :-)

One configuration for flowed text should be sufficient.

Post by Ligesh
Anyway I think it is a bug in procmail. I saw someone complain in usenet too, but I couldn't find a resolution. Also he claims the configuration works on one machine but not on the other. So there has to be some problem with procmail. I changed spamassassin to 'spamc' and it works fine.

I'm suspicious of your -a flag -- it's not in my manpage. Maybe you're
using an older version of SA though.

Post by Ligesh
Ok so I think I will just ask one question and go away. I need to sort the spam mails into different mailboxes according to their score..
If 4 <= spam-score < 5
mail/mboxes/spam4
If 5 <= spam-score < 6
mail/mboxes/spam5
And so on. That way I can schedule checking these mailboxes accordingly. The larger the value, the less frequently you check it - say once a month, and remove everything. I think something like this is needed, since the spamassassin is not very perfect, and the frigging spammers are now using random statements in their mails to confuse the spam blockers.

Count the number of *s.

Post by Ligesh
Finally the ideal solution for spam is not through software, but through hardware. As in, get a gun a shoot those f*cking bastards.

Russian spammer found beaten to death
One billion email users under suspicion as police launch enquiry
http://www.itweek.co.uk/vnunet/news/2140340/russian-spammer-murdered

Damian Menscher

--
-=#| Physics Grad Student & SysAdmin @ U Illinois Urbana-Champaign |#=-
-=#| 488 LLP, 1110 W. Green St, Urbana, IL 61801 Ofc:(217)333-0038 |#=-
-=#| 4602 Beckman, VMIL/MS, Imaging Technology Group:(217)244-3074 |#=-
-=#| <***@uiuc.edu> www.uiuc.edu/~menscher/ Fax:(217)333-9819 |#=-
-=#| The above opinions are not necessarily those of my employers. |#=-

Dallman Ross

2005-07-31 18:44:53 UTC

Post by Ligesh
Anyway I think it is a bug in procmail. I saw someone complain in
usenet too, but I couldn't find a resolution. Also he claims the
configuration works on one machine but not on the other. So there
has to be some problem with procmail. I changed spamassassin to
'spamc' and it works fine.

I very highly doubt it's a bug in procmail.

To know more, I think it would be good to see the headers of the
message you sent through that didn't do what you expected as
per that log extract you sent. Also, what version of SA and
what is the -a flag?

Post by Ligesh
Ok so I think I will just ask one question and go away. I need
to sort the spam mails into different mailboxes according to
their score..

Well, my goal is not to send you away, and that would be too bad on
account of formatting discussions. How hard is it to type on your
keyboard and pound on the Enter key every once in a while? We have
some great, and helpful, and experienced procmail users on this
list.

Post by Ligesh
If 4 <= spam-score < 5 mail/mboxes/spam4
If 5 <= spam-score < 6 mail/mboxes/spam5
And so on. That way I can schedule checking these mailboxes
accordingly. The larger the value, the less frequently you
check it - say once a month, and remove everything. I think
something like this is needed,

Someone just posted to test for asterisks. Aside from that,
there have been extensive SA recipes discussed here in the past,
findable through the archive, which link I gave earlier. I myself
contributed at least one such recipe set with a threshold number,
etc., in the past. I just don't want to reinvent the wheel every
time a new member shows up -- that's what the archives are best
used for.

Post by Ligesh
since the spamassassin is not
very perfect, and the frigging spammers are now using random
statements in their mails to confuse the spam blockers.
Finally the ideal solution for spam is not through software,
but through hardware. As in, get a gun a shoot those f*cking
bastards.

Yeah. Well, I've been coding something called Spam Snaggers for
three months straight, and it's just about ready for a broad beta.
I will open it up to local beta testers on my system this week.
I'm pretty damn happy with how it's behaving for me right now.
I'll announce more about it as it develops.

Dallman

Ligesh

2005-07-31 20:41:14 UTC

Post by Ligesh
Anyway I think it is a bug in procmail. I saw someone complain in
usenet too, but I couldn't find a resolution. Also he claims the
configuration works on one machine but not on the other. So there
has to be some problem with procmail. I changed spamassassin to
'spamc' and it works fine.

I very highly doubt it's a bug in procmail.
To know more, I think it would be good to see the headers of the
message you sent through that didn't do what you expected as
per that log extract you sent. Also, what version of SA and
what is the -a flag?

No it is some form of bug somewhere. I am actually testing this by piping messages to procmail. The funny thing is, if I pipe the messages to 'strace -f procmail', instead of just 'procmail', it works fine, and the mail is delivered properly. Also, when the spamassassin line is removed, the SAME mail - the one which did not get detected when there was the spamassassin rule - gets properly filtered. Maybe I can get the source of procmail, compile, and run it through idevim and find out what's happening.

But seeing the behaviour with strace, I think if I run it through gdb, the bug may not show up. This is one of the most frustrating kind of bugs; ones that disappear in the presence of a debugger. It is most likely a timing issue. The wait() function may not be working properly, and the checks are happening before spamassassin properly returns or something. I do not know the architecture of procmail so I cannot comment more.

Post by Ligesh
Ok so I think I will just ask one question and go away. I need
to sort the spam mails into different mailboxes according to
their score..

It is a bit difficult. I write a lot and the actions sort of come subconsciously. I use the <C-w>, <c-e>, <C-a> f,d,w, e, etc and make a lot of changes to my message before final posting. If you add hard newlines, editing the message becomes difficult since the paragraph indentation will get screwed up when you edit it. This is, in fact, the primary reason I use single line paragraphs. Indentation becomes automatic. You can add/delete full sentences in the midst of paragraphs without needing to bother about the indentation.

I can configure vim to do a forced hard newline wrap to files before it is saved, or I have to try configuring mutt or msmtp to wrap the lines before mails are sent.

thanks.

--
:: Ligesh :: http://ligesh.com

Dallman Ross

2005-07-31 21:36:03 UTC

Post by Dallman Ross
I very highly doubt it's a bug in procmail.
To know more, I think it would be good to see the headers of the
message you sent through that didn't do what you expected as
per that log extract you sent. Also, what version of SA and
what is the -a flag?

No it is some form of bug somewhere. I am actually testing
this by piping messages to procmail. The funny thing is,
if I pipe the messages to 'strace -f procmail', instead of
just 'procmail', it works fine, and the mail is delivered
properly. Also, when the spamassassin line is removed, the SAME
mail - the one which did not get detected when there was the
spamassassin rule - gets properly filtered. Maybe I can get the
source of procmail, compile, and run it through idevim and find
out what's happening.
But seeing the behaviour with strace, I think if I run it
through gdb, the bug may not show up. This is one of the most
frustrating kind of bugs; ones that disappear in the presence
of a debugger. It is most likely a timing issue. The wait()
function may not be working properly, and the checks are
happening before spamassassin properly returns or something. I
do not know the architecture of procmail so I cannot comment
more.

Think about what you're saying: you're saying procmail doesn't
work with spamassassin. Well, tens of thousands of uses run
spamassassin from procmail hundreds or thousands of times a day.
Where are these thousands of frustrated users, then, with their
bug reports? SA has been around for four years now. I've been
on this list for eleven, myself. I'm not the oldest user
around here, by any means, either.

Though I don't use SA much at all, I do have it hanging off the end
of my legacy .procmailrc that still runs off the back of my new one.
It has caught three messages in July that my beta self-rolled
spam trap missed. Anyway, I'm running spamd, not SA proper,
but I used to run SA proper as a last resort, and didn't find
any such bug. But many here rely on SA much more than I do,
anyway.

Post by Dallman Ross
How hard is it to type on your keyboard and pound on the Enter
key every once in a while? We have some great, and helpful,

Whatever. I'm pretty sure I'm a fair bit older than you, and
I can assure you that I'm grounded in my own set ways as much
as anybody is. That's how humans are, and the older they get, the
more like this they become. But -- and I've mentioned this twice
now today, this marking the third time -- get yourself a copy
of the "par" program. I reformatted this entire posting in
one keystroke. When I edited my paragraphs, I reformatted them,
again with one keystroke. That's how good software's supposed
to work.

You can even write your uninterrupted paragraph-lines as you've
been doing, then, when you're done, hit the macro key and invoke
par. You could even have it automatically happen when you save
your edits! Though one does have to be careful when quoting
procmail recipes, etc.

Post by Ligesh
I can configure vim to do a forced hard newline wrap to files
before it is saved, or I have to try configuring mutt or msmtp
to wrap the lines before mails are sent.

Apparently either one is pretty easy, though.

Dallman

Ligesh

2005-07-31 23:58:59 UTC

Post by Ligesh
But seeing the behaviour with strace, I think if I run it
through gdb, the bug may not show up. This is one of the most
frustrating kind of bugs; ones that disappear in the presence
of a debugger. It is most likely a timing issue. The wait()
function may not be working properly, and the checks are
happening before spamassassin properly returns or something. I
do not know the architecture of procmail so I cannot comment
more.

I am not saying that procmail doesn't work with spamassassin, but rather there is a very specific bug which gets triggered on very particular circumstances. I am not the only one. There were 2 other people who complained, but neither of them has later attributed it to config mistakes. And one guy specifically said the config works on another machine, but on one machine somehow this weird behaviour is displayed.

Anyway, here's what's happening in the code. The "themail.p" variable becomes just an "F" once spamasssin is executed. So there must be some problem with spamasssin. But the thing is a mail IS delivered to 'rest' (the default folder), and it even contains the extra spam headers and all.

I tried 'cat test.procmail | spamasssin' and it gave no output. The problem was with spamasssin. But now the real issue is is how did new mail get into the 'default' folder, every time I piped it through procmail.

Anyway, I will do somemore gdbing some time and get back. I have to download and configure 'par' too.

thanks..
bye.

--
:: Ligesh :: http://ligesh.com

Ligesh

2005-07-31 20:45:43 UTC

Btw, I am sorry, I never mentioned my spamassassin version - I am using redhat 8, with spamassasin version 2.31-16. I am testing by piping mails from mutt to procmail, and I have tried quite a lot of permutations - as I had said in my earlier mail, including running it through 'strace -f procmail'.

Thanks.

--
:: Ligesh :: http://ligesh.com

Dallman Ross

2005-07-31 21:37:36 UTC

Post by Ligesh
Btw, I am sorry, I never mentioned my spamassassin version -
I am using redhat 8, with spamassasin version 2.31-16. I am
testing by piping mails from mutt to procmail, and I have tried
quite a lot of permutations - as I had said in my earlier mail,
including running it through 'strace -f procmail'.

Okay, that's pretty damn old.

11:36pm [~/Mail] 947[1]> spamassassin --version
SpamAssassin version 3.0.4
running on Perl version 5.8.0

Still, I used to run SA 2.x from procmail.

Dallman

Jonesy

2005-07-31 22:08:33 UTC

Okay, that's pretty damn old.
11:36pm [~/Mail] 947[1]> spamassassin --version
SpamAssassin version 3.0.4
running on Perl version 5.8.0
Still, I used to run SA 2.x from procmail.

Did you run `spamassassin`? Or, did you run `spamc`? -- _from_ procmail...

Jonesy

Dallman Ross

2005-07-31 22:51:48 UTC

Post by Dallman Ross
Still, I used to run SA 2.x from procmail.

Did you run `spamassassin`? Or, did you run `spamc`? -- _from_
procmail...

Before our system admins installed the daemon, we used to run
spamassassin directly. That was about two years ago.
As I said before, I don't use SA all that much, since I
run it way at the back end just for grins after I run
all my own spam traps. But I did run it, and it worked.

--
dman

Ligesh

2005-08-01 00:19:56 UTC

Post by Ligesh
Btw, I am sorry, I never mentioned my spamassassin version - I am using redhat 8, with spamassasin version 2.31-16. I am testing by piping mails from mutt to procmail, and I have tried quite a lot of permutations - as I had said in my earlier mail, including running it through 'strace -f procmail'.

Ok here's what I think is happening. Procmail is using the output of the spamassasin rule for further comparisons, but it is delivering the original mail itself. So the variable that is used to make regexp search is getting corrupted, the the procmail doesn't let the mail get corrupted. So the mail gets delivered to the wrong mailbox.

That's it. Thanks for the help.

--
:: Ligesh :: http://ligesh.com

Dallman Ross

2005-08-01 00:17:22 UTC

Post by Ligesh
Ok here's what I think is happening. Procmail is using the
output of the spamassasin rule for further comparisons, but it
is delivering the original mail itself. So the variable that is
used to make regexp search is getting corrupted, the the procmail
doesn't let the mail get corrupted. So the mail gets delivered to
the wrong mailbox.

I don't know about that. Procmail delivers what you filtered the
message to with the fw flags.

I do remember a SA stupid peculiarity from a couple of years ago
where it would change the spaces in headers, though. Could that
be it? I.e., it would either do tab expansion or add extra
spaces after the header colons or something.

Dallman

Google Kreme

2005-08-01 12:01:03 UTC

Post by Ligesh
spamassasin version 2.31-16.

A spam tool that is how many years old? And, for the record, is
pretty much completely useless at this point.

The current version is 3.0.4. Install that and then come back.

When you come back, wrap your messages correctly and be user and post
a log of both the "working" and "non-working" procmail runs, as well
as what the "non-working" email looks like when it gets dumped into
the mbox at the end.

Damian Menscher

2005-07-31 16:40:04 UTC

Post by Dallman Ross
http://www.xray.mpe.mpg.de/mailing-lists/procmail/2005-07/msg00245.html
Do you think that looks all so readable?

Of course, badly written softwares will screw up; but that can't be helped. As a matter of fact, I have a 'set wrapmargin=40' in my muttrc, which will soft wrap the lines at around 80 chars itself. The comfortable reading length is actually 65 chars, but I can't have such a setting because a lot of people arbitrarily put hard newlines at 75 chars. :-)
- A single line paragraph will work in all circumstances. (You can configure your display software to wrap the line at whatever you wish. And it will work very fine in PDAs, and also will look good in forums.
- 75 char word wrap will screw up entirely in a lot of situations.
To me it trivially makes a formidable case for the former. Couple
this with the ease of editing at the sender end, I don't understand
why people still get into some religious fervor about the word
wrapping issue. 80 char is simply a historical relic from the days of
the dumb terminals. It is not only difficult to edit at the sender
end, it will screw up the display unless the screen width is more than
75 chars.

My mail software (pine) is configured to auto-wrap at 72 chars (74 is
the default, but I tend to get quoted a lot, and all the indentations
make it grow past the 80-col limit a bit too quickly. The nice thing,
though, is that now there's this thing called "flowed" text. You put an
extra space at the end of each line, and the other end's reader is
supposed to understand that it is a continuation of the same paragraph,
and therefore resize it to fit their column-width. The big problem with
it is that I often want specific formatting in my text, so I have to
disable it for about half my mails.

Your text was NOT flowed, hence when I reply it will probably still
quote you as a single long line (my MUA wrapped it for me). Just as a
test, I've flowed your last paragraph.

It should be interesting to see how *this* message shows up in the
archives. If people complain about it, then they need to read RFC 2646
(and then RFC 3676).

Damian Menscher

Professional Software Engineering

2005-07-31 20:22:09 UTC

Post by Ligesh
I have configured all my mailing lists to munge the 'Reply-To' header
and add the mailing list address there, so I always end up pressing just
'r'. I cannot understand the logic behind the other way, since most often
the person would want to reply to the mailing list, rather than to the
sender. I guess I have a problem with rules and practices which doesn't
make logical sense to me.

One logical rule I have is to use sensible subject lines. Another is to
not use replies to other messages (including list-s*bscription welcome
messages) to springboard a new message to the list because I'm too lazy to
type the list address and a proper subject on my posts. Besides having
subject lines which are nonsensical to the content of the message, the
"References:" and "In-Reply-To:" headers are inserted by modern mailers to
manage threads, and you end up having a thread relationship with an
unaffiliated topic.

But I'm sure YOUR "logic" makes sense to you. Somehow.

Post by Ligesh
Of course, badly written softwares will screw up; but that can't be helped.

Uhm, don't use badly written software? From most accounts here, it's
_your_ posts that are messed up, not everyone elses'.

Post by Ligesh
- A single line paragraph will work in all circumstances. (You can
configure your display software to wrap the line at whatever you wish.
And it will work very fine in PDAs, and also will look good in forums.

Not if that software expects to render the message AS SENT. There are
plenty of usenet posts and forum messages which require scrolling endlessly
to the right because they weren't wrapped by the sender.

FTR, the Eudora MUA has a reasonable way of formatting paragraphs and
sending them with hard coded line breaks - the ends of lines in a paragraph
have a trailing space. If a receiving program wants to re-flow the
paragraph, they can do so by reconstituting the paragraphs. I do this in
my blogger tool (to which I submit entries via email), which allows the web
representation to utilize the width and font style and size to
effect. Messages with this line wrapping style have a header like so:

Content-Type: text/plain; charset="us-ascii"; format=flowed

Despite being flagged as "flowed", the message DOES CONTAIN LINEBREAKS
WITHIN THE PARAGRAPHS. IOW, the software sends it as it expects to be
viewed, but provides hints for smart agents to be able to reflow it if they
so choose.

Further, since reply attributions are marked as attributions in so many
different ways, how 'zactly is the receiving software to know how to WRAP
the continuation marker (esp. if it doesn't KNOW that the line was a
quote)? If the SENDER has broken it into widely recognized "reasonable"
line lengths and tagged the beginning of each with an attribution marker
(historically ">", but there are others), then EACH LINE OF A PARAGRAPH
will properly have an attribution marker. When you get a long paragraph
without any linebreak, I can assure you that 99% or more of email clients
and forum programs which are tacked onto the back end of conventional
mailing lists, will render the message without any indication of quoting
added to the wrapped lines.

Post by Ligesh
- 75 char word wrap will screw up entirely in a lot of situations.

Not if the sender word wrapped WHAT THEY SENT.

Post by Ligesh
80 char is simply a historical relic from the days of the dumb terminals.

I view mail in a fixed width font, because proportional width fonts lose
ASCII ART characteristics (text diagrams, following line emphasis, etc),
commonly employed by technical types who've been using the internet since
before you were born.

Post by Ligesh
It is not only difficult to edit at the sender end, it will screw up the
display unless the screen width is more than 75 chars.

I've never had a problem editing a message.

Perhaps - and I'm going way out on a limb with my logic here - since in the
early days, internet mail was edited on computer terminals with 80 columns,
and nowadays, so many computers can display much more than that (using
graphic displays instead of strict text displays), perhaps dumbed down
devices like PDAs and cellphones could be the ones to add logic for
reflowing text to meet their SUBSTANDARD DISPLAYS. You can't actually surf
the REAL web on your cellphone or PDA, why should you expect email would be
the same?

---
Sean B. Straw / Professional Software Engineering

Procmail disclaimer: <http://www.professional.org/procmail/disclaimer.html>
Please DO NOT carbon me on list replies. I'll get my copy from the list.

Ligesh

2005-08-01 00:03:07 UTC

Post by Professional Software Engineering
One logical rule I have is to use sensible subject lines. Another is to
not use replies to other messages (including list-s*bscription welcome
messages) to springboard a new message to the list because I'm too lazy to
type the list address and a proper subject on my posts. Besides having
subject lines which are nonsensical to the content of the message, the
"References:" and "In-Reply-To:" headers are inserted by modern mailers to
manage threads, and you end up having a thread relationship with an
unaffiliated topic.

I am sorry, you are blabbering about something you have no clue about. What does 'In-Reply-To' has to do with munging 'Reply-To'? At least, I run a full mailing list which keeps the thread structure intact, but replaces any Reply-To with the mailing list. Mailman has a proper configuration option itself which allows you to replace any Reply-Tos with that of the mailing list. And your own signature at the end says that a copy shouldn't be sent to you. Exactly!!! If you configure mailman in the way I do, pressing reply will send a mail only to the list, and never to the sender. See. You have no idea of what the heck you are talking about.

Post by Professional Software Engineering
But I'm sure YOUR "logic" makes sense to you. Somehow.

Post by Ligesh
Of course, badly written softwares will screw up; but that can't be helped.

Uhm, don't use badly written software?

Of course, you shouldn't use badly written software. Is that a rhetorical question?

A display software has to take care of the limitations at its end. It shouldn't really make assumptions about the sender. If it is a single line paragraph, you just wrap it as you wish. And Wrapping is a trivial task. I think it is default in html to wrap and the idiotic mail archiver is forcibly adding a nowrap or something. That is pathetic design. And now because of bad software design you want everyone to follow some outdated convention?

Post by Professional Software Engineering
FTR, the Eudora MUA has a reasonable way of formatting paragraphs and
Despite being flagged as "flowed", the message DOES CONTAIN LINEBREAKS
WITHIN THE PARAGRAPHS. IOW, the software sends it as it expects to be
viewed, but provides hints for smart agents to be able to reflow it if they
so choose.
Further, since reply attributions are marked as attributions in so many
different ways, how 'zactly is the receiving software to know how to WRAP
the continuation marker (esp. if it doesn't KNOW that the line was a
quote)? If the SENDER has broken it into widely recognized "reasonable"
line lengths and tagged the beginning of each with an attribution marker
(historically ">", but there are others), then EACH LINE OF A PARAGRAPH
will properly have an attribution marker. When you get a long paragraph
without any linebreak, I can assure you that 99% or more of email clients
and forum programs which are tacked onto the back end of conventional

I don't understand why you need multiple attribution marker per paragraph. The purpose of attribution marker is to make it easy to distinguish your statements from others, and all normal MUAs do this by showing them in different colors. So whether you have a single attribution character or many, the the MUA will show the whole paragraph in a different color. If you are in a monchrome system, you can configure it to use underline or bold attribute or something.
The primary function of the 'attribution' marker is distinguish your messages from the other person, and I think this is served whether you have one or many per paragraph.

Post by Professional Software Engineering
mailing lists, will render the message without any indication of quoting
added to the wrapped lines.

Post by Ligesh
- 75 char word wrap will screw up entirely in a lot of situations.

Not if the sender word wrapped WHAT THEY SENT.

What if my screen is smaller than 75 chars. What if I am viewing the mail in vim that is split in the middle? And it looks quite ugly in forums.

Post by Ligesh
80 char is simply a historical relic from the days of the dumb terminals.

"Before I was born"? I think this statement should be grouped along with the 'famous' Nazi statement. That is, when someone starts talking 'before you were born', that means he has flipped, and is in need of serious medical treatment. I am running an fvwm with xterm -fn 10x20. with style "Style "nxterm" NoTitle, NoHandles, BorderWidth 0"; this means that this primary xterm is without handles or titles or borders or anything. So my X looks exactly like a dumb terminal. My primary mailer is vim, I use emacs as my primary web browser. The only graphical program I run is galeon 1.25 for websites that do not get displayed properly in emacs-w3m. My total configuration files comes to more than 35,000 lines (Without any comments). If you want to argue for being 'technical', I am sorry, you lose. Badly.

Post by Ligesh
It is not only difficult to edit at the sender end, it will screw up the
display unless the screen width is more than 75 chars.

I've never had a problem editing a message.
Perhaps - and I'm going way out on a limb with my logic here - since in the
early days, internet mail was edited on computer terminals with 80 columns,
and nowadays, so many computers can display much more than that (using
graphic displays instead of strict text displays), perhaps dumbed down
devices like PDAs and cellphones could be the ones to add logic for
reflowing text to meet their SUBSTANDARD DISPLAYS. You can't actually surf
the REAL web on your cellphone or PDA, why should you expect email would be
the same?

Your attitude is absurd. You are treating a random number - 75 in this case - as if it was ordained by God. Is this number found in Bible? Maybe next time you will start claiming so. Why 75? Why should every device in the world follow some arbitrary convention laid by the limitations of some terminal in some old age? It is pure insanity.

Post by Professional Software Engineering
Sean B. Straw / Professional Software Engineering
Procmail disclaimer: <http://www.professional.org/procmail/disclaimer.html>
Please DO NOT carbon me on list replies. I'll get my copy from the list.

Exactly. Please go configure the mailman like I had explained above. And you can avoid such ugly/arrogant disclaimers at the end of the mail.

--
:: Ligesh :: http://ligesh.com

Dallman Ross

2005-08-01 00:15:28 UTC

Post by Ligesh
A display software has to take care of the limitations at its
end. It shouldn't really make assumptions about the sender. If
it is a single line paragraph, you just wrap it as you wish. And
Wrapping is a trivial task. I think it is default in html to
wrap and the idiotic mail archiver is forcibly adding a nowrap
or something. That is pathetic design. And now because of bad
software design you want everyone to follow some outdated
convention?

My mutt wraps your lines. But I don't want to look all the way over
to the 124th column (where I currently have my terminal emulation
program set for cols) to see the end of the line -- and that, with
no right whitespace. That is very tiring. I want to be able
to read the text edge-to-edge without moving my head all around.

Post by Ligesh
I don't understand why you need multiple attribution marker per
paragraph. The purpose of attribution marker is to make it easy
to distinguish your statements from others, and all normal MUAs
do this by showing them in different colors. So whether you

Ex*cuse* me? *Color*? and that's *normal*?

I am not seeing any color here. Nor do I want to.

I also use Outlook, and I don't see any color there, either.
Nor do I see any MIME/HTML stuff, as I have a COM plug-in
to turn it off. I want plain text.

If I want color, I'll go look at some web pages. Please!

P.S. Do you know how many system resources it burns to run
color with GNU ls?

Dallman

Professional Software Engineering

2005-08-01 02:22:46 UTC

Post by Ligesh
I am sorry, you are blabbering about something you have no clue about.
What does 'In-Reply-To' has to do with munging 'Reply-To'?

Drop the crayons and back away from the computer.

Refer to your original post to this list. Note the subject, note the Re:,
when it wasn't in reply to a discussion here, and how the subject is
obviously from a s*bscription welcome message and is therefore meaningless
babble, note the References: which didn't relate to any post on this list.

I referred to rules and logic. You obviously don't share the same ones WRT
meaningful group communication, yet you question the chosen protocols of
this group, which weren't decided overnight, and certainly not by a single
individual.

Post by Ligesh
sender. See. You have no idea of what the heck you are talking about.

I'll gladly sell you half a dozen clues, but you'll have to pay for them in
USD via PayPal in advance of delivery (I've been cheated way too many times
- some people take the clues and run). If you need more, you'll need to
find them elsewhere, since I like to keep a healthy number of clues in
reserve for some of the other clueless people I deal with.

Post by Ligesh
What if my screen is smaller than 75 chars. What if I am viewing the
mail in vim that is split in the middle? And it looks quite ugly in forums.

Inadequate systems should compensate for their own inadequacies, not
redefine how the rest of the internet should fuction so that they may be
included. The converse is reasonably true as well - just because some
"nifty" new thing like yellow or cyan on white text is available doesn't
mean it should be utilized, or that it is a good idea. Colour in email
communication leaves a lot to be desired, and I've yet to see *ANY* RFC's
defining what colours mean what.

Perhaps you'd like to take up the torch for top posting too?

Look at all the different web forum implementations out there, no two using
quite the same way to present data. Is this really a BENEFIT to
users? Having to learn to navigate so many different wannabe standards in
order to communicate?

Email has remained the effective communication tool that it is because it
is simple. Technical mailing lists have by and large chosen to remain
plain text because it doesn't clutter communications with someone's idea of
"style" and colour preferences.

Post by Ligesh
Your attitude is absurd. You are treating a random number - 75 in this case

No, actually 80, and it isn't random - I already explained the origins of
the 80 column limitation, and you know that, but you seem bent on arguing.
75 or so allows for easy indentation for a few replies without changing
formatting (and anyone quoting something five itarations back probably
shouldn't be doing that anyway).

Post by Ligesh
- as if it was ordained by God.

Kind 'o hard for me to do since I'm an athiest, not that it has anything do
do with my technical outlook.

(Wow, bringing up war criminals AND religion in justification of your
position. How mature and well reasoned.)

Post by Ligesh
Exactly. Please go configure the mailman like I had explained above.
And you can avoid such ugly/arrogant disclaimers at the end of the mail.

Requesting people to not carbon me is arrogant? Actually, the ONLY people
who manage to CC: me are either entering addresses manually, or have buggy
MUAs, and therefore changed to the mailman configuration would have NO
effect. But since you're mangling your email on arrival, I'm guessing
you're unaware of that.

Also, as it happens, I _can_ change the mailman options for this list, but
I wouldn't dream of doing so unless there was a consensus of the
contributing users, and prior to your arrival, this hasn't been much of a
hot topic, since people here are usually geared towards discussing how to
resolve problems with procmail.

It'd be awfully nice if the people who plop in here would respect the
protocols of this group rather than telling us we're all idjuts for
choosing efficiency over glitz. But hey, we all suffer from short
attention spans, so nobody is going to remember your actions the next time
you seek assistance, right?

Lessee, address in subject, sign with my public key, and send to my
killfile processor... Okay, done.

---
Sean B. Straw / Professional Software Engineering

Procmail disclaimer: <http://www.professional.org/procmail/disclaimer.html>
Please DO NOT carbon me on list replies. I'll get my copy from the list.

Ligesh

2005-08-01 13:15:04 UTC

Post by Professional Software Engineering
Drop the crayons and back away from the computer.

So said the old man wobbling unsteadily on his keyboard. :-).

You need to use crayons where crayons are needed, and you shouldn't shy away from them feeling that you are too old. Anyway, you call using xterm -fn 10x20 -bg black -fg gray72 as glitzy stuff? You need to learn to use to your xterm more effectively. You can configure all the 16 colors of xterm with the following configuration in your xresource file.

------------------------

*foreground: grey73
*background: black
*highlight : red

XTerm*color0: #000000
XTerm*color1: #b54f6f
XTerm*color2: #00bb00
XTerm*color3: #bcb300
XTerm*color4: #0090bf
XTerm*color5: #56a656
XTerm*color6: #00bbbf
XTerm*color7: #bfbfbf
XTerm*color8: #458325
XTerm*color9: #f04050
XTerm*color10: #a0b2c5
XTerm*color11: #cf0f00
XTerm*color12: #a070ad
XTerm*color13: #cf8085
XTerm*color14: #af93af
XTerm*color15: #90bb9f
XTerm*color16: #00d500
---------------------

The advantage of the above set of colors is that ALL of them have absolutely the same intensity, and have the same as the gray73. So none of the colors will ever hit your eye. I have, in effect, removed all the bold and intense colors and replaced them with the lighter versions.

You can configure mutt this way.

------------------- Muttrc
color markers magenta black
color tree color3 black
color search green black
color signature red black
color quoted2 cyan black
color quoted1 color5 black
color bold color6 black
color underline color6 black
------------------

Vim config.

------------- syntax/mail.vim
hi MailQuoted1 ctermfg=DarkCyan
hi MailQuoted2 ctermfg=none
hi MailQuoted3 ctermfg=11
hi MailQuoted4 ctermfg=12
hi mailHeaderKey ctermfg=DarkGreen
hi mailHeader ctermfg=DarkBlue
hi mailSubject ctermfg=Brown
hi Special ctermfg=13
--------------

The above is not for creating glitz, but rather helps tremendously in easily identifying the required sections of the mail. After sometime your brain will subconsciously do the identifying, and it will come very naturally.

Post by Ligesh
sender. See. You have no idea of what the heck you are talking about.

You have a talent to be pedantic without being informative. Maybe in some quarters it is something that can fetch money. But not with any place I know.

Post by Ligesh
What if my screen is smaller than 75 chars. What if I am viewing the
mail in vim that is split in the middle? And it looks quite ugly in forums.

- The important point you miss: The single line paragraph will work everywhere. So why not just use something that will work in all circumstances, rather than penalize poor people who don't have enough money to buy all those large systems.

The point I am trying to get across for so long is that: A Single line paragraph will work in all situations. So instead of using something that will help everyone, you are vehemently arguing for continuing with your barbaric ways. Pardon me if I find it absurd. Anyway, your idea is to just ignore everything that is happening around you, and do exactly the same thing regardless. I don't think that is a good idea. You have to evolve. Just ask the dinosaurs.

Post by Professional Software Engineering
Perhaps you'd like to take up the torch for top posting too?
Look at all the different web forum implementations out there, no two using
quite the same way to present data. Is this really a BENEFIT to
users? Having to learn to navigate so many different wannabe standards in
order to communicate?

There are a lot of people using forums. I have no issues with forums as long as they support lynx. And as for my own forums, I have integrated everything tightly with mailman. I use fudforum, which has in-built mailing list synchronizatin logic. So I get the best of both worlds. This is exactly the problem with people. On one side we have utter newbies who don't care about efficiency at all, and on the other end we have people hard set in their ways ignoring any of the new technologies. I am trying to bridge a gap between the two. I want to have the efficiency of the mailing system (For my own use), and also want the ease of use and glitz that a forum provides. (For the other people).

Post by Professional Software Engineering
Email has remained the effective communication tool that it is because it
is simple. Technical mailing lists have by and large chosen to remain
plain text because it doesn't clutter communications with someone's idea of
"style" and colour preferences.

I don't remember arguing for sending color emails. Just that you can configure your mailer to show the paragraphs in different colors which will make it very easy to distinguish your message from quoted text. It is there in all text mailers - mutt/vim and even vm - the mailer for emacs.

--
:: Ligesh :: http://ligesh.com

David W. Tamkin

2005-08-01 05:32:09 UTC

Post by Ligesh
You are treating a random number - 75 in this case - as if it was
ordained by God. Is this number found in Bible? Maybe next time you
will start claiming so.

Genesis 12:4.

Google Kreme

2005-08-01 12:07:29 UTC

Post by David W. Tamkin

Post by Ligesh
You are treating a random number - 75 in this case - as if it was
ordained by God. Is this number found in Bible? Maybe next time you
will start claiming so.

Genesis 12:4.

LOL! Oh, that's a good one, that is.

Peter Jones

2005-08-01 15:52:45 UTC

Post by David W. Tamkin

Post by Ligesh
You are treating a random number - 75 in this case - as if it was
ordained by God. Is this number found in Bible? Maybe next time you
will start claiming so.

Genesis 12:4.

Nice! :-)

Dallman Ross

2005-08-01 11:03:29 UTC

Ligesh, you are claiming to apply careful logic; but (to the degree
your logic is not flawed, which point I won't go into right this sec)
you are doing so in a vaccuum, as if no other well-followed practice
had ever evolved and you are free simply to invent your own way,
irrespective of other users that have spent time, effort, and energy
designing their own interactive tools to comply with what's out there
already.

You're acting like a hermit. That has its attractions -- I have
fantasized more than once about becoming a hermit, and I dare say
many others have, as well. But to carry it out in actuality, yet
then interact with others on your hermit's terms nonetheless, just
doesn't work, and is bound to create friction.

[L to Sean]

Post by Ligesh
Exactly. Please go configure the mailman like I had explained
above. And you can avoid such ugly/arrogant disclaimers at the
end of the mail.

Regarding how the list server sets follow-ups, this list does
it correctly. If it would be set to follow up by default to the
list just with the normal "reply" action, then we'd see all
manner of bounce messages and "out-of-office" crap, etc., not
to mention worm and bot stuff.

Patrick Shanahan: what was that FAQ URL you once showed me in
this regard, again?

Dallman

Ligesh

2005-08-01 13:22:37 UTC

A google search for par is not giving me any meaningful result. Can you just give me the package name or something.

thanks.

--
:: Ligesh :: http://ligesh.com

Patrick Shanahan

2005-08-01 14:05:49 UTC

Post by Dallman Ross
Patrick Shanahan: what was that FAQ URL you once showed me in
this regard, again?

Sorry if memory doesn't serve, but wasn't a URL or FAQ, but quoting
from the Fine Mutt Manual repeated here:

In answer to your questions above, a couple of quotes from TFM for mutt
1.5.6i:

The other method some mailing list admins use is to generate a
Reply-To'' field which points back to the mailing list address rather
than the author of the message. This can create problems when trying to
reply directly to the author in private, since most mail clients will
automatically reply to the address given in the Reply- To'' field. Mutt
uses the $reply_to'' variable to help decide which address to use. If
set, you will be prompted as to whether or not you would like to use the
address given in the Reply-To'' field, or reply directly to the address
given in the From'' field. When unset, the Reply-To'' field will be
used when present.

note: I have reply_to SET

6.3.72. ignore_list_reply_to

Type: boolean
Default: no

Affects the behaviour of the reply function when replying to messages
from mailing lists. When set, if the Reply-To:'' field is set to the
same value as the To:'' field, Mutt assumes that the Reply- To:'' field
was set by the mailing list to automate responses to the list, and will
ignore this field. To direct a response to the mailing list when this
option is set, use the list-reply function; group-reply will reply to
both the sender and the list.

note: I have ignore_list_reply_to SET

list-reply (default: L)

Reply to the current or tagged message(s) by extracting any addresses
which match the addresses given by the lists or subscribe'' commands, but
also honor any Mail-Followup-To header(s) if the $honor_followup_to''
configuration variable is set. Using this when replying to messages
posted to mailing lists helps avoid duplicate copies being sent to the
author of the message you are replying to.

This last, list-reply, I consider the most important. The way I am
configured, <L> ignores all addresses except those listed in .muttrc as
<subscribe>, in which you actively participate, and <list> which you may
watch and allows a list-address identification, from TFM:

Mutt has a few nice features for handling mailing lists''. In order to
take advantage of them, you must specify which addresses belong to
mailing lists, and which mailing lists you are subscribed to. Once you
have done this, the list-reply'' function will work for all known lists.
Additionally, when you send a message to a subscribed list, mutt will add
a Mail-Followup-To header to tell other users' mail user agents not to
send copies of replies to your personal address. Note that the
Mail-Followup-To header is a non-standard extension which is not
supported by all mail user agents. Adding it is not bullet-proof against
receiving personal CCs of list messages. Also note that the generation of
the Mail-Followup-To header is controlled by the $followup_to''
configuration variable.

--
Patrick Shanahan Registered Linux User #207535
http://wahoo.no-ip.org @ http://counter.li.org
HOG # US1244711 Photo Album: http://wahoo.no-ip.org/gallery

Ligesh

2005-08-01 15:25:46 UTC

Post by Patrick Shanahan
Reply to the current or tagged message(s) by extracting any addresses
which match the addresses given by the lists or subscribe'' commands, but
also honor any Mail-Followup-To header(s) if the $honor_followup_to''
configuration variable is set. Using this when replying to messages
posted to mailing lists helps avoid duplicate copies being sent to the
author of the message you are replying to.
This last, list-reply, I consider the most important. The way I am
configured, <L> ignores all addresses except those listed in .muttrc as
<subscribe>, in which you actively participate, and <list> which you may
Mutt has a few nice features for handling mailing lists''. In order to
take advantage of them, you must specify which addresses belong to
mailing lists, and which mailing lists you are subscribed to. Once you
have done this, the list-reply'' function will work for all known lists.
Additionally, when you send a message to a subscribed list, mutt will add

See, this is workable only for people who are regular members of a list. I have been in and out of so many mailing lists, I decided to give it a miss. I am sure most regular people are also the same. I have 'reply-to' set to ask-yes in general mailboxes, and in mailing-list specific mailboxes, set to yes. But it doesn't work for mailing list where the 'reply-to' is not added.

--
:: Ligesh :: http://ligesh.com

Peter Jones

2005-08-01 15:41:46 UTC

Post by Ligesh
This is the spamassassin rule.
---------
| spamassassin
------------

FWIW, my recipe for passing mail through spamassassin looks something like
this:

:0 fw: spamassassin.lock
* < 256000
| spamc

I cannot remember exactly where I got this from, but I recall that it was
specifically only looking at messages smaller than 256000 bytes; larger
mails may well cause problems with SA, apparently.

I also use spamc rather than spamassassin because it was recommended at
the time; whether this should make a difference I don't know, but I seem
to recall you yourself said somewhere that if you switched to spamc it
all worked; why not just do that?

Finally, I got the impression that a lock file was required; I'm seeing
others here saying it isn't? Which is correct?

Cheers,
Pete.

Professional Software Engineering

2005-08-01 17:14:20 UTC

Post by Peter Jones
I also use spamc rather than spamassassin because it was recommended at
the time; whether this should make a difference I don't know, but I seem
to recall you yourself said somewhere that if you switched to spamc it
all worked; why not just do that?

He doesn't have buggy software - the problem is US and our software.

Post by Peter Jones
Finally, I got the impression that a lock file was required; I'm seeing
others here saying it isn't? Which is correct?

Well, procmail itself isn't at risk of overwriting a file when you pipe to
a process (unless you have a series and eventually redirect to a file*),
but if that process writes to files and doesn't perform locking
(translation: it's poorly written, AKA buggy), then having procmail wrap
the operation in a lockfile will help protect against that (so long as it's
a per-user file that the buggy process is trying to manipulate -- if it
tries writing to a file that OTHER users might also be able to write to,
unless they are also using the SAME lockfile as yourself (which is terribly
unlikely), then YOUR lockfile is meaningless to them). Additionally, a
lockfile will throttle the system load generated by SA - granted, it'll
hold up ALL of your mail (at least that which should be piped through SA),
until the prior SA process completes.

I work for an ISP, and we have about a dozen servers running spamd (the
daemon interface for SpamAssassin, allowing other mail servers to submit
messages across the network to check them for spammishness), and the
loading on them is astounding.

IMO, SA is a poorly written program, lacking in optimizations, and which
assumes that there is unlimited resources available. It doesn't throttle well.

* like so:

| someprocess > yourfile

In which case, you should definatley have a lockfile for
yourfile. Procmail should manage this fine with an implicit (unnamed)
lockfile flag.

---
Sean B. Straw / Professional Software Engineering

Procmail disclaimer: <http://www.professional.org/procmail/disclaimer.html>
Please DO NOT carbon me on list replies. I'll get my copy from the list.

Peter Jones

2005-08-02 12:57:38 UTC

Post by Professional Software Engineering
He doesn't have buggy software - the problem is US and our software.

That's right; silly of me to forget...

Post by Professional Software Engineering
Additionally, a lockfile will throttle the
system load generated by SA - granted, it'll hold up ALL of your mail
(at least that which should be piped through SA), until the prior SA
process completes.

That sounds a good enough reason for me to keep the lockfile, at least for
now. Holding up my mail a little isn't a big issue for me; it's only
*my* mail, and the volume isn't huge.

Thanks for clarifying the lockfile issue for me! (And for your thoughts
on spamassassin too -- at the moment it's doing a pretty good job of
minimising the spam I see, so I'll stick with it, but based on what
you've said I may do a little research into alternatives...

Cheers,
Pete.

Google Kreme

2005-08-04 02:26:30 UTC

Post by Peter Jones
:0 fw: spamassassin.lock
* < 256000
| spamc
I cannot remember exactly where I got this from, but I recall that it was
specifically only looking at messages smaller than 256000 bytes; larger
mails may well cause problems with SA, apparently.

Your recipe is a few years out of date. Spamassassin will not
process mail over 256K in size since at least 3.x, and I think it was
changed to do this in 2.6...

Post by Peter Jones
I also use spamc rather than spamassassin because it was
recommended at
the time; whether this should make a difference I don't know, but I seem
to recall you yourself said somewhere that if you switched to spamc it
all worked; why not just do that?

spamc requires spamd to be running on the system. It is much faster
than running spamassassin.

(and no, you never need a lock file for a filter)

--
Nothing like grilling a kosher dog over human hair to bring out the
subtle flavors.

Damian Menscher

2005-08-04 02:32:34 UTC

Post by Google Kreme

Post by Peter Jones
:0 fw: spamassassin.lock
* < 256000
| spamc
I cannot remember exactly where I got this from, but I recall that it
was specifically only looking at messages smaller than 256000 bytes;
larger mails may well cause problems with SA, apparently.

Your recipe is a few years out of date. Spamassassin will not
process mail over 256K in size since at least 3.x, and I think it was
changed to do this in 2.6...

It's 250k by default, and can be changed with the -s flag. If you can
handle the load, it may be worth increasing, now that spammers have
started including pictures of their wares.

Damian Menscher

G.W. Haywood

2005-08-01 12:01:04 UTC

Hi there,

Post by David W. Tamkin

Post by Ligesh
You are treating a random number - 75 in this case - as if it was
ordained by God. Is this number found in Bible? Maybe next time you
will start claiming so.

Genesis 12:4.

Nice try, but that's the wrong Bible.

Please, everyone, read RFC2822, particularly section 2.1.1.

And then *please* shut up on this subject. People are trying to work.

73,
Ged.

Damian Menscher

2005-08-01 15:28:56 UTC