What's wrong with this fancy split?

What's wrong with this fancy split?

[Tommy Kelly]

This fancy split is meant to match emails with "[some-list]" in the
subject.

("subject" "\\[some-list\\]" "some-list")

But it isn't. Any ideas why?

Tommy

Re: What's wrong with this fancy split?

[Andreas Schwab]

Tommy Kelly <tommy.kelly@verilab.com> writes:

> This fancy split is meant to match emails with "[some-list]" in the
> subject.
>
> ("subject" "\\[some-list\\]" "some-list")
>
> But it isn't. Any ideas why?

Because [ and ] are not word characters.  See
nnmail-split-fancy-syntax-table.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: What's wrong with this fancy split?

[Russ Allbery]

Andreas Schwab <schwab@linux-m68k.org> writes:
> Tommy Kelly <tommy.kelly@verilab.com> writes:

>> This fancy split is meant to match emails with "[some-list]" in the
>> subject.
>>
>> ("subject" "\\[some-list\\]" "some-list")
>>
>> But it isn't. Any ideas why?

> Because [ and ] are not word characters.  See
> nnmail-split-fancy-syntax-table.

Practically speaking, what this means is that you need to write:

    ("subject" ".*\\[some-list\\].*" "some-list")

instead.  The regular expressions only match at word boundaries, which
usually isn't what you want when they start or end with punctuation.

-- 
Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>

Re: What's wrong with this fancy split?

[Tommy Kelly]

Russ Allbery <rra@stanford.edu> writes:

> Practically speaking, what this means is that you need to write:
>
>     ("subject" ".*\\[some-list\\].*" "some-list")
>
> instead.  The regular expressions only match at word boundaries, which
> usually isn't what you want when they start or end with punctuation.

Excellent! A thousand thank-yous. May piles of gold and silver be
discovered in your front yard along with the legal paperwork proving
that it belongs to you. 

thanks,
Tommy

Re: What's wrong with this fancy split?

[Andreas Schwab]

Russ Allbery <rra@stanford.edu> writes:

> Andreas Schwab <schwab@linux-m68k.org> writes:
>> Tommy Kelly <tommy.kelly@verilab.com> writes:
>
>>> This fancy split is meant to match emails with "[some-list]" in the
>>> subject.
>>>
>>> ("subject" "\\[some-list\\]" "some-list")
>>>
>>> But it isn't. Any ideas why?
>
>> Because [ and ] are not word characters.  See
>> nnmail-split-fancy-syntax-table.
>
> Practically speaking, what this means is that you need to write:
>
>     ("subject" ".*\\[some-list\\].*" "some-list")

Or make them word characters.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: What's wrong with this fancy split?

[Steinar Bang]

>>>>> Andreas Schwab <schwab@linux-m68k.org>:

> Russ Allbery <rra@stanford.edu> writes:

>> Practically speaking, what this means is that you need to write:

>> ("subject" ".*\\[some-list\\].*" "some-list")

> Or make them word characters.

Making [ and ] word characters for something like this, sounds like a
receipe for confusion.

I'd say keep it simple and go with Russ' suggestion.

Re: What's wrong with this fancy split?

[Andreas Schwab]

Steinar Bang <sb@dod.no> writes:

>>>>>> Andreas Schwab <schwab@linux-m68k.org>:
>
>> Russ Allbery <rra@stanford.edu> writes:
>
>>> Practically speaking, what this means is that you need to write:
>
>>> ("subject" ".*\\[some-list\\].*" "some-list")
>
>> Or make them word characters.
>
> Making [ and ] word characters for something like this, sounds like a
> receipe for confusion.

Why?  That's the whole point of nnmail-split-fancy-syntax-table.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: What's wrong with this fancy split?

[Steinar Bang]

>>>>> Andreas Schwab <schwab@linux-m68k.org>:

> Why?

In three words: keep it simple

Re: What's wrong with this fancy split?

[Andreas Schwab]

Steinar Bang <sb@dod.no> writes:

> In three words: keep it simple

Then you should not use Gnus.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: What's wrong with this fancy split?

[Bjørn Mork]

Andreas Schwab <schwab@linux-m68k.org> writes:
> Steinar Bang <sb@dod.no> writes:
>
>> In three words: keep it simple
>
> Then you should not use Gnus.

¿què?


Bjørn

Re: What's wrong with this fancy split?

[Tommy Kelly]

Andreas Schwab <schwab@linux-m68k.org> writes:

> Steinar Bang <sb@dod.no> writes:
>
>> In three words: keep it simple
>
> Then you should not use Gnus.

That seems like a bit of a non sequitur. And in fact to be honest it
risks being simply a cop  out. Being as simple *as possible* is a sign of
enginering and programming excellence. As Fred Brooks said, complexity
is our business. So there's no need to add more of it ourselves.

If that's not the case then here are some suggestions:

a. Remove the documentation. All of it.
b. Remove all the code comments
c. Obfuscate all the lisp
d. Encrypt it
e. Let everyone answer all questions on this list in Esperanto. Do that
until a majority are fluent in that language, then switch to
Aramaic. Then Coptic. Keep changing as long as it takes to ensure that
things are never simple.

Tommy

Re: What's wrong with this fancy split?

[Tassilo Horn]

Tommy Kelly <tommy.kelly@verilab.com> writes:

Saluton Tommy,

> e. Let everyone answer all questions on this list in Esperanto.

Bonvolu skribi vian poŝtantan En Esperanto.

SCNR,
Tassilo
-- 
Sendita de mia Emacs

Re: What's wrong with this fancy split?

[Tommy Kelly]

Tassilo Horn <tassilo@member.fsf.org> writes:

> Bonvolu skribi vian poŝtantan En Esperanto.

Andvarp. Ég gæti hafa vitað að það væri að minnsta kosti einn esperantó
náungi hér. :-)

Tommy

Re: What's wrong with this fancy split?

[Steinar Bang]

>>>>> Tommy Kelly <tommy.kelly@verilab.com>:

> Andvarp. Ég gæti hafa vitað að það væri að minnsta kosti einn
> esperantó náungi hér. :-)

I tillegg er det opptil flere som kan gjøre seg forstått på norsk.

Re: What's wrong with this fancy split?

[Tommy Kelly]

Steinar Bang <sb@dod.no> writes:

>>>>>> Tommy Kelly <tommy.kelly@verilab.com>:
>
>> Andvarp. Ég gæti hafa vitað að það væri að minnsta kosti einn
>> esperantó náungi hér. :-)
>
> I tillegg er det opptil flere som kan gjøre seg forstått på norsk.

Aiblins a' fowk shoods write in their natife tongue an' 'en Andreas kin
whip up some lisp tae create a kin' ay babel fesh hin', eh?

Re: What's wrong with this fancy split?

[Steinar Bang]

>>>>> Tommy Kelly <tommy.kelly@verilab.com>:

> Aiblins a' fowk shoods write in their natife tongue an' 'en Andreas
> kin whip up some lisp tae create a kin' ay babel fesh hin', eh?

Eg e' heilt eni'!

Re: What's wrong with this fancy split?

[Tommy Kelly]

Steinar Bang <sb@dod.no> writes:

>>>>>> Tommy Kelly <tommy.kelly@verilab.com>:
>
>> Aiblins a' fowk shoods write in their natife tongue an' 'en Andreas
>> kin whip up some lisp tae create a kin' ay babel fesh hin', eh?
>
> Eg e' heilt eni'!

Utmärkt! Glad att vi är överens. Vi återvänder nu till dig vår vanliga
programtablån. Så, komme tilbake til denne fancy-split ting. Tror du
Lars var beruset da han skrev det, eller er det bare hans onde Norske
sans for humor?  

Tomi

Re: What's wrong with this fancy split?

[Steinar Bang]

>>>>> Tommy Kelly <tommy.kelly@verilab.com>:

> Utmärkt! Glad att vi är överens. Vi återvänder nu till dig vår vanliga
> programtablån. Så, komme tilbake til denne fancy-split ting. Tror du
> Lars var beruset da han skrev det, eller er det bare hans onde Norske
> sans for humor?

Jeg tror det er så lenge sia han skrev det at han ikke en gang husker
hva han tenkte.  Vi snakker om et eller annet sted mellom 1997 og 2000,
tror jeg.

Re: What's wrong with this fancy split?

[Tommy Kelly]

Steinar Bang <sb@dod.no> writes:

> Jeg tror det er så lenge sia han skrev det at han ikke en gang husker
> hva han tenkte.  Vi snakker om et eller annet sted mellom 1997 og 2000,
> tror jeg.

1994 tror jeg. Jeg tror jeg først brukte det ikke så lenge etter at den
kom ut i 1995. Hvor bra er dette norske likevel - Jeg har ingen anelse
om hva jeg sier, jeg er å stole helt på Google translate. 

Re: What's wrong with this fancy split?

[Steinar Bang]

>>>>> Tommy Kelly <tommy.kelly@verilab.com>:

> 1994 tror jeg. Jeg tror jeg først brukte det ikke så lenge etter at
> den kom ut i 1995.

På den tida brukte jeg Gnus bare til njus (hah! den tar ikke google).

> Hvor bra er dette norske likevel

Forståbar, men med litt kreativ grammatikk.

> - Jeg har ingen anelse om hva jeg sier, jeg er å stole helt på Google
> translate.

Jau det...

Re: What's wrong with this fancy split?

[Tommy Kelly]

Steinar Bang <sb@dod.no> writes:

> Forståbar ...

??? 

Re: What's wrong with this fancy split?

[Steinar Bang]

>>>>> Tommy Kelly <tommy.kelly@verilab.com>:

> Steinar Bang <sb@dod.no> writes:
>> Forståbar ...

> ??? 

What?  Google translate isn't up to a little "nyord" challenge?

But tomorrow is Friday.  And Friday is pun day.  So... let's say it is
"sjekkested med sjenkebevilling".

Re: What's wrong with this fancy split?

[Richard Riley]

Andreas Schwab <schwab@linux-m68k.org> writes:

> Steinar Bang <sb@dod.no> writes:
>
>> In three words: keep it simple
>
> Then you should not use Gnus.
>
> Andreas.

What a strange and negative reply.

Or things could be made simpler.

Re: What's wrong with this fancy split?

[Tommy Kelly]

In response to
>> Making [ and ] word characters for something like this, sounds like a
>> receipe for confusion.

Andreas Schwab <schwab@linux-m68k.org> writes:
> Why?  

Because the original writer of "sounds like a recipe for confusion" is
understanding the reader where you Andreas, with respect, are not.

This is not about the code. It's about the people using the code. You
clearly get the former. Maybe not so much the latter when they are less
knowledgeable about this stuff than you.

> That's the whole point of nnmail-split-fancy-syntax-table.

If that's really the case. If the whole point of
nnmail-split-fancy-table is to ensure no confusion in this area, then
it's failing to achieve its point with what I'm guessing is a reasonable
subset of those who could suffer from said confusion.

But that's OK. That's what documentation is for :-)

Tommy

Just shoot me now (was Re: What's wrong with this fancy split?)

[Tommy Kelly]

Andreas Schwab <schwab@linux-m68k.org> writes:

> Because [ and ] are not word characters.  See
> nnmail-split-fancy-syntax-table.

Thanks Andreas. Unfortunately that 6,500 line variable, consisting
mostly of pairs of parens enclosing numbers, is as clear as mud :-)

This to no one in particular:

Thus far, like my adventures in figuring out what version of gnus is
current, my attempts Just To Split My Mail have been ... interesting:

1. Try to figure out multi-line regexps with simple splitting
2. Advised to move to fancy splitting
3. Hours of trying to get even the simplest fancy split working (and I'm
not alone)
4. Finally get it moving -- but only after direct advice. But still
don't understand why (specifically, why is nnmail-split-methods settable
on a per-server basis, but nnmail-split-fancy is done on a ... shrug
.. a per gnus basis?)
5. Begin building more simple splits then stumble on
"\\[stuff\\]". Splitting isn't as keen on \\[ as it is on, say, \\. .
6. Told that nnmail-split-fancy-syntax-table is the clue
7. C-h v on that variable. Look. Puke. Look again.
8. Google for "emacs lisp syntax table"
9. Fall pretty much at the first fence on being told that a syntax table
is a "char-table" about which I know very little
10. Start reading about what a char-table is
11. Consider reading about "aref" and "aset" since part of the basic
description of char-tables mentions those
12. Stop and think "Seriously!? All this Just To Split My Mail!?"

I'm not remotely demanding, requesting or even begging for change. I
acknowledge that Gnus is a labour of love, open source, and is there to
be taken or left as each sees fit.

I'm just saying.

Tommy

Re: Just shoot me now

[Reiner Steib]

On Wed, Dec 22 2010, Tommy Kelly wrote:

> Andreas Schwab <schwab@linux-m68k.org> writes:
>
>> Because [ and ] are not word characters.  See
>> nnmail-split-fancy-syntax-table.
>
> Thanks Andreas. Unfortunately that 6,500 line variable, consisting
> mostly of pairs of parens enclosing numbers, is as clear as mud :-)
[...]
> 5. Begin building more simple splits then stumble on
> "\\[stuff\\]". Splitting isn't as keen on \\[ as it is on, say, \\. .
> 6. Told that nnmail-split-fancy-syntax-table is the clue
> 7. C-h v on that variable. Look. Puke. Look again.
> 8. Google for "emacs lisp syntax table"
> 9. Fall pretty much at the first fence on being told that a syntax table
> is a "char-table" about which I know very little
> 10. Start reading about what a char-table is
> 11. Consider reading about "aref" and "aset" since part of the basic
> description of char-tables mentions those
> 12. Stop and think "Seriously!? All this Just To Split My Mail!?"

Right, much too complicated.  The solution is described in the manual:

,----[ (info "(gnus)Fancy Mail Splitting") ]
|    Normally, VALUE in these splits must match a complete _word_
| according to the fundamental mode syntax table.  In other words, all
| VALUE's will be implicitly surrounded by `\<...\>' markers, which are
| word delimiters.  Therefore, if you use the following split, for
| example,
| 
|      (any "joe" "joemail")
| 
| messages sent from `joedavis@foo.org' will normally not be filed in
| `joemail'.  If you want to alter this behavior, you can use any of the
| following three ways:
| 
|   1. You can set the `nnmail-split-fancy-match-partial-words' variable
|      to non-`nil' in order to ignore word boundaries and instead the
|      match becomes more like a grep.  This variable controls whether
|      partial words are matched during fancy splitting.  The default
|      value is `nil'.
| 
|      Note that it influences all VALUE's in your split rules.
| 
|   2. VALUE beginning with `.*' ignores word boundaries in front of a
|      word.  Similarly, if VALUE ends with `.*', word boundaries in the
|      rear of a word will be ignored.  For example, the VALUE
|      `"@example\\.com"' does not match `foo@example.com' but
|      `".*@example\\.com"' does.
| 
|   3. You can set the INVERT-PARTIAL flag in your split rules of the
|      `(FIELD VALUE ...)' types, aforementioned in this section.  If the
|      flag is set, word boundaries on both sides of a word are ignored
|      even if `nnmail-split-fancy-match-partial-words' is `nil'.
|      Contrarily, if the flag is set, word boundaries are not ignored
|      even if `nnmail-split-fancy-match-partial-words' is non-`nil'.
|      (New in Gnus 5.10.7)
`----

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Just shoot me now

[Richard Riley]

Reiner Steib <reinersteib+gmane@imap.cc> writes:

> On Wed, Dec 22 2010, Tommy Kelly wrote:
>
>> Andreas Schwab <schwab@linux-m68k.org> writes:
>>
>>> Because [ and ] are not word characters.  See
>>> nnmail-split-fancy-syntax-table.
>>
>> Thanks Andreas. Unfortunately that 6,500 line variable, consisting
>> mostly of pairs of parens enclosing numbers, is as clear as mud :-)
> [...]
>> 5. Begin building more simple splits then stumble on
>> "\\[stuff\\]". Splitting isn't as keen on \\[ as it is on, say, \\. .
>> 6. Told that nnmail-split-fancy-syntax-table is the clue
>> 7. C-h v on that variable. Look. Puke. Look again.
>> 8. Google for "emacs lisp syntax table"
>> 9. Fall pretty much at the first fence on being told that a syntax table
>> is a "char-table" about which I know very little
>> 10. Start reading about what a char-table is
>> 11. Consider reading about "aref" and "aset" since part of the basic
>> description of char-tables mentions those
>> 12. Stop and think "Seriously!? All this Just To Split My Mail!?"
>
> Right, much too complicated.  The solution is described in the manual:
>
> ,----[ (info "(gnus)Fancy Mail Splitting") ]
> |    Normally, VALUE in these splits must match a complete _word_
> | according to the fundamental mode syntax table.  In other words, all
> | VALUE's will be implicitly surrounded by `\<...\>' markers, which are
> | word delimiters.  Therefore, if you use the following split, for
> | example,
> | 
> |      (any "joe" "joemail")
> | 
> | messages sent from `joedavis@foo.org' will normally not be filed in
> | `joemail'.  If you want to alter this behavior, you can use any of the
> | following three ways:
> | 
> |   1. You can set the `nnmail-split-fancy-match-partial-words' variable
> |      to non-`nil' in order to ignore word boundaries and instead the
> |      match becomes more like a grep.  This variable controls whether
> |      partial words are matched during fancy splitting.  The default
> |      value is `nil'.
> | 
> |      Note that it influences all VALUE's in your split rules.
> | 
> |   2. VALUE beginning with `.*' ignores word boundaries in front of a
> |      word.  Similarly, if VALUE ends with `.*', word boundaries in the
> |      rear of a word will be ignored.  For example, the VALUE
> |      `"@example\\.com"' does not match `foo@example.com' but
> |      `".*@example\\.com"' does.
> | 
> |   3. You can set the INVERT-PARTIAL flag in your split rules of the
> |      `(FIELD VALUE ...)' types, aforementioned in this section.  If the
> |      flag is set, word boundaries on both sides of a word are ignored
> |      even if `nnmail-split-fancy-match-partial-words' is `nil'.
> |      Contrarily, if the flag is set, word boundaries are not ignored
> |      even if `nnmail-split-fancy-match-partial-words' is non-`nil'.
> |      (New in Gnus 5.10.7)
> `----
>
> Bye, Reiner.

So the question is really "why is this missed by everyone who trys
splitting"?

Can it be simplified? Moved? 

Is there an nnimap-split-fancy-match... equivalent? 

Point 3 is non  parseable by anyone without a degree in elisp.

By, R.

Re: Just shoot me now

[Tommy Kelly]

In response to Reiner pointing out:
>> |   3. You can set the INVERT-PARTIAL flag in your split rules of the
>> |      `(FIELD VALUE ...)' types, aforementioned in this section.  If the
>> |      flag is set, word boundaries on both sides of a word are ignored
>> |      even if `nnmail-split-fancy-match-partial-words' is `nil'.
>> |      Contrarily, if the flag is set, word boundaries are not ignored
>> |      even if `nnmail-split-fancy-match-partial-words' is non-`nil'.
>> |      (New in Gnus 5.10.7)

Richard Riley <rileyrg@googlemail.com> writes:
> Point 3 is non  parseable by anyone without a degree in elisp.

Actually for me the problem is more subtle. *Once* Reiner pointed it out
I could see what was going on. Worse, I'd already seen that part of the
docs. And therefore:

> So the question is really "why is this missed by everyone who trys
> splitting"?

Exactly. So I reckon that in my case at least part of the reason is
that the split documentation is *in general* sufficiently tough-going,
and enough of it is simply obscure to anyone not already familiar with
what it's saying, that I (quite reasonably) end up not trusting it and
not reverting to it even when it would in fact help me. (Hence, for
example, me googling for info on syntax tables rather than going first
to the emacs help). 

Some of the documentation reminds me of signage in London Heathrow
Airport. Once you know where you're going, the signs make perfect
sense. 

But I think if we keep chipping away -- and with occasional assistance
from the ueber-coders (genuflects in Andreas's direction :-p ) -- we'll
come up with some useful improvements. 

Tommy

Re: Just shoot me now

[Sivaram Neelakantan]

On Wed, Dec 22 2010,Tommy Kelly wrote:

[snipped 19 lines]

> Exactly. So I reckon that in my case at least part of the reason is
> that the split documentation is *in general* sufficiently tough-going,
> and enough of it is simply obscure to anyone not already familiar with
> what it's saying, that I (quite reasonably) end up not trusting it and
> not reverting to it even when it would in fact help me. (Hence, for
> example, me googling for info on syntax tables rather than going first
> to the emacs help). 

Funnily, mail splitting worked for me the first time I tried apart
from a few minor misses.  But mine were simple ones

(setq nnmail-split-methods
      '(("sas-l" "^Subject:.*SAS-L*")
;; 	("ctan-ann" "^From:.*ctan-ann*")
 	("mail.misc" "")))

is all I did.


>
> Some of the documentation reminds me of signage in London Heathrow
> Airport. Once you know where you're going, the signs make perfect
> sense. 

Right, you're on the same path as me.  Things don't make sense for a
long time and then suddenly everything falls into place.  Group
Parameters, for some reason took me a long time to get sorted.  But
the documentation is ummm....very clear now.  WhyTF, I didn't
understand it the first time around or for that matter many other
times before, I don't know.

On a few topics, I go through searching-for-lightswitch-in-the-dark
stumbles. 

Read the stuff in the manual and try the cryptic incantation (was Lars
drunk when he wrote that?) ;)
Doesn't work
Google and try again with someone's random code on the same thingy
it works; I read the documentation again
tweak it based on what I *think* is the understanding of what it
should do 
works again
reread the documentation; fcuk, it makes perfect sense now
2 days later,completely forget how I went from noob to semi literate
guy on the topic.  This prevents me with making any constructive
suggestion on how to make the documentation better.

Other than "Documentation is/was not clear" which only serves to
exasperate the developers.

Of course, the fault is mostly mine that I make feeble attempts at
learning elisp.  Hopefully, that and a modicum of understanding of the
English language is all that is needed.


>
> But I think if we keep chipping away -- and with occasional
> assistance from the ueber-coders (genuflects in Andreas's direction
> :-p ) -- we'll come up with some useful improvements.

Look, if you're using Gnus,then you're already tagged as a weirdo in
office. :-)
It's unlikely that a text based mail/news reader is going to win new
converts.  All I got for my personal efforts in using Gnus is just
that; personal satisfaction of working with large number of articles
and mailing lists without getting overloaded.

Not one bloke expressed interest in it apart from

"you still working on DOS?" (white on black Gnus)
"Is that Pine/Unix?" 
"What is a newsreader?"
"why is there no colour like MSLookout?"

What I'm saying is, 6 months into this Gnus, you're likely to think
that documentation is doing just fine. :)

 sivaram
 -- 

Re: Just shoot me now

[Tommy Kelly]

Sivaram Neelakantan <nsivaram.net@gmail.com> writes:
> Funnily, mail splitting worked for me the first time I tried apart
> from a few minor misses.  But mine were simple ones
>
> (setq nnmail-split-methods
>       '(("sas-l" "^Subject:.*SAS-L*")
> ;; 	("ctan-ann" "^From:.*ctan-ann*")
>  	("mail.misc" "")))
>
> is all I did.

Ah but see that's simple splitting. That's for girls (cue slaps for
sexism and threats to get Sacha Chua to kick my ass). Real Men use
*fancy* splitting.  And it's fancy splitting that's destroying my will
to live. :-)

> WhyTF, I didn't
> understand it the first time around or for that matter many other
> times before, I don't know.

Exactly right. It's normal, once you become more expert to forget what
it was like to be a newb.

But a key component of writing good docs, much as in good teaching, is
precisely the ability to rewind your mind to that first time around,
remember what it was like to not know what you now know, and to plot a
path forwards for the newbie. 

Most people find that hard. The few that can are called good
teachers. The very best of those are called Richard Feynman.

> Of course, the fault is mostly mine that I make feeble attempts at
> learning elisp....

Well yes and no. To me it's not unreasonable that if you want to get to
grips with the dark underbelly of Gnus (or anything in emacs) you have
to roll your sleeves up and dig into lisp. But that doesn't mean the
documentation is fine.

Again -- lest Lars, Andreas and the other Noble Writers Of Code get fed
up with what may seem like whining, just keep hacking guys. You are much 
appreciated. 

Tommy

Re: Just shoot me now

[Richard Riley]

Tommy Kelly <tommy.kelly@verilab.com> writes:

> Sivaram Neelakantan <nsivaram.net@gmail.com> writes:
>> Funnily, mail splitting worked for me the first time I tried apart
>> from a few minor misses.  But mine were simple ones
>>
>> (setq nnmail-split-methods
>>       '(("sas-l" "^Subject:.*SAS-L*")
>> ;; 	("ctan-ann" "^From:.*ctan-ann*")
>>  	("mail.misc" "")))
>>
>> is all I did.
>
> Ah but see that's simple splitting. That's for girls (cue slaps for
> sexism and threats to get Sacha Chua to kick my ass). Real Men use
> *fancy* splitting.  And it's fancy splitting that's destroying my will
> to live. :-)
>
>> WhyTF, I didn't
>> understand it the first time around or for that matter many other
>> times before, I don't know.
>
> Exactly right. It's normal, once you become more expert to forget what
> it was like to be a newb.
>
> But a key component of writing good docs, much as in good teaching, is
> precisely the ability to rewind your mind to that first time around,
> remember what it was like to not know what you now know, and to plot a
> path forwards for the newbie. 
>
> Most people find that hard. The few that can are called good
> teachers. The very best of those are called Richard Feynman.
>
>> Of course, the fault is mostly mine that I make feeble attempts at
>> learning elisp....
>
> Well yes and no. To me it's not unreasonable that if you want to get to
> grips with the dark underbelly of Gnus (or anything in emacs) you have
> to roll your sleeves up and dig into lisp. But that doesn't mean the
> documentation is fine.
>
> Again -- lest Lars, Andreas and the other Noble Writers Of Code get fed
> up with what may seem like whining, just keep hacking guys. You are much 
> appreciated. 
>
> Tommy

Abd just in case anyone is reading this, I still havent got an example
of fancy splitting working, Here is my setup:-

,----
|    (setq spam-use-bogofilter t)
|    (setq spam-use-spamassassin t)
|    
|    (require 'spam)
|    (spam-initialize)
|    
|    (setq nnimap-split-inbox '("INBOX"))
|    (setq nnimap-split-fancy '(| (: spam-split) "INBOX"))
| 
|    (setq nnmail-split-inbox '("INBOX"))
|    (setq nnmail-split-fancy '(| (: spam-split) "SYS-INBOX"))
|    
|    (setq spam-split-disabled nil)
|    
|    (setq spam-autodetect-recheck-messages nil)
|    (setq spam-mark-ham-unread-before-move-from-spam-group t)
|    (setq spam-mark-new-messages-in-spam-group-as-spam t)
|    (setq spam-split-group "Gnus-Spam")
`----
   
And no splitting is happening on the INBOX. I cant think what else might
be inhibiting it so if no one can help I'll try, gulp, to debug it...

Re: Just shoot me now

[Tommy Kelly]

Richard Riley <rileyrg@googlemail.com> writes:

> |    (setq nnimap-split-inbox '("INBOX"))
> |    (setq nnimap-split-fancy '(| (: spam-split) "INBOX"))
> ...    
> And no splitting is happening on the INBOX. I cant think what else might
> be inhibiting it so if no one can help I'll try, gulp, to debug it...

Can you get splitting to work *at all* -- i.e. if you remove that
(: spam-split) spit and maybe just have something like:

(setq nnimap-split-fancy '(| "TEMP-GROUP-TO-PROVE-IM-NOT-INSANE"))

does that at least work?

Tommy

Re: Just shoot me now

[Lawrence Mitchell]

Richard Riley wrote:

[...]


> Abd just in case anyone is reading this, I still havent got an example
> of fancy splitting working, Here is my setup:-

> ,----
> |    (setq spam-use-bogofilter t)
> |    (setq spam-use-spamassassin t)
> |
> |    (require 'spam)
> |    (spam-initialize)
> |
> |    (setq nnimap-split-inbox '("INBOX"))

This variable no longer exists in gnus.  As of Lars' recent
nnimap rewrite.

You want:

(setq nnimap-split-methods 'nnimap-split-fancy)
> |    (setq nnimap-split-fancy '(| (: spam-split) "INBOX"))
> |
> |    (setq nnmail-split-inbox '("INBOX"))

Similarly here.

(setq nnmail-split-methods 'nnmail-split-fancy)

> |    (setq nnmail-split-fancy '(| (: spam-split) "SYS-INBOX"))
> |
> |    (setq spam-split-disabled nil)
> |
> |    (setq spam-autodetect-recheck-messages nil)
> |    (setq spam-mark-ham-unread-before-move-from-spam-group t)
> |    (setq spam-mark-new-messages-in-spam-group-as-spam t)
> |    (setq spam-split-group "Gnus-Spam")
> `----

> And no splitting is happening on the INBOX. I cant think what else might
> be inhibiting it so if no one can help I'll try, gulp, to debug it...

The manual hasn't been updated inline with the most recent set of
nnimap changes, maybe?

Cheers,
Lawrence
-- 
Lawrence Mitchell <wence@gmx.li>

Re: Just shoot me now

[Richard Riley]

Lawrence Mitchell <wence@gmx.li> writes:

> Richard Riley wrote:
>
> [...]
>
>> Abd just in case anyone is reading this, I still havent got an example
>> of fancy splitting working, Here is my setup:-
>
>> ,----
>> |    (setq spam-use-bogofilter t)
>> |    (setq spam-use-spamassassin t)
>> |
>> |    (require 'spam)
>> |    (spam-initialize)
>> |
>> |    (setq nnimap-split-inbox '("INBOX"))
>
> This variable no longer exists in gnus.  As of Lars' recent
> nnimap rewrite.
>
> You want:
>
> (setq nnimap-split-methods 'nnimap-split-fancy)
>> |    (setq nnimap-split-fancy '(| (: spam-split) "INBOX"))
>> |
>> |    (setq nnmail-split-inbox '("INBOX"))
>
> Similarly here.
>
> (setq nnmail-split-methods 'nnmail-split-fancy)
>
>> |    (setq nnmail-split-fancy '(| (: spam-split) "SYS-INBOX"))
>> |
>> |    (setq spam-split-disabled nil)
>> |
>> |    (setq spam-autodetect-recheck-messages nil)
>> |    (setq spam-mark-ham-unread-before-move-from-spam-group t)
>> |    (setq spam-mark-new-messages-in-spam-group-as-spam t)
>> |    (setq spam-split-group "Gnus-Spam")
>> `----
>
>> And no splitting is happening on the INBOX. I cant think what else might
>> be inhibiting it so if no one can help I'll try, gulp, to debug it...
>
> The manual hasn't been updated inline with the most recent set of
> nnimap changes, maybe?
>

Ye gods. I was blinded. I reverted changes. I know that
nnimap-split-inbox is history. 

I could cry.

Thank you so much. I'd been looking so long I missed the most obvious
issue. One I only recently raised as a documentation error in the oracle
section of the Gnus manual too!

cheers

r.

Re: Just shoot me now

[Tommy Kelly]

Lawrence Mitchell <wence@gmx.li> writes:

> The manual hasn't been updated inline with the most recent set of
> nnimap changes, maybe?

Certainly what you get from C-h v on nnmail-split-fancy is different
from what's in the current info. It's not hugely different but the C-h v
stuff is slightly better IMO.

Tommy

Re: Just shoot me now

[Richard Riley]

Lawrence Mitchell <wence@gmx.li> writes:

> Richard Riley wrote:
>
> [...]
>
>> Abd just in case anyone is reading this, I still havent got an example
>> of fancy splitting working, Here is my setup:-
>
>> ,----
>> |    (setq spam-use-bogofilter t)
>> |    (setq spam-use-spamassassin t)
>> |
>> |    (require 'spam)
>> |    (spam-initialize)
>> |
>> |    (setq nnimap-split-inbox '("INBOX"))
>
> This variable no longer exists in gnus.  As of Lars' recent
> nnimap rewrite.
>
> You want:
>
> (setq nnimap-split-methods 'nnimap-split-fancy)
>> |    (setq nnimap-split-fancy '(| (: spam-split) "INBOX"))
>> |
>> |    (setq nnmail-split-inbox '("INBOX"))
>
> Similarly here.
>
> (setq nnmail-split-methods 'nnmail-split-fancy)
>
>> |    (setq nnmail-split-fancy '(| (: spam-split) "SYS-INBOX"))
>> |
>> |    (setq spam-split-disabled nil)
>> |
>> |    (setq spam-autodetect-recheck-messages nil)
>> |    (setq spam-mark-ham-unread-before-move-from-spam-group t)
>> |    (setq spam-mark-new-messages-in-spam-group-as-spam t)
>> |    (setq spam-split-group "Gnus-Spam")
>> `----
>
>> And no splitting is happening on the INBOX. I cant think what else might
>> be inhibiting it so if no one can help I'll try, gulp, to debug it...
>
> The manual hasn't been updated inline with the most recent set of
> nnimap changes, maybe?
>
> Cheers,
> Lawrence

ok corrections made. It worked the first time as it did before I
reverted to the faulty config earlier. I enabled spam autodetection for
the group. And the second time I entered the INBOX I got:-

Debugger entered--Lisp error: (wrong-type-argument listp nnimap-split-fancy)
  nnmail-article-group(nnimap-dummy-active-number nil nnimap-save-mail-spec)
  nnmail-check-duplication("<4Oz1cgmcivDcq0rLec2NA9Braic.mv.1293040975222@OMS09.techtargetlists.com>" nnimap-save-mail-spec nnimap-dummy-active-number nnimap-save-mail-spec)
  nnmail-process-mmdf-mail-format(nnimap-save-mail-spec nnimap-dummy-active-number nnimap-save-mail-spec)
  nnmail-split-incoming(#<buffer *nnimap offlineimap nil  *nntpd**<3>> nnimap-save-mail-spec nil nil nnimap-dummy-active-number nnimap-save-mail-spec)
  nnimap-split-incoming-mail()
  nnimap-request-scan(nil "riley")
  gnus-request-scan(nil (nnimap "riley" (nnir-search-engine imap) (nnimap-address "offlineimap")))
  gnus-get-unread-articles(nil)
  gnus-group-get-new-news(nil)
  call-interactively(gnus-group-get-new-news nil nil)

Q: Can someone explain how the group parameters compare to the global
settings such as nnimap-inbox? 

config:-

,----
|    (setq spam-use-bogofilter t)
|    (setq spam-use-spamassassin t)
|    
|    (require 'spam)
|    (spam-initialize)
|    
|    (setq nnimap-inbox '("INBOX"))
|    (setq nnimap-split-fancy '(| (: spam-split) "INBOX"))
|    (setq nnimap-split-methods 'nnimap-split-fancy)
|    (setq nnimap-split-download-body t)
| 
|    (setq nnmail-inbox '("INBOX"))
|    (setq nnmail-split-fancy '(| (: spam-split) "SYS-INBOX"))
|    (setq nnmail-split-methods 'nnmail-split-fancy)
|    
|    (setq spam-split-disabled nil)
|    
|    (setq spam-autodetect-recheck-messages nil)
|    (setq spam-mark-ham-unread-before-move-from-spam-group t)
|    (setq spam-mark-new-messages-in-spam-group-as-spam t)
|    (setq spam-split-group "Gnus-Spam")
`----

Re: Just shoot me now

[Eric S Fraga]

Tommy Kelly <tommy.kelly@verilab.com> writes:

[...]

> Some of the documentation reminds me of signage in London Heathrow
> Airport. Once you know where you're going, the signs make perfect
> sense. 

Although, this past week, signs have been the least of the
problems at Heathrow... ;-)

-- 
Eric S Fraga (GnuPG: 0xC89193D8FFFCF67D)

Re: Just shoot me now (was Re: What's wrong with this fancy split?)

[Richard Riley]

Tommy Kelly <tommy.kelly@verilab.com> writes:

> Andreas Schwab <schwab@linux-m68k.org> writes:
>
>> Because [ and ] are not word characters.  See
>> nnmail-split-fancy-syntax-table.
>
> Thanks Andreas. Unfortunately that 6,500 line variable, consisting
> mostly of pairs of parens enclosing numbers, is as clear as mud :-)
>
> This to no one in particular:
>
> Thus far, like my adventures in figuring out what version of gnus is
> current, my attempts Just To Split My Mail have been ... interesting:
>
> 1. Try to figure out multi-line regexps with simple splitting
> 2. Advised to move to fancy splitting
> 3. Hours of trying to get even the simplest fancy split working (and I'm
> not alone)
> 4. Finally get it moving -- but only after direct advice. But still
> don't understand why (specifically, why is nnmail-split-methods settable
> on a per-server basis, but nnmail-split-fancy is done on a ... shrug
> .. a per gnus basis?)
> 5. Begin building more simple splits then stumble on
> "\\[stuff\\]". Splitting isn't as keen on \\[ as it is on, say, \\. .
> 6. Told that nnmail-split-fancy-syntax-table is the clue
> 7. C-h v on that variable. Look. Puke. Look again.
> 8. Google for "emacs lisp syntax table"
> 9. Fall pretty much at the first fence on being told that a syntax table
> is a "char-table" about which I know very little
> 10. Start reading about what a char-table is
> 11. Consider reading about "aref" and "aset" since part of the basic
> description of char-tables mentions those
> 12. Stop and think "Seriously!? All this Just To Split My Mail!?"
>
> I'm not remotely demanding, requesting or even begging for change. I
> acknowledge that Gnus is a labour of love, open source, and is there to
> be taken or left as each sees fit.
>
> I'm just saying.
>
> Tommy

Heh. This made me smile ;) It's nice to know I'm not alone as someone
who thinks themselves moderately able that also bangs their head against
the wall wondering if I have gone crazy ... ;)

This could be a great start for the new "Make Gnus Accessible" effort!

Re: Just shoot me now (was Re: What's wrong with this fancy split?)

[Andreas Schwab]

Tommy Kelly <tommy.kelly@verilab.com> writes:

> Andreas Schwab <schwab@linux-m68k.org> writes:
>
>> Because [ and ] are not word characters.  See
>> nnmail-split-fancy-syntax-table.
>
> Thanks Andreas. Unfortunately that 6,500 line variable, consisting
> mostly of pairs of parens enclosing numbers, is as clear as mud :-)

Read about modify-syntax-table.

> 8. Google for "emacs lisp syntax table"

Did you know that Emacs comes with a searchable manual?

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: Just shoot me now (was Re: What's wrong with this fancy split?)

[Tommy Kelly]

Andreas Schwab <schwab@linux-m68k.org> writes:

>> 8. Google for "emacs lisp syntax table"
>
> Did you know that Emacs comes with a searchable manual?

Absolutely. I use it often. And as Reiner pointed out, the answer to my
question can in fact be figured out from the manual.

So where does that leave us? 

A user who knows about and uses the searchable manual; who has shown
willing to dive into the code and to extend his searching beyond the
manual; who in general acknowledges that one doesn't use emacs or Gnus
if what you want in your editor is a set of clicky buttons and an
eye-candy-filled GUI, IS STILL HAVING DIFFICULTY.

Either the user is thick, or the documentation is thick.

Thing is -- so let's call a spade a spade -- I so *know* I'm not thick
that I wouldn't offer that as a choice if I thought I was the problem.

It's the documentation. It's fine for you -- but you're not the
audience, or at least not the whole audience.

Again, we're in Open Source land, and Lars, you, and the rest of the
coders are just Cool Dudes to whom I'm grateful for the work you've put
into this thing. I'm not *expecting* anything. 

Look at it this way. I'm merely telling you all this to save you reading
"the code" -- being the enormous literature on psychology, learning
styles, philosophy of language, and, to a good extent, literate
programming.

But, if you prefer, see Tractatus Logic-Philosophicus. ;-)

Tommy

Re: Just shoot me now (was Re: What's wrong with this fancy split?)

[Adam Sjøgren]

On Wed, 22 Dec 2010 06:53:07 -0600, Tommy wrote:

> So where does that leave us? 

It is much easier for a developer deep in the code to receive a
documentation patch, look at it, and say "Ok, that sounds fine,
applied", than it is to receieve a suggestion in words that something -
while technically correct - is slightly confusing to a (new) user.

(E.g. I agree that not escaping '.' in a regexp matching a domain is
confusing to a user who is trying to check his "world-view" by looking
at the examples, but it is also plain to see that it isn't wrong - and a
user thinking 'regular expressions' will recognize that (but then the
clue of "flawless" examples go down the drain)).

What I am trying to say is that I think you will find much less argument
if you dig out git and start supplying your suggestions for improvements
as patches to the manual.

It is quite natural to say "Meh, but it isn't wrong" if you're asking
someone to do... the horror... work!

Appreciating this type of feedback isn't always easy to do cheerfully,
sometimes it will be curmudgeonly.

> Either the user is thick, or the documentation is thick.

Or the user is trying to achieve to many things at once. There is a
learning curve, even if you aren't thick ;-)


  Best regards,

    Adam

-- 
 "Accept the mystery!"                                        Adam Sjøgren
                                                         asjo@koldfront.dk

Re: Just shoot me now (was Re: What's wrong with this fancy split?)

[Tommy Kelly]

asjo@koldfront.dk (Adam Sjøgren) writes:

> On Wed, 22 Dec 2010 06:53:07 -0600, Tommy wrote:
>
>> So where does that leave us? 
>
> It is much easier for a developer deep in the code to receive a
> documentation patch...

Totally understand.

> What I am trying to say is that I think you will find much less argument
> if you dig out git and start supplying your suggestions for improvements
> as patches to the manual.

Absolutely agree.

> It is quite natural to say "Meh, but it isn't wrong" if you're asking
> someone to do... the horror... work!

Yep. So just to be clear. I've already begun (and posted to this group)
the beginnings of a rewrite of the bit in question, and I plan to take
that to completion (if I don't shoot myself in the head in the process
:-) ). Lars has pointed me at the git patch  process and I'm aware of
the need for FSF copyright assignment.  

So my questions at the moment are definitely not veiled demands for
someone to fix things for me. Rather they are mostly requests for
information to allow *me* to do the fixing.

> Or the user is trying to achieve to many things at once. There is a
> learning curve, even if you aren't thick ;-)

It's a good observation. In fact, one of the things that tells me that
no matter what else is an issue, the documentation itself could be
improved, is the extent to which simply to do one thing the newbie needs
to dive into so many other things. Of course some pre-requisite
knowledge really is pre-requisite. You can't do physics without
calculus. But sometimes the lower levels are poking through
needlessly. You *don't* need to understand Goedel's Incompleteness
Theorem to do physics, even though it is profoundly important.

Regardless, in many cases I think a bit of extra wordiness in the
docs will allow the new user to get moving faster, thereby helping their
learning when the time comes for them to dive into the guts and,
conversely, lowering the question burden on the experts. Reuse an' all
that :-)

Tommy

Re: Just shoot me now (was Re: What's wrong with this fancy split?)

[Adam Sjøgren]

On Wed, 22 Dec 2010 13:12:52 -0600, Tommy wrote:

> So my questions at the moment are definitely not veiled demands for
> someone to fix things for me. Rather they are mostly requests for
> information to allow *me* to do the fixing.

Great!

> Of course some pre-requisite knowledge really is pre-requisite. You
> can't do physics without calculus. But sometimes the lower levels are
> poking through needlessly. You *don't* need to understand Goedel's
> Incompleteness Theorem to do physics, even though it is profoundly
> important.

"Just" Emacs is kind of a beast to get into, even just terminology-wise.
It is not an easy balance: How much can you as a manual-writer of Gnus
assume that the reader is familiar with Emacs? With Emacs lisp?

I used to be exceedingly annoyed about examples in Emacs manuals,
because I could never figure the quoting out. I seem to have learned
something by osmosis, because I am not as annoyed any more.

> Regardless, in many cases I think a bit of extra wordiness in the
> docs will allow the new user to get moving faster, thereby helping their
> learning when the time comes for them to dive into the guts and,
> conversely, lowering the question burden on the experts. Reuse an' all
> that :-)

Agreed - I think your patches will be welcomed with open arms :-)


  Best regards,

    Adam

-- 
 "Accept the mystery!"                                        Adam Sjøgren
                                                         asjo@koldfront.dk