gnus-registry-split-fancy-with-parent

gnus-registry-split-fancy-with-parent

[Jake Colman]

I tried using this and gnus crashed out.  What is the correct way use this
function?

gnus-registry is apparently not loaded by default so I 'require' it early on
in my .gnus file (right after I (require 'spam).

My nnimap-split-fancy starts like this:

(setq nnimap-split-fancy
      '(|
        (: gnus-registry-split-fancy-with-parent)
        (: spam-split 'spam-use-regex-headers)
))

Is there something else I should have done?

Do I understand correctly that will scan my mailbox and immediately file this
message together with its parent?  If so, it makes sense to do this as one of
the first checks, correct?

-- 
Jake Colman                     

Principia Partners LLC                    Phone: (201) 209-2467
Harborside Financial Center                 Fax: (201) 946-0320
902 Plaza Two                          E-mail: colman@ppllc.com
Jersey City, NJ 07311                 www.principiapartners.com

Re: gnus-registry-split-fancy-with-parent

[Ted Zlatanov]

On Wed, 29 Oct 2003, colman@ppllc.com wrote:

> I tried using this and gnus crashed out.  What is the correct way
> use this function?
> 
> gnus-registry is apparently not loaded by default so I 'require' it
> early on in my .gnus file (right after I (require 'spam).

That's not enough, you need to do (gnus-registry-initialize).  The
registry won't be a Gnus feature formally until after the current
release is out of beta, that's why it's not in the manual.

Before you do that, can you send me a trace of the problem?  Do M-x
toggle-debug-on-error and you'll get a trace.

> My nnimap-split-fancy starts like this:
> 
> (setq nnimap-split-fancy
>       '(|
>         (: gnus-registry-split-fancy-with-parent)
>         (: spam-split 'spam-use-regex-headers)
> ))
> 
> Is there something else I should have done?

No, this is fine.  You may also want tracking by subject, but that's
optional and experimental.

> Do I understand correctly that will scan my mailbox and immediately
> file this message together with its parent?  If so, it makes sense
> to do this as one of the first checks, correct?

Usually, yes, this will group threads together.  It doesn't work with
some MUAs that don't preserve the References header, though.

Ted

Re: gnus-registry-split-fancy-with-parent

[Jake Colman]

>>>>> "TZ" == Ted Zlatanov <tzz@lifelogs.com> writes:

   TZ> On Wed, 29 Oct 2003, colman@ppllc.com wrote:
   >> I tried using this and gnus crashed out.  What is the correct way
   >> use this function?
   >> 
   >> gnus-registry is apparently not loaded by default so I 'require' it
   >> early on in my .gnus file (right after I (require 'spam).

   TZ> That's not enough, you need to do (gnus-registry-initialize).  The
   TZ> registry won't be a Gnus feature formally until after the current
   TZ> release is out of beta, that's why it's not in the manual.

I should do (gnus-registry-initialize) right after I 'require' it, correct?

   TZ> Before you do that, can you send me a trace of the problem?  Do M-x
   TZ> toggle-debug-on-error and you'll get a trace.

I tried this but it doesn't want to crash anymore.  Yes, I tried it before I
added (gnus-registry-initialize).  If I see any more errors I'll post a
backtrace.

When the registry is initilalized, it tracks where all articles have been
filed?  And that's what allows (gnus-registry-split-fancy-with-parent) to
know what to do?  Why doesn't that cause alot of unacceptable overhead?

   >> My nnimap-split-fancy starts like this:
   >> 
   >> (setq nnimap-split-fancy
   >> '(|
   >> (: gnus-registry-split-fancy-with-parent)
   >> (: spam-split 'spam-use-regex-headers)
   >> ))
   >> 
   >> Is there something else I should have done?

   TZ> No, this is fine.  You may also want tracking by subject, but that's
   TZ> optional and experimental.

How would one do that?

   >> Do I understand correctly that will scan my mailbox and immediately
   >> file this message together with its parent?  If so, it makes sense
   >> to do this as one of the first checks, correct?

   TZ> Usually, yes, this will group threads together.  It doesn't work with
   TZ> some MUAs that don't preserve the References header, though.

But nothiung bad would happen, right?  It would just not be split where I
might have wanted it go, right?

I think the latest gnus package (as available through XEmacs) is 5.10.2.
Will the registry be part of 5.10.3?  Will there be a 5.10.3?

-- 
Jake Colman                     

Principia Partners LLC                    Phone: (201) 209-2467
Harborside Financial Center                 Fax: (201) 946-0320
902 Plaza Two                          E-mail: colman@ppllc.com
Jersey City, NJ 07311                 www.principiapartners.com

Re: gnus-registry-split-fancy-with-parent

[Ted Zlatanov]

On Wed, 29 Oct 2003, colman@ppllc.com wrote:

>>>>>> "TZ" == Ted Zlatanov <tzz@lifelogs.com> writes:
> 
>    TZ> On Wed, 29 Oct 2003, colman@ppllc.com wrote:
>    >> I tried using this and gnus crashed out.  What is the correct
>    >> way use this function?
>    >> 
>    >> gnus-registry is apparently not loaded by default so I
>    >> 'require' it early on in my .gnus file (right after I (require
>    >> 'spam).
> 
>    TZ> That's not enough, you need to do (gnus-registry-initialize).
>    TZ> The registry won't be a Gnus feature formally until after the
>    TZ> current release is out of beta, that's why it's not in the
>    TZ> manual.
> 
> I should do (gnus-registry-initialize) right after I 'require' it,
> correct?

I'm pretty sure you don't need to 'require' anything because of the
magic of autoloading, and it works that way for me.

> When the registry is initilalized, it tracks where all articles have
> been filed?  And that's what allows
> (gnus-registry-split-fancy-with-parent) to know what to do?  

Yes.

> Why doesn't that cause alot of unacceptable overhead?

Well, you can look at the source if you are interested in that level
of detail.  Basically we keep a fast index of the article IDs and the
groups they are in.

>    TZ> No, this is fine.  You may also want tracking by subject, but
>    TZ> that's optional and experimental.
> 
> How would one do that?

Do M-x customize-group gnus-registry and you'll see that option and
some others.  Subject tracking is in CVS only, I think.

>    TZ> Usually, yes, this will group threads together.  It doesn't
>    TZ> work with some MUAs that don't preserve the References
>    TZ> header, though.
> 
> But nothiung bad would happen, right?  It would just not be split
> where I might have wanted it go, right?

Right, it will go to the next splitting rule.

> I think the latest gnus package (as available through XEmacs) is
> 5.10.2.  Will the registry be part of 5.10.3?  Will there be a
> 5.10.3?

I'm not sure about the exact release plans.  In the worst case, you
can always get just gnus-registry.el from CVS - it doesn't depend on
anything and nothing depends on it.  I think it will be a part of
Gnus, though.

Ted

Re: gnus-registry-split-fancy-with-parent

[Jake Colman]

(gnus-registry-initialize) is not part of the Gnus 5.10.2 package in
XEmacs. Since I am going to want your support, I decided to bite the bullet
and went with the latest cvs gnus.  This way, things like (spam-initialize)
and (gnus-registry-initialize) all work as you say.

Does (gnus-registry-initialize) have to be executed before (spam-initialize)
or is there no interdependency?  BTW is it required that I "(setq
spam-use-bogofilter t)" before doing (spam-initialize)?

Using gnus-registry-split-fancy-with-parent I am seeing messages "traced
... to group nil".  Why is it never finding a group to put it in?  I'm
assuming that this test, when it works, will split faster than forcing the
email through all my 'any' splits.  Is this correct?

>>>>> "TZ" == Ted Zlatanov <tzz@lifelogs.com> writes:

   TZ> On Wed, 29 Oct 2003, colman@ppllc.com wrote:
   >>>>>>> "TZ" == Ted Zlatanov <tzz@lifelogs.com> writes:
   >> 
   TZ> On Wed, 29 Oct 2003, colman@ppllc.com wrote:
   >> >> I tried using this and gnus crashed out.  What is the correct
   >> >> way use this function?
   >> >> 
   >> >> gnus-registry is apparently not loaded by default so I
   >> >> 'require' it early on in my .gnus file (right after I (require
   >> >> 'spam).
   >> 
   TZ> That's not enough, you need to do (gnus-registry-initialize).
   TZ> The registry won't be a Gnus feature formally until after the
   TZ> current release is out of beta, that's why it's not in the
   TZ> manual.
   >> 
   >> I should do (gnus-registry-initialize) right after I 'require' it,
   >> correct?

   TZ> I'm pretty sure you don't need to 'require' anything because of the
   TZ> magic of autoloading, and it works that way for me.

   >> When the registry is initilalized, it tracks where all articles have
   >> been filed?  And that's what allows
   >> (gnus-registry-split-fancy-with-parent) to know what to do?  

   TZ> Yes.

   >> Why doesn't that cause alot of unacceptable overhead?

   TZ> Well, you can look at the source if you are interested in that level
   TZ> of detail.  Basically we keep a fast index of the article IDs and the
   TZ> groups they are in.

   TZ> No, this is fine.  You may also want tracking by subject, but
   TZ> that's optional and experimental.
   >> 
   >> How would one do that?

   TZ> Do M-x customize-group gnus-registry and you'll see that option and
   TZ> some others.  Subject tracking is in CVS only, I think.

   TZ> Usually, yes, this will group threads together.  It doesn't
   TZ> work with some MUAs that don't preserve the References
   TZ> header, though.
   >> 
   >> But nothiung bad would happen, right?  It would just not be split
   >> where I might have wanted it go, right?

   TZ> Right, it will go to the next splitting rule.

   >> I think the latest gnus package (as available through XEmacs) is
   >> 5.10.2.  Will the registry be part of 5.10.3?  Will there be a
   >> 5.10.3?

   TZ> I'm not sure about the exact release plans.  In the worst case, you
   TZ> can always get just gnus-registry.el from CVS - it doesn't depend on
   TZ> anything and nothing depends on it.  I think it will be a part of
   TZ> Gnus, though.

   TZ> Ted

-- 
Jake Colman                     

Principia Partners LLC                    Phone: (201) 209-2467
Harborside Financial Center                 Fax: (201) 946-0320
902 Plaza Two                          E-mail: colman@ppllc.com
Jersey City, NJ 07311                 www.principiapartners.com

Re: gnus-registry-split-fancy-with-parent

[Ted Zlatanov]

On Thu, 30 Oct 2003, colman@ppllc.com wrote:

> Does (gnus-registry-initialize) have to be executed before
> (spam-initialize) or is there no interdependency?  

They are completely unrelated.

> BTW is it required that I "(setq spam-use-bogofilter t)" before
> doing (spam-initialize)?

No, you can set it at any time when you're using spam-initialize.

> Using gnus-registry-split-fancy-with-parent I am seeing messages
> "traced ... to group nil".  Why is it never finding a group to put
> it in?  I'm assuming that this test, when it works, will split
> faster than forcing the email through all my 'any' splits.  Is this
> correct?

Yes.  You can turn up the Gnus verbosity and you'll see messages like

gnus-registry-split-fancy-with-parent traced <v9ptgfpy12.fsf@marauder.physik.uni-ulm.de> to group ding

Registry: article <4nznfividq.fsf@lockgroove.bwh.harvard.edu> spooled to ding

Registry: Registering article 4219 with group sys

You can also do M-: (gnus-registry-fetch-group "x") where "x" is the
Message-ID of a recently received message.  You should get back the
group that the message is in.

Finally, you can look at ~/.gnus.registry.eld and see if it does
indeed hold valid data.

Please don't top-post.

Thanks
Ted

Re: gnus-registry-split-fancy-with-parent

[Jake Colman]

>>>>> "TZ" == Ted Zlatanov <tzz@lifelogs.com> writes:

   >> Using gnus-registry-split-fancy-with-parent I am seeing messages
   >> "traced ... to group nil".  Why is it never finding a group to put
   >> it in?  I'm assuming that this test, when it works, will split
   >> faster than forcing the email through all my 'any' splits.  Is this
   >> correct?

   TZ> Yes.  You can turn up the Gnus verbosity and you'll see messages like

   TZ> gnus-registry-split-fancy-with-parent traced <v9ptgfpy12.fsf@marauder.physik.uni-ulm.de> to group ding

   TZ> Registry: article <4nznfividq.fsf@lockgroove.bwh.harvard.edu> spooled to ding

   TZ> Registry: Registering article 4219 with group sys

But that's just it.  I am only seeing "traced ... to group nil".  It never
seems to trace it to a particular group.

   TZ> You can also do M-: (gnus-registry-fetch-group "x") where "x" is the
   TZ> Message-ID of a recently received message.  You should get back the
   TZ> group that the message is in.

M-: executes eval-expression.  How can I use it check the registry.

   TZ> Finally, you can look at ~/.gnus.registry.eld and see if it does
   TZ> indeed hold valid data.

It does have data in it.  It looks valid to me.

   TZ> Please don't top-post.

Sorry.

   TZ> Thanks

I have not said it yet but I really appreciate the time and effort you are
spending across my various posts.  I tend to have many questions as I strive
to master new techniques and I appreciate your patience.

-- 
Jake Colman                     

Principia Partners LLC                    Phone: (201) 209-2467
Harborside Financial Center                 Fax: (201) 946-0320
902 Plaza Two                          E-mail: colman@ppllc.com
Jersey City, NJ 07311                 www.principiapartners.com

Re: gnus-registry-split-fancy-with-parent

[Ted Zlatanov]

On Thu, 30 Oct 2003, colman@ppllc.com wrote:

>>>>>> "TZ" == Ted Zlatanov <tzz@lifelogs.com> writes:
> 
>    TZ> You can also do M-: (gnus-registry-fetch-group "x") where "x"
>    TZ> is the Message-ID of a recently received message.  You should
>    TZ> get back the group that the message is in.
> 
> M-: executes eval-expression.  How can I use it check the registry.

Hit M-:, then at the prompt, type (gnus-registry-fetch-group "x") and
hit ENTER.  You'll have to get "x" from the message ID of a message
you know was registered.

>    TZ> Finally, you can look at ~/.gnus.registry.eld and see if it
>    TZ> does indeed hold valid data.
> 
> It does have data in it.  It looks valid to me.

Something is strange then, we'll figure it out in time.

> I have not said it yet but I really appreciate the time and effort
> you are spending across my various posts.  I tend to have many
> questions as I strive to master new techniques and I appreciate your
> patience.

No problem at all.

Ted