new marks code is breaking a pre_marks setup

new marks code is breaking a pre_marks setup

[Harry Putnam]

First let me ask, what is the full purpose of the new `marks' code.
I didn't really follow all the discussion.

Now:
Here is one circumstance where something involving the new `Marks'
stuff breaks an exisiting setup, that worked error free under pre
marks Oort.

I have one unusal nnml group that works as a preview of incoming mail
before splitting.  Like this:

Procmail sorts some mail to a few specific procmail spools and whats
left goes to ~/spool/prinb.in . When gnus scans the
~/spool/*.in files, that mail goes to nnml:prinb.  

There I review it and may respond to some messages, delete some, etc
and then call respool (`B r') on whats left, which sends it on to
final resting places in whatever groups.

May sound a little convoluted, but I find it handier than tracking
down all new messages once they've been split asunder.  I usually know
where some things are from, or what needs attention at a glance even
if there are 120 messages in the buffer, (usually more like 50 at
once) and the pile of secondary mailing list interests gets respooled
for later.  I may receive 250-300 messages in a day but usually not
more than 50-70 at once since scanning happens every 5 minutes, and
usually less than 10 or so.

Once the respool is done the group is then empty until next mail scan.

Well, something about the new marks stuff causes a situation where
gnus won't let me in prinb after having been there and wanting to
reenter.

It gives this error:
wrong-type-argument listp (3541 . 3575)

I thought maybe I could just rm -f .marks .overview on that group and
say  nnml-generate-nov-databases-1 ~/Mail/prinb would fix it by
letting gnus regenerate nov and letting the marks bootstrap thingy
take care of .marks.

It didn't work.  Still get the message.  Here is the backtrace

Debugger entered--Lisp error: (wrong-type-argument listp (3541 . 3575))
  delq(3031 (3541 . 3575))

  gnus-update-missing-marks((3031 3032 3033 3034 3035 3036 3037 3038
  3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051
  3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064
  3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077
  3078 3079 3080 ...))

  gnus-select-newsgroup("nnml:prinb" nil nil)
  gnus-summary-read-group-1("nnml:prinb" nil t nil nil nil)
  gnus-summary-read-group("nnml:prinb" nil t nil nil nil nil)
  gnus-group-read-group(nil t)
  gnus-group-select-group(nil)
  call-interactively(gnus-group-select-group)

========================================

$ ls Mail/prinb

  3541 3545 3549 3553 3557 3561 3565 3569 3573 3577 3581 3585 3589 3542
  3546 3550 3554 3558 3562 3566 3570 3574 3578 3582 3586 3543 3547 3551
  3555 3559 3563 3567 3571 3575 3579 3583 3587 3544 3548 3552 3556 3560
  3564 3568 3572 3576 3580 3584 3588

Re: new marks code is breaking a pre_marks setup

[Harry Putnam]

Harry Putnam <reader@newsguy.com> writes:

> Well, something about the new marks stuff causes a situation where
> gnus won't let me in prinb after having been there and wanting to
> reenter.

Seems to have broken `M-c' (`gnus-group-clear-data') too.

Re: new marks code is breaking a pre_marks setup

[Lars Magne Ingebrigtsen]

Harry Putnam <reader@newsguy.com> writes:

> First let me ask, what is the full purpose of the new `marks' code.
> I didn't really follow all the discussion.

[...]

> Well, something about the new marks stuff causes a situation where
> gnus won't let me in prinb after having been there and wanting to
> reenter.
>
> It gives this error:
> wrong-type-argument listp (3541 . 3575)

No, this isn't related to Simon's new nnml marks code.  This is my new
seen/unseen code.  :-)

It's fixed in CVS now.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: new marks code is breaking a pre_marks setup

[Kai Großjohann]

Harry Putnam <reader@newsguy.com> writes:

> First let me ask, what is the full purpose of the new `marks' code.
> I didn't really follow all the discussion.

With the new marks code, you can move a directory like
~/Mail/mail/misc from one machine to the other and the marks will go
along.  Or you can save some of your mail groups on CD and then access
them later on, keeping all the marks and stuff.

Previously, all the marks were in .newsrc.eld, and it wasn't easy to
back up only that part of it which was relevant to a specific group.

kai
-- 
Symbol's function definition is void: signature

Re: new marks code is breaking a pre_marks setup

[Colin Marquardt]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> Harry Putnam <reader@newsguy.com> writes:
>
> With the new marks code, you can move a directory like
> ~/Mail/mail/misc from one machine to the other and the marks will go
> along.  Or you can save some of your mail groups on CD and then access
> them later on, keeping all the marks and stuff.
>
> Previously, all the marks were in .newsrc.eld, and it wasn't easy to
> back up only that part of it which was relevant to a specific group.

Cool. So SOUP is not really necessary anymore (not to mention that I
never got it to work), or its function could change to only compress
the stuff...

Colin

Re: new marks code is breaking a pre_marks setup

[Daniel Pittman]

On Wed, 22 Aug 2001, Colin Marquardt wrote:
> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
> 
>> Harry Putnam <reader@newsguy.com> writes:
>>
>> With the new marks code, you can move a directory like
>> ~/Mail/mail/misc from one machine to the other and the marks will go
>> along. Or you can save some of your mail groups on CD and then access
>> them later on, keeping all the marks and stuff.
>>
>> Previously, all the marks were in .newsrc.eld, and it wasn't easy to
>> back up only that part of it which was relevant to a specific group.
> 
> Cool. So SOUP is not really necessary anymore (not to mention that I
> never got it to work), or its function could change to only compress
> the stuff...

Actually, there are presumably still a few die-hard Fido Technology
Network users out there who use SOUP for access to Internet email and
all; I know that I did for a while when I was dealing primarily with FTN
mail...

...and SOUP keeps that information it's own way and interoperates with
other tools that speak it.[1] :)

        Daniel


Footnotes: 
[1]  Well, I know of three -- the SOUP packager I had, GoldEd and Gnus. :)

-- 
You come for me now with a cake that you've made
Ravaged avenger with a clip in your hair
Full of glass and bleach and my old razorblades
O where do we go now but nowhere
        -- Nick Cave, _Where Do We Go Now But Nowhere?_

Re: new marks code is breaking a pre_marks setup

[Kai Großjohann]

Colin Marquardt <colin.marquardt@usa.alcatel.com> writes:

> Cool. So SOUP is not really necessary anymore (not to mention that I
> never got it to work), or its function could change to only compress
> the stuff...

SOUP can download news, too.

kai
-- 
Symbol's function definition is void: signature