miscellaneous Gnus problems/questions

miscellaneous Gnus problems/questions

[Joe Wells]

GNU Emacs 19.31.1 (sparc-sun-solaris2.4, X toolkit) of Wed Jun  5 1996 on csb
Gnus v5.1; nntp 4.0; nnml 1.0; nnmh 1.0; nndir 1.0; nndoc 1.0

This is a list of random assorted problems with Gnus and questions about
Gnus.  I am using the version of Gnus that comes with the latest Emacs, so
these may be known and/or fixed already.

1.  In the summary buffer, "X v u" seems to be refetching the article even
    if it is already in the article buffer.  Why?

2.  Using "o" on a pseudo-article in an attempt to save it doesn't seem to
    work.  I could use "X U" on the original article, but this downloads
    the article from the news server and runs uudecode on it again,
    wasting a lot of time.  I can also switch to the temporary directory
    and copy the file using the shell, but this is also time-consuming.

3.  It would be nice if there were an option to create the pseudo-article
    in the summary *AND* to immediately view it.

4.  What does the name "eething" mean?  Why "ee"?

5.  In summary, "I ?" splits the windows really funny.  Pressing C-g to abort
    does not clean up window split.  Only proceeding all the way through
    cleans up the windows.

6.  The help text for "I a ?" lists match types that are not legal in
    combination with "a".  It also has a duplicate description for match
    type "r".

7.  For an "eval" keyword in a score file, how often is the form evaluated?
    Once per entering into the group?  Before or after the group select
    hook?  This needs some documentation.

8.  The "P" command at beginning of buffer does not work in a group read via
    the "nndir" back end which uses a directory reached via ange-ftp.
    This is because it calls gnus-read-header which eventually ends up
    calling insert-file-contents-literally which disables ange-ftp.  I
    encountered this problem reading:

      nndir+/ftp@ftp.hpc.uh.edu:/pub/emacs/ding-list-recent/:ding.recent

9.  nnmail-request-post-buffer ignores its follow-to argument if its post
    argument is non-nil.  This happens when using the "a" command in a
    mail group.  This prevents the user's setting of
    gnus-followup-to-function from having any effect for mail groups.  The
    same problem also seems to occur with nndir groups.

    (Yes, I know that I can get the desired result by setting a group
    parameter.  This is what I have done.)

10. In the summary buffer, if you use "x" then "T T", the removed article
    lines come back.

12. nnmail-split-it adds "\\<" and "\\>" around the patterns the user
    supplies in nnmail-split-fancy.  This breaks any patterns which don't
    begin and end at word boundaries.  This is undocumented.  It would be
    nice if this was optional.

13. "*" has word syntax in the article buffer.  This interacts with the
    use of nnmail-split-fancy.  This needs to be documented, because it
    can cause very confusing results.

14. It would be nice if threading did not use garbage subjects like
    "(none)", "hi", or "hello".  Is there a garbage-subject regexp which I
    can set to tell the threading to ignore those subjects?

15. A bug in nnmail-split it: "\\>\\)" should be "\\)\\>".

16. Editing a SCORE file directly doesn't get noticed.  What does one have
    to do to get the changes noticed by Gnus?

17. gnus-followup-to-function can be called with a argument of nil when
    the "a" command is used in the group buffer.  This should be
    documented.  Even better, use the group whose line point is on.

18. I have noticed when reading the group

      nndir+/ftp@ftp.hpc.uh.edu:/pub/emacs/ding-list-recent/:ding.recent

    that some of the read marks often get lost when I revisit the group.
    I can't figure out why this is happening.

-- 
Joe Wells <jbw@cs.bu.edu>

Re: miscellaneous Gnus problems/questions

[Jason L Tibbitts III]

>>>>> "JW" == Joe Wells <jbw@cs.bu.edu> writes:

JW> This is a list of random assorted problems with Gnus and questions
JW> about Gnus.

I'll give it a whirl.

JW> I am using the version of Gnus that comes with the latest Emacs, so
JW> these may be known and/or fixed already.

You should look at 5.2.x.  It's nice.  It's not just a little bug-fix
release, either.  There's major new stuff.

JW> 1.  In the summary buffer, "X v u" seems to be refetching the article
JW> even if it is already in the article buffer.  Why?

I believe it's to get a pristine copy of the article, before any
destructive decoding has been done.  5.2.x saves the original article
in a separate buffer, and can also store copies of the last n articles.  In
any case, it no longer refetches the article.

JW> 2.  Using "o" on a pseudo-article in an attempt to save it doesn't seem
JW> to work.

This seems to work in 5.2.x.

JW> 3.  It would be nice if there were an option to create the
JW> pseudo-article in the summary *AND* to immediately view it.

Hmmm.  You might be able to do something with
gnus-uu-grabbed-file-functions.  You could probably also write it in a few
lines of code.  I can see why this isn't in there: there are way too many
key bindings as it is.  But hey, Red gnus will start up soon.  Bug Lars and
get it on the ToDo list.

JW> 4.  What does the name "eething" mean?  Why "ee"?

We couldn't think of a better name.  Really.  Lars probably has some
rationalization made up just for this occasion, though.

JW> 5.  In summary, "I ?" splits the windows really funny.  Pressing C-g to
JW> abort does not clean up window split.  Only proceeding all the way
JW> through cleans up the windows.

Still does.  This could be considered a bug.  A non-recognized key will
escape and clean the display, though.

JW> 6.  The help text for "I a ?" lists match types that are not legal in
JW> combination with "a".  It also has a duplicate description for match
JW> type "r".

Still does.  Most likely a left-out feature (have to write bunches of help
texts to fix it).

JW> 7.  For an "eval" keyword in a score file, how often is the form
JW> evaluated?  Once per entering into the group?  Before or after the
JW> group select hook?  This needs some documentation.

I agree that it's poorly documented.  The documentation for
gnus-score-alist says:

eval:    Sexp to be evaluated when the score file is loaded.

>From the code, gnus-possible-score-headers, which eventually loads the
score file, is called soon after gnus-select-group-hook is run.

JW> 8.  The "P" command at beginning of buffer does not work in a group
JW> read via the "nndir" back end which uses a directory reached via
JW> ange-ftp.

This seems to work fine in 5.2.x.  (Just tried it.  Getting to
ding-list-recent is quick from my machine!)

JW> 9.  nnmail-request-post-buffer ignores its follow-to argument if its
JW> post argument is non-nil.  This happens when using the "a" command in a
JW> mail group.  This prevents the user's setting of
JW> gnus-followup-to-function from having any effect for mail groups.

nnmail-request-post-buffer doesn't exist anymore; the entire mail-post
dichotomy has gone away.

JW> 10. In the summary buffer, if you use "x" then "T T", the removed
JW> article lines come back.

Seems to work in 5.2.x.

[Can't comment on splitting since I use Procmail.]

JW> 14. It would be nice if threading did not use garbage subjects like
JW> "(none)", "hi", or "hello".  Is there a garbage-subject regexp which I
JW> can set to tell the threading to ignore those subjects?

`gnus-summary-gather-exclude-subject'
     Since loose thread gathering is done on subjects only, that might
     lead to many false hits, especially with certain common subjects
     like `' and `(none)'.  To make the situation slightly better, you
     can use the regexp `gnus-summary-gather-exclude-subject' to say
     what subjects should be excluded from the gathering process.  The
     default is `^ *$\\|^(none)$'.

Also, `gnus-summary-thread-gathering-function' can be use to do gathering
by the references header instead of by the subject.

JW> 16. Editing a SCORE file directly doesn't get noticed.  What does one
JW> have to do to get the changes noticed by Gnus?

`V R'
     Run the current summary through the scoring process
     (`gnus-summary-rescore').  This might be useful if you're playing
     around with your score files behind Gnus' back and want to see the
     effect you're having.

JW> 17. gnus-followup-to-function can be called with a argument of nil when
JW> the "a" command is used in the group buffer.  This should be
JW> documented.  Even better, use the group whose line point is on.

Hmmm.  gnus-followup-to-function doesn't exist anymore.  Ah, it's
message-followup-to-function, which seems to get called with no arguments.
I'm not quire sure how this works.

JW> 18. I have noticed when reading the group

JW> nndir+/ftp@ftp.hpc.uh.edu:/pub/emacs/ding-list-recent/:ding.recent

That's me.

JW>     that some of the read marks often get lost when I revisit the
JW> group.  I can't figure out why this is happening.

Odd.  Nobody else has reported it anywhere I've seen.  I don't renumber the
articles, just delete the old ones, so I can't say.  Anyone else seen this
one?  If upgrading doesn't cure it, let me know.
-- 
      Jason L. Tibbitts III - tibbs@uh.edu - 713/743-8684 - 221SR1
System Manager:  University of Houston High Performance Computing Center
                1994 PC800 "Kuroneko"      DoD# 1723

Re: miscellaneous Gnus problems/questions

[Per Abrahamsen]

>>>>> "JW" == Joe Wells <jbw@cs.bu.edu> writes:

JW> 4.  What does the name "eething" mean?  Why "ee"?

When I try to pronounce `nneething' is sounds just like `anything'.

JW> 12. nnmail-split-it adds "\\<" and "\\>" around the patterns the user
JW>     supplies in nnmail-split-fancy.  This breaks any patterns which don't
JW>     begin and end at word boundaries.  This is undocumented.  

No, it is documented in the section about fancy mail splitting:

Gnus>   In these splits, FIELD must match a complete field name.  VALUE must
Gnus> match a complete word according to the fundamental mode syntax table.
Gnus> You can use `.*' in the regexps to match partial field names or words.

JW>     It would be nice if this was optional.

I think `.*' is easy enough to add the few places it is needed.

JW> 13. "*" has word syntax in the article buffer.  

It has symbol syntax in Gnus 5.2.

JW> 15. A bug in nnmail-split it: "\\>\\)" should be "\\)\\>".

I don't see how that can make a difference.

JW> 16. Editing a SCORE file directly doesn't get noticed.  What does one have
JW>     to do to get the changes noticed by Gnus?

gnus-score-flush-cache

Re: miscellaneous Gnus problems/questions

[Joe Wells]

In article <rjhgsho4cl.fsf@kleene.dina.kvl.dk> Per Abrahamsen <abraham@dina.kvl.dk> writes:

   JW> 15. A bug in nnmail-split it: "\\>\\)" should be "\\)\\>".

   I don't see how that can make a difference.

Suppose the user-supplied pattern is "foo\\|bar".  Inside nnmail-split-it,
this becomes "...\\<\\(foo\\|bar\\>\\)".  The "\\>" now only applies to the
"bar" alternative.

-- 
Joe Wells <jbw@cs.bu.edu>

Re: miscellaneous Gnus problems/questions

[Per Abrahamsen]

>>>>> "JW" == Joe Wells <jbw@cs.bu.edu> writes:

JW> 15. A bug in nnmail-split it: "\\>\\)" should be "\\)\\>".

>>>>> "PA" ==  Per Abrahamsen <abraham@dina.kvl.dk> writes:

PA>    I don't see how that can make a difference.

JW> Suppose the user-supplied pattern is "foo\\|bar".  Inside nnmail-split-it,
JW> this becomes "...\\<\\(foo\\|bar\\>\\)".  The "\\>" now only applies to the
JW> "bar" alternative.

You are right, it is a bug.

Re: miscellaneous Gnus problems/questions

[Ken Olstad]

JW == Joe Wells
Jason == Jason L Tibbitts III

JW> 4.  What does the name "eething" mean?  Why "ee"?

Jason> We couldn't think of a better name.  Really.  Lars probably has
Jason> some rationalization made up just for this occasion, though.

It must be pronounced "anything", eh?
--
Ken Olstad <ken@mn.cheyenne.com>                      Burnsville, MN, USA

Re: miscellaneous Gnus problems/questions

[Joe Wells]

In article <ufa4tohpnic.fsf@sina.hpc.uh.edu> Jason L Tibbitts III <tibbs@hpc.uh.edu> writes:

   JW> 6.  The help text for "I a ?" lists match types that are not legal in
   JW> combination with "a".  It also has a duplicate description for match
   JW> type "r".

   Still does.  Most likely a left-out feature (have to write bunches of help
   texts to fix it).

Just write one piece of help text for each letter and then insert the help
texts in the help buffer only for the letters that apply.

   JW> 18. I have noticed when reading the group

   JW> nndir+/ftp@ftp.hpc.uh.edu:/pub/emacs/ding-list-recent/:ding.recent

   That's me.

   JW>     that some of the read marks often get lost when I revisit the
   JW> group.  I can't figure out why this is happening.

   Odd.  Nobody else has reported it anywhere I've seen.  I don't renumber the
   articles, just delete the old ones, so I can't say.  Anyone else seen this
   one?  If upgrading doesn't cure it, let me know.

I seem to notice this happening when I switch from home to my office.  I
never exit from Emacs; I leave it running for months.  At home, I use an
Emacs running under iScreen.  At my office, I use an Emacs connected to my
X display.  When I leave my office (or home), I quit Gnus ("q" in *Group*
buffer).  Then I restart Gnus (M-x gnus) when I get home (or to my
office).

-- 
Joe Wells <jbw@cs.bu.edu>

Re: miscellaneous Gnus problems/questions

[Jason L Tibbitts III]

>>>>> "JW" == Joe Wells <jbw@cs.bu.edu> writes:

[Archive group]
JW> that some of the read marks often get lost when I revisit the group.  I
JW> can't figure out why this is happening.

JW> I seem to notice this happening when I switch from home to my office.
JW> I never exit from Emacs; I leave it running for months.

I do the same, so I doubt that's it.  (I'd love fancy buttons, colors,
x-faces and smiley icons, but I prefer speed.)  I do recall that this was a
bug in earlier versions of sgnus (though I could be wrong); perhaps it was
also in 5.[0.x|1].

Oops, there's Lars with yet another version.  We will not stop until we can
run emacs in gnus!  Hmmmm...
-- 
      Jason L. Tibbitts III - tibbs@uh.edu - 713/743-8684 - 221SR1
System Manager:  University of Houston High Performance Computing Center
                1994 PC800 "Kuroneko"      DoD# 1723

Re: miscellaneous Gnus problems/questions

[Kai Grossjohann]

>>>>> On Wed, 12 Jun 1996 00:11:20 -0400, jbw@cs.bu.edu (Joe Wells)
>>>>> said:

  Joe> 4.  What does the name "eething" mean?  Why "ee"?

All names begin with `nn', so `nething' was out of the question, and
would have been pronounced differently.  `nneetthhiinngg' would is TOO
strange, don't you think?

Anyhow, I can't think of a better way to spell `anything' if the
prefix `nn' is a given, can you?

kai "bowing to Lars, the Creative"
-- 
Life is hard and then you die.