Re: message-shorten-references hardwired 21 references

Re: message-shorten-references hardwired 21 references

[Reiner Steib]

[ Redirecting to ding@gnus ...]

On Wed, Jan 04 2006, Dan Jacobson wrote:

> There should be a variable to allow users to have
> message-shorten-references use less than the hardwired 21 references.

Please elaborate why you think this would be useful.

> Also are you sure you want the oldest 21 references, not the latest?

I'm quite sure that Gnus does the right thing.  See section
"6.5. References" in [1].

> I want to just include the single very last reference but setting
> maxcount to 1 gives just the earliest.

Mail or news?

Bye, Reiner.

[1] http://www.karlsruhe.org/rfc/son1036.txt
    (expired, but AFAIK still best practice WRT shorting references) 
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: message-shorten-references hardwired 21 references

[Dan Jacobson]

>> There should be a variable to allow users to have
>> message-shorten-references use less than the hardwired 21 references.

R> Please elaborate why you think this would be useful.

I want full control of my references. No "mother knows best". I feel
ashamed to send out bloated headers. I want to have tidy tiny spartan
headers.

>> Also are you sure you want the oldest 21 references, not the latest?

R> I'm quite sure that Gnus does the right thing.  See section
R> "6.5. References" in [1].

    It is permissible to not include the entire previous "References"
    line if it is too long.  An attempt should be made to include a
    reasonable number of backwards references.

He just says "reasonable". I say that the newest 21 is the only way to
keep threads intact, vs. the oldest 21, no?

>> I want to just include the single very last reference but setting
>> maxcount to 1 gives just the earliest.

R> Mail or news?

mail. I didn't try news.

Re: message-shorten-references hardwired 21 references

[Reiner Steib]

On Mon, Jan 09 2006, Dan Jacobson wrote:

> R> I'm quite sure that Gnus does the right thing.  See section
> R> "6.5. References" in [1].
>
>     It is permissible to not include the entire previous "References"
>     line if it is too long.  An attempt should be made to include a
>     reasonable number of backwards references.
>
> He just says "reasonable". 

And "if it is too long".

> I say that the newest 21 is the only way to keep threads intact,
> vs. the oldest 21, no?

You've misread the code.  It keeps the first MID and the last 20 MIDs:

ELISP> (setq rs-test "<id-1@invalid> <id-2@invalid> ...
  ... <id-89@invalid> <id-90@invalid>")

ELISP> (with-temp-buffer
	  (message-shorten-references 'References rs-test)
	  (buffer-string))
"References: <id-1@invalid> <id-71@invalid> <id-72@invalid> ...
         ... <id-88@invalid> <id-89@invalid> <id-90@invalid>"

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