Proposed default change: Show the lapsed time instead of the date

Proposed default change: Show the lapsed time instead of the date

[Lars Ingebrigtsen]

I think reading the raw date header is pretty much useless, so I propose
to make the following default change:

(setq gnus-treat-date-lapsed 'head)

And, in addition, default to keep it updated.  It'll be like all Web
2.0, only in Emacs:

(gnus-start-date-timer)

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

Re: Proposed default change: Show the lapsed time instead of the date

[Robert Pluim]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> I think reading the raw date header is pretty much useless, so I propose
> to make the following default change:
>
> (setq gnus-treat-date-lapsed 'head)
>
> And, in addition, default to keep it updated.  It'll be like all Web
> 2.0, only in Emacs:
>
> (gnus-start-date-timer)

-1 I like my dates to be absolute, it causes less mental shifting for
me.

Robert

(unless this is a Lars-posting-drunk-joke, in which case: very funny ;) )

Re: Proposed default change: Show the lapsed time instead of the date

[Julien Danjou]

[-- Attachment #1: Type: text/plain, Size: 472 bytes --]

On Tue, Jan 25 2011, Lars Ingebrigtsen wrote:

> I think reading the raw date header is pretty much useless, so I propose
> to make the following default change:
>
> (setq gnus-treat-date-lapsed 'head)
>
> And, in addition, default to keep it updated.  It'll be like all Web
> 2.0, only in Emacs:
>
> (gnus-start-date-timer)

+1 I like that \o/
-1 I'd like to still have absolute date

Maybe have both?

-- 
Julien Danjou
❱ http://julien.danjou.info

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> I think reading the raw date header is pretty much useless, so I propose
> to make the following default change:
>
> (setq gnus-treat-date-lapsed 'head)
>

hmm, setting this variable and I don't see anymore the "Date:" field at
all in my articles...

-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[David Edmondson]

* francis.moro@gmail.com [2011-01-25 Tue 11:03]
> Lars Ingebrigtsen <larsi@gnus.org> writes:
>
>> I think reading the raw date header is pretty much useless, so I propose
>> to make the following default change:
>>
>> (setq gnus-treat-date-lapsed 'head)
>
> hmm, setting this variable and I don't see anymore the "Date:" field at
> all in my articles...

You must add the `X-Sent' header to those displayed. If you want both
X-Sent and Date:
	(setq gnus-article-date-lapsed-new-header t)

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

David Edmondson <dme@dme.org> writes:

> * francis.moro@gmail.com [2011-01-25 Tue 11:03]
>> Lars Ingebrigtsen <larsi@gnus.org> writes:
>>
>>> I think reading the raw date header is pretty much useless, so I propose
>>> to make the following default change:
>>>
>>> (setq gnus-treat-date-lapsed 'head)
>>
>> hmm, setting this variable and I don't see anymore the "Date:" field at
>> all in my articles...
>
> You must add the `X-Sent' header to those displayed. If you want both
> X-Sent and Date:
> 	(setq gnus-article-date-lapsed-new-header t)

Thanks !

-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> I think reading the raw date header is pretty much useless, so I propose
> to make the following default change:
>
> (setq gnus-treat-date-lapsed 'head)
>

Well, I think it makes sense FWIW.

But I also think that displaying the second is quite useless, since
there're no real point to know this information. Are there any way to
skip them ?

Also maybe it would be nice to simplify the elapsed time format. For
example, displaying this:

         5 weeks, 1 day, 3 hours, 51 minutes, 26 seconds ago

is quite ugly.

What is important is 5 weeks, the rest becomes irrevelant. So it would
be nice if we could round the elapsed time.

BTW are there any ways to make such change in the group buffer ?

Thanks
-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[Julien Danjou]

[-- Attachment #1: Type: text/plain, Size: 688 bytes --]

On Tue, Jan 25 2011, Francis Moreau wrote:

> What is important is 5 weeks, the rest becomes irrevelant. So it would
> be nice if we could round the elapsed time.

Good idea.

> BTW are there any ways to make such change in the group buffer ?

I use this, FWIW:

(setq gnus-user-date-format-alist
      '(((gnus-seconds-today) . "Today, %H:%M")
        ((+ 86400 (gnus-seconds-today)) . "Yesterday, %H:%M")
        (604800 . "%A %H:%M")     ; That's one week
        ((gnus-seconds-month) . "%A %d")
        ((gnus-seconds-year) . "%B %d")
        (t . "%B %d %Y")))       ; This one is used when no other does match

-- 
Julien Danjou
❱ http://julien.danjou.info

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: Proposed default change: Show the lapsed time instead of the date

[Adam Sjøgren]

On Mon, 24 Jan 2011 16:48:53 -0800, Lars wrote:

> And, in addition, default to keep it updated.  It'll be like all Web
> 2.0, only in Emacs:

> (gnus-start-date-timer)

I turned that on once and had to turn it off again, because the text was
changing at the periphery of where I was focusing, making me dizzy
(and/or alert).

YMMV of course.


   :-),

   Adam

-- 
 "Här kommer rädslan, gamle vän                               Adam Sjøgren
  När alla fjärilar i magen vaknar upp                   asjo@koldfront.dk
  Viskar välkommen hem"

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

asjo@koldfront.dk (Adam Sjøgren) writes:

> On Mon, 24 Jan 2011 16:48:53 -0800, Lars wrote:
>
>> And, in addition, default to keep it updated.  It'll be like all Web
>> 2.0, only in Emacs:
>
>> (gnus-start-date-timer)
>
> I turned that on once and had to turn it off again, because the text was
> changing at the periphery of where I was focusing, making me dizzy
> (and/or alert).

I did too, since IMHO displaying seconds is totaly useless, and
displaying changing seconds is incredibly useless.

-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[Ted Zlatanov]

On Tue, 25 Jan 2011 17:55:58 +0100 Francis Moreau <francis.moro@gmail.com> wrote: 

FM> asjo@koldfront.dk (Adam Sjøgren) writes:
>> On Mon, 24 Jan 2011 16:48:53 -0800, Lars wrote:
>> 
>>> And, in addition, default to keep it updated.  It'll be like all Web
>>> 2.0, only in Emacs:
>> 
>>> (gnus-start-date-timer)
>> 
>> I turned that on once and had to turn it off again, because the text was
>> changing at the periphery of where I was focusing, making me dizzy
>> (and/or alert).

FM> I did too, since IMHO displaying seconds is totaly useless, and
FM> displaying changing seconds is incredibly useless.

I like displaying the seconds and hope it's an option.

Ted

Re: Proposed default change: Show the lapsed time instead of the date

[Adam Sjøgren]

On Tue, 25 Jan 2011 17:55:58 +0100, Francis wrote:

> I did too, since IMHO displaying seconds is totaly useless, and
> displaying changing seconds is incredibly useless.

My humble opinion is that seconds are nice, but updating is slightly
annoying.

This seems to really be one of those "taste" sheds.


  Best regards,

    Adam

-- 
 "Här kommer rädslan, gamle vän                               Adam Sjøgren
  När alla fjärilar i magen vaknar upp                   asjo@koldfront.dk
  Viskar välkommen hem"

Re: Proposed default change: Show the lapsed time instead of the date

[Leonidas Tsampros]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> I think reading the raw date header is pretty much useless, so I propose
> to make the following default change:
>
> (setq gnus-treat-date-lapsed 'head)
>
> And, in addition, default to keep it updated.  It'll be like all Web
> 2.0, only in Emacs:
>
> (gnus-start-date-timer)

I think with widescreens and all the modern stuff it'd be nice to get
both values displayed (raw header + elapsed).

my 0.2c

Re: Proposed default change: Show the lapsed time instead of the date

[Lars Ingebrigtsen]

Julien Danjou <julien@danjou.info> writes:

> +1 I like that \o/
> -1 I'd like to still have absolute date
>
> Maybe have both?

I'm using a Vaio P, and I find anything that takes vertical screen real
estate to be an annoyance...

But using a 8 inch 1600x800 screen isn't a very common case, though.

But I'm wondering: Why would you want to see the Date header at all?
I'm curious.  Personally, it tells me nothing, especially what with all
the different time zones and stuff.  What's interesting to me is how old
the message is, not whether it was posted at 9:55 pm on a Tuesday.

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

Re: Proposed default change: Show the lapsed time instead of the date

[Richard Riley]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Julien Danjou <julien@danjou.info> writes:
>
>> +1 I like that \o/
>> -1 I'd like to still have absolute date
>>
>> Maybe have both?
>
> I'm using a Vaio P, and I find anything that takes vertical screen real
> estate to be an annoyance...
>
> But using a 8 inch 1600x800 screen isn't a very common case, though.

More and more so : most 10" netbooks (like my new asus 1015) are only 600
pixels high. So +1 on the vertical space comment.

Re: Proposed default change: Show the lapsed time instead of the date

[Andreas Schwab]

> But I'm wondering: Why would you want to see the Date header at all?
> I'm curious.  Personally, it tells me nothing, especially what with all
> the different time zones and stuff.  What's interesting to me is how old
> the message is, not whether it was posted at 9:55 pm on a Tuesday.

At least knowing the time zone of the sender is useful.  Also, the
absolute date can sometimes be easier to grok than the the relative one.

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: Proposed default change: Show the lapsed time instead of the date

[Ted Zlatanov]

On Tue, 25 Jan 2011 13:18:22 -0800 Lars Ingebrigtsen <larsi@gnus.org> wrote: 

LI> But I'm wondering: Why would you want to see the Date header at all?
LI> I'm curious.  Personally, it tells me nothing, especially what with all
LI> the different time zones and stuff.  What's interesting to me is how old
LI> the message is, not whether it was posted at 9:55 pm on a Tuesday.

Both are useful.  The local date tells me someone was up at 3 AM their
time, while the relative time tells me that was 8 hours ago.

Ted

Re: Proposed default change: Show the lapsed time instead of the date

[Leonidas Tsampros]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Julien Danjou <julien@danjou.info> writes:
>
>> +1 I like that \o/
>> -1 I'd like to still have absolute date
>>
>> Maybe have both?
>
> I'm using a Vaio P, and I find anything that takes vertical screen real
> estate to be an annoyance...
>
> But using a 8 inch 1600x800 screen isn't a very common case, though.
>
> But I'm wondering: Why would you want to see the Date header at all?
> I'm curious.  Personally, it tells me nothing, especially what with all
> the different time zones and stuff.  What's interesting to me is how old
> the message is, not whether it was posted at 9:55 pm on a Tuesday.

You have a really strong point, although and more importantly for work
emails, having BOTH options is far more helpful.

It would be great if the Date: header could be converted to the
localtime zone also. Less thinking.

Re: Proposed default change: Show the lapsed time instead of the date

[Reiner Steib]

On Tue, Jan 25 2011, Leonidas Tsampros wrote:

> It would be great if the Date: header could be converted to the
> localtime zone also. Less thinking.

,----[ <f1> v gnus-treat-date-local RET ]
| gnus-treat-date-local is a variable defined in `gnus-art.el'.
| Its value is nil
| 
| 
| Documentation:
| Display the Date in the local timezone.
| Valid values are nil, t, `head', `first', `last', an integer or a
| predicate.  See Info node `(gnus)Customizing Articles'.
| 
| You can customize this variable.
`----

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

Re: Proposed default change: Show the lapsed time instead of the date

[Reiner Steib]

On Tue, Jan 25 2011, Lars Ingebrigtsen wrote:

> Julien Danjou <julien@danjou.info> writes:
>
>> +1 I like that \o/
>> -1 I'd like to still have absolute date
>>
>> Maybe have both?
>
> I'm using a Vaio P, and I find anything that takes vertical screen real
> estate to be an annoyance...
>
> But using a 8 inch 1600x800 screen isn't a very common case, though.
>
> But I'm wondering: Why would you want to see the Date header at all?
> I'm curious.  Personally, it tells me nothing, especially what with all
> the different time zones and stuff.  What's interesting to me is how old
> the message is, not whether it was posted at 9:55 pm on a Tuesday.

How about using a Gmane-like format by default?
(`gnus-article-date-lapsed-merged-header'?)

| Date: 2010-11-29 04:47:28 GMT (8 weeks, 2 days, 2 hours and 44 minutes ago)

As other pointed out, if weeks is > 0, we should probably drop
(hours,) minutes, seconds.

| Date: 2010-11-29 04:47:28 GMT (8 weeks and 2 days ago)
| Date: 2011-01-25 21:18:22 GMT (10 hours and 17 minutes ago)

As we see that people have different preferences, the whole thing
should be a customizable format string, so that people can e.g. use
local or original time zone or drop seconds, etc.

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

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

Julien Danjou <julien@danjou.info> writes:

> On Tue, Jan 25 2011, Francis Moreau wrote:
>
>> BTW are there any ways to make such change in the group buffer ?
>
> I use this, FWIW:
>
> (setq gnus-user-date-format-alist
>       '(((gnus-seconds-today) . "Today, %H:%M")
>         ((+ 86400 (gnus-seconds-today)) . "Yesterday, %H:%M")
>         (604800 . "%A %H:%M")     ; That's one week
>         ((gnus-seconds-month) . "%A %d")
>         ((gnus-seconds-year) . "%B %d")
>         (t . "%B %d %Y")))       ; This one is used when no other does match

Thanks, I'll try to cook up something similar with my incredibly limited
elisp skills.

-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[Štěpán Němec]

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

> How about using a Gmane-like format by default?
> (`gnus-article-date-lapsed-merged-header'?)
>
> | Date: 2010-11-29 04:47:28 GMT (8 weeks, 2 days, 2 hours and 44 minutes ago)
>
> As other pointed out, if weeks is > 0, we should probably drop
> (hours,) minutes, seconds.
>
> | Date: 2010-11-29 04:47:28 GMT (8 weeks and 2 days ago)
> | Date: 2011-01-25 21:18:22 GMT (10 hours and 17 minutes ago)
>
> As we see that people have different preferences, the whole thing
> should be a customizable format string, so that people can e.g. use
> local or original time zone or drop seconds, etc.

+1
(I would go for the local time zone as the default, but as long as there's
an option...)

  Štěpán

Re: Proposed default change: Show the lapsed time instead of the date

[Wes Hardaker]

>>>>> On Tue, 25 Jan 2011 13:18:22 -0800, Lars Ingebrigtsen <larsi@gnus.org> said:

LI> But I'm wondering: Why would you want to see the Date header at all?
LI> I'm curious.  Personally, it tells me nothing, especially what with all
LI> the different time zones and stuff.  What's interesting to me is how old
LI> the message is, not whether it was posted at 9:55 pm on a Tuesday.

I've had the X-Sent support on for years.  But occasionally I do convert
back to the original header, frequently to quote to someone when they
wrote something.  Or for other reasons I can't remember.  I don't do it
*often* but I do read it.

Here's another option though: combine them into a single line.

  X-Sent: 18 hours, 21 minutes, 16 seconds ago
  Date: Tue, 25 Jan 2011 13:18:22 -0800

=>

  X-Sent: Tue, 25 Jan 2011 13:18:22 -0800 (18 hours, 21 minutes, 16 seconds ago)

     or

  X-Sent: 18 hours, 21 minutes, 16 seconds ago (Tue, 25 Jan 2011 13:18:22 -0800)

All fits on one line.  Though I think for the combined mode I'd drop the
seconds part since it's a bit too long.  Maybe always use the right two
significant units (eg 14 years, 2 months) and drop everything after that?



[As for time-zone concerns: a huge amount of my mail is distant and
 timezones do affect things.  However, I have a huge amount that is all
 local too so they're in the same timezone.]
-- 
Wes Hardaker                                     
My Pictures:  http://capturedonearth.com/
My Thoughts:  http://pontifications.hardakers.net/

Re: Proposed default change: Show the lapsed time instead of the date

[Eric Abrahamsen]

On Wed, Jan 26 2011, Wes Hardaker wrote:

>>>>>> On Tue, 25 Jan 2011 13:18:22 -0800, Lars Ingebrigtsen <larsi@gnus.org> said:
>
> LI> But I'm wondering: Why would you want to see the Date header at all?
> LI> I'm curious.  Personally, it tells me nothing, especially what with all
> LI> the different time zones and stuff.  What's interesting to me is how old
> LI> the message is, not whether it was posted at 9:55 pm on a Tuesday.
>
> I've had the X-Sent support on for years.  But occasionally I do convert
> back to the original header, frequently to quote to someone when they
> wrote something.  Or for other reasons I can't remember.  I don't do it
> *often* but I do read it.
>
> Here's another option though: combine them into a single line.
>
>   X-Sent: 18 hours, 21 minutes, 16 seconds ago
>   Date: Tue, 25 Jan 2011 13:18:22 -0800
>
> =>
>
>   X-Sent: Tue, 25 Jan 2011 13:18:22 -0800 (18 hours, 21 minutes, 16 seconds ago)
>
>      or
>
>   X-Sent: 18 hours, 21 minutes, 16 seconds ago (Tue, 25 Jan 2011 13:18:22 -0800)
>
> All fits on one line.  Though I think for the combined mode I'd drop the
> seconds part since it's a bit too long.  Maybe always use the right two
> significant units (eg 14 years, 2 months) and drop everything after that?

Good!

Re: Proposed default change: Show the lapsed time instead of the date

[Ted Zlatanov]

On Wed, 26 Jan 2011 08:40:54 +0100 Reiner Steib <reinersteib+gmane@imap.cc> wrote: 

RS> As we see that people have different preferences, the whole thing
RS> should be a customizable format string, so that people can e.g. use
RS> local or original time zone or drop seconds, etc.

The problem is a) establishing a default, and b) dropping useless info.
Otherwise we could have just used strftime() style formatting.  I'd
rather merge Date and X-Sent and use a funcall to display them (with the
default being what Wes Hardaker proposed).

Ted

Re: Proposed default change: Show the lapsed time instead of the date

[Lars Ingebrigtsen]

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

> How about using a Gmane-like format by default?
> (`gnus-article-date-lapsed-merged-header'?)
>
> | Date: 2010-11-29 04:47:28 GMT (8 weeks, 2 days, 2 hours and 44 minutes ago)

I really like that idea.

> As other pointed out, if weeks is > 0, we should probably drop
> (hours,) minutes, seconds.

Yeah, makes sense.

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

Re: Proposed default change: Show the lapsed time instead of the date

[Lars Ingebrigtsen]

Wes Hardaker <wes@hardakers.net> writes:

>   X-Sent: Tue, 25 Jan 2011 13:18:22 -0800 (18 hours, 21 minutes, 16 seconds ago)
>
>      or
>
>   X-Sent: 18 hours, 21 minutes, 16 seconds ago (Tue, 25 Jan 2011 13:18:22 -0800)

Would the default be to show the original date header, or to use a more
iso 8601-ey format?  The latter is more compact, but I'm not sure what
people would expect to see here.

If we stick to rfc822-date strings, then we could get away with saying
that it's a Date header, and the "lapsed" stuff would just be a
comment.  So no more need for the kinda awkward X-Sent header.

So it'd be:

Date: Tue, 25 Jan 2011 13:18:22 -0800 (18 hours, 21 minutes, 16 seconds ago)

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

Re: Proposed default change: Show the lapsed time instead of the date

[Lars Ingebrigtsen]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Date: Tue, 25 Jan 2011 13:18:22 -0800 (18 hours, 21 minutes, 16 seconds ago)

This might be too wide, though.  We should aim for less than 80
characters wide...

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

Re: Proposed default change: Show the lapsed time instead of the date

[Philipp Haselwarter]

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

RS> On Tue, Jan 25 2011, Lars Ingebrigtsen wrote:

>> Julien Danjou <julien@danjou.info> writes:
>> 
>>> +1 I like that \o/ -1 I'd like to still have absolute date
>>> 
>>> Maybe have both?
>> 
>> I'm using a Vaio P, and I find anything that takes vertical screen
>> real estate to be an annoyance...
>> 
>> But using a 8 inch 1600x800 screen isn't a very common case, though.

1024x600 here. Columns? You can have 'em. But _don't you dare touch my lines_..
---8<---[snipped 12 lines]---8<---

RS> As other pointed out, if weeks is > 0, we should probably drop
RS> (hours,) minutes, seconds.

RS> | Date: 2010-11-29 04:47:28 GMT (8 weeks and 2 days ago)
RS> | Date: 2011-01-25 21:18:22 GMT (10 hours and 17 minutes ago)

I like this. It is informative, readable and it makes sense. *want*

RS> As we see that people have different preferences, the whole thing
RS> should be a customizable format string, so that people can e.g. use
RS> local or original time zone or drop seconds, etc.

RS> Bye, Reiner.

+1

-- 
Philipp Haselwarter

Re: Proposed default change: Show the lapsed time instead of the date

[Lars Ingebrigtsen]

Philipp Haselwarter <philipp.haselwarter@gmx.de> writes:

> RS> | Date: 2011-01-25 21:18:22 GMT (10 hours and 17 minutes ago)
>
> I like this. It is informative, readable and it makes sense. *want*

I've now pushed out a variant of this.  Probably needs some
bikeshedding before it's perfect.  :-)

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

Re: Proposed default change: Show the lapsed time instead of the date

[Eric Abrahamsen]

On Wed, Jan 26 2011, Lars Ingebrigtsen wrote:

> Philipp Haselwarter <philipp.haselwarter@gmx.de> writes:
>
>> RS> | Date: 2011-01-25 21:18:22 GMT (10 hours and 17 minutes ago)
>>
>> I like this. It is informative, readable and it makes sense. *want*
>
> I've now pushed out a variant of this.  Probably needs some
> bikeshedding before it's perfect.  :-)

Ahhhh!!!!!! No seconds, no seconds!!!! I love it, except please for the
love of god don't make my new messages tick at me…

Otherwise, dope.

Re: Proposed default change: Show the lapsed time instead of the date

[Robert Pluim]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:

> On Wed, Jan 26 2011, Lars Ingebrigtsen wrote:
>
>> Philipp Haselwarter <philipp.haselwarter@gmx.de> writes:
>>
>>> RS> | Date: 2011-01-25 21:18:22 GMT (10 hours and 17 minutes ago)
>>>
>>> I like this. It is informative, readable and it makes sense. *want*
>>
>> I've now pushed out a variant of this.  Probably needs some
>> bikeshedding before it's perfect.  :-)
>
> Ahhhh!!!!!! No seconds, no seconds!!!! I love it, except please for the
> love of god don't make my new messages tick at me…

Seconded

Robert

Re: Proposed default change: Show the lapsed time instead of the date

[Julien Danjou]

[-- Attachment #1: Type: text/plain, Size: 415 bytes --]

On Thu, Jan 27 2011, Lars Ingebrigtsen wrote:

> I've now pushed out a variant of this.  Probably needs some
> bikeshedding before it's perfect.  :-)

Looks good, but the date is now impossible to mark. The cursor is kicked
out each time I navigate through the Date: header. Just a minor
annoyance if you want to yank the date, that's now impossible. :)

-- 
Julien Danjou
❱ http://julien.danjou.info

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

Julien Danjou <julien@danjou.info> writes:

> On Thu, Jan 27 2011, Lars Ingebrigtsen wrote:
>
>> I've now pushed out a variant of this.  Probably needs some
>> bikeshedding before it's perfect.  :-)
>
> Looks good, but the date is now impossible to mark. The cursor is kicked
> out each time I navigate through the Date: header. Just a minor
> annoyance if you want to yank the date, that's now impossible. :)

Well don't update the seconds if it's really annoying, specially since
it's really useless.

-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Philipp Haselwarter <philipp.haselwarter@gmx.de> writes:
>
>> RS> | Date: 2011-01-25 21:18:22 GMT (10 hours and 17 minutes ago)
>>
>> I like this. It is informative, readable and it makes sense. *want*
>
> I've now pushed out a variant of this.  Probably needs some
> bikeshedding before it's perfect.  :-)

Could have something like

  (setq gnus-date-lapsed-format 'default)
  (setq gnus-date-lapsed-format 'round)


where dense is the date round that I proposed previously: for example:

      10 hours and 17 minutes ago => 10 hours ago
      34 minutes and 54 seconds ago => 34 mins ago
      1 hour, 30 minutes and 20 seconds ago => 1 hour, 30 mins ago

There 2 things showed:

  - you can use maybe use mins, secs instead of minutes, seconds
  - for the rounding: you drop any (n+1)th field if n+1 > 10% * n

where 10% can be customizable.

Thanks
-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[James Cloos]

>>>>> "LI" == Lars Ingebrigtsen <larsi@gnus.org> writes:

LI> Philipp Haselwarter <philipp.haselwarter@gmx.de> writes:
RS> | Date: 2011-01-25 21:18:22 GMT (10 hours and 17 minutes ago)
>> 
>> I like this. It is informative, readable and it makes sense. *want*

LI> I've now pushed out a variant of this.  Probably needs some
LI> bikeshedding before it's perfect.  :-)

I use this (grep(1)ed from .emacs-custom):

 '(gnus-article-date-lapsed-new-header t)
 '(gnus-treat-date-iso8601 nil)
 '(gnus-treat-date-lapsed (quote head))
 '(gnus-treat-date-local (quote head))

and want to continue to see separate Date: and X-Sent: headers (with
Date in my local tz).

What settings will I need to fix after I next update gnus to continue
to do so?

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: Proposed default change: Show the lapsed time instead of the date

[Wes Hardaker]

>>>>> On Wed, 26 Jan 2011 15:14:37 -0800, Lars Ingebrigtsen <larsi@gnus.org> said:

>> Date: Tue, 25 Jan 2011 13:18:22 -0800 (18 hours, 21 minutes, 16 seconds ago)

LI> This might be too wide, though.  We should aim for less than 80
LI> characters wide...

That's sort of why I suggested limiting it to the "two most significant
units".

However...  This is code.  So it's actually very possible to say 

  Date: Tue, 25 Jan 2011 13:18:22 -0800

Hmm...  That's 37 chars.

I can add to that...  Lets try:

  Date: Tue, 25 Jan 2011 13:18:22 -0800 (18 hours)

Yep, still fits...

  [time marches on and ()s are continued to be matched internally]

  Date: Tue, 25 Jan 2011 13:18:22 -0800 (18 hours, 21 minutes, 16 seconds ago)

Yep, that still fits lets try:

  Date: Tue, 25 Jan 2011 13:18:22 -0800 (18 hours, 21 minutes, 16 seconds ago, 14 milliseconds)

Nope, that doesn't fit.  Revert to the previous and display.


*that* would be the best solution.
-- 
Wes Hardaker                                     
My Pictures:  http://capturedonearth.com/
My Thoughts:  http://pontifications.hardakers.net/

Re: Proposed default change: Show the lapsed time instead of the date

[Peter Münster]

Wes Hardaker <wes@hardakers.net> writes:

> That's sort of why I suggested limiting it to the "two most significant
> units".

*that* would be the best solution.


>   Date: Tue, 25 Jan 2011 13:18:22 -0800 (18 hours, 21 minutes, 16 seconds ago,
> 14 milliseconds)

Then, please update this in real-time! ;)

-- 
           Peter

Re: Proposed default change: Show the lapsed time instead of the date

[Lars Ingebrigtsen]

Julien Danjou <julien@danjou.info> writes:

> Looks good, but the date is now impossible to mark. The cursor is kicked
> out each time I navigate through the Date: header. Just a minor
> annoyance if you want to yank the date, that's now impossible. :)

Yeah, the code should try harder to keep point "where it is".  It's not
totally trivial, because the way to do it is:

1) if point is outside the region that's being killed/inserted, then the
best way to ensure that points remains "where it is" is to put a mark
at point and then put point at that mark afterwards.

2) if point is inside the region that's being killed/inserted, the best
way is to put point at the same numerical point position as before.

I'll try to write a macro that implements this...

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

Re: Proposed default change: Show the lapsed time instead of the date

[Lars Ingebrigtsen]

Francis Moreau <francis.moro@gmail.com> writes:

> Could have something like
>
>   (setq gnus-date-lapsed-format 'default)
>   (setq gnus-date-lapsed-format 'round)

I don't really see the point in making this customisable.  At present it
shows the three most significant segments, so this all works
automatically.

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

Re: Proposed default change: Show the lapsed time instead of the date

[Lars Ingebrigtsen]

James Cloos <cloos@jhcloos.com> writes:

> I use this (grep(1)ed from .emacs-custom):
>
>  '(gnus-article-date-lapsed-new-header t)
>  '(gnus-treat-date-iso8601 nil)
>  '(gnus-treat-date-lapsed (quote head))
>  '(gnus-treat-date-local (quote head))
>
> and want to continue to see separate Date: and X-Sent: headers (with
> Date in my local tz).
>
> What settings will I need to fix after I next update gnus to continue
> to do so?

I found some slight bugs in relation to keeping all these things working
at the same time...

However, the date settings are kinda confusing.  I think it may have
been a design mistake to keep these in separate variables.  I think it
should have been

(defcustom gnus-article-date-headers '(local lapsed))

or something.  That is, just a list of all the different Date headers
you want to see.  That would have been a lot simpler and more flexible.

Now gnus-treat-date-combined-lapsed defaults to 'head, which means that
there are two Date headers (local and combined-lapsed) that are
clobbering each other, and because combined-lapsed happens to be last,
it wins.  (Well, that and the timer that updates it.)

So what you'll see with your settings is the new combined-lapsed, and
the old lapsed in addition.  So you'll get two updating lines.

This is obviously not ideal.

And if somebody has set gnus-treat-date-iso8601, they'll just get the
new combined-lapsed format, anyway.

Ideas, all y'all?

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

Re: Proposed default change: Show the lapsed time instead of the date

[Lars Ingebrigtsen]

James Cloos <cloos@jhcloos.com> writes:

> What settings will I need to fix after I next update gnus to continue
> to do so?

Oh, and to answer your question.  :-)

You need to set gnus-treat-date-combined-lapsed to nil.

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

Re: Proposed default change: Show the lapsed time instead of the date

[Lars Ingebrigtsen]

Wes Hardaker <wes@hardakers.net> writes:

> *that* would be the best solution.

Great idea.  I'll implement it.

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

Re: Proposed default change: Show the lapsed time instead of the date

[Wes Hardaker]

>>>>> On Thu, 27 Jan 2011 15:23:42 -0800, Lars Ingebrigtsen <larsi@gnus.org> said:

LI> Wes Hardaker <wes@hardakers.net> writes:
>> *that* would be the best solution.

LI> Great idea.  I'll implement it.

Slick.  I may have to do a 'git pull' yet.  I'm still pre-imap-rewrite
until my mail becomes less critical (probably in the next few weeks).
-- 
Wes Hardaker                                     
My Pictures:  http://capturedonearth.com/
My Thoughts:  http://pontifications.hardakers.net/

Re: Proposed default change: Show the lapsed time instead of the date

[Steinar Bang]

>>>>> Wes Hardaker <wes@hardakers.net>:

> Slick.  I may have to do a 'git pull' yet.  I'm still pre-imap-rewrite
> until my mail becomes less critical (probably in the next few weeks).

What kind of IMAP server are you talking to?  dovecot?  Gmail?
Exchange?  Something else?

I'm using an IMAP server, and apart from a few scares in the early days,
my experience with the rewrite have been positive.

But I do my mail splitting outside of emacs.  It seems like that's the
place where people have had most of their issues...?

When you do upgrade: the new search functionality is nice and something
to look forward to (if your IMAP server supports search, that is).

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Francis Moreau <francis.moro@gmail.com> writes:
>
>> Could have something like
>>
>>   (setq gnus-date-lapsed-format 'default)
>>   (setq gnus-date-lapsed-format 'round)
>
> I don't really see the point in making this customisable.  At present
> it shows the three most significant segments, so this all works
> automatically.

But it's not what I try to show you with the 'round' method.

You're currently displaying the 3 most significant segments which I
still find too verbose IMO when the 2 last segments are too small
compare to the first one.

What I'd like is when the 2 last are too small, just drop them:

For example:

   10 years, 3 months, 23 days ago

becomes

   10 years ago

That is, the revelant information is 10 years, the rest is irrevelant
and if I really need to know when this happen exactly, I'll look at the
date !

Thanks
-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[Richard Riley]

Francis Moreau <francis.moro@gmail.com> writes:

> Lars Ingebrigtsen <larsi@gnus.org> writes:
>
>> Francis Moreau <francis.moro@gmail.com> writes:
>>
>>> Could have something like
>>>
>>>   (setq gnus-date-lapsed-format 'default)
>>>   (setq gnus-date-lapsed-format 'round)
>>
>> I don't really see the point in making this customisable.  At present
>> it shows the three most significant segments, so this all works
>> automatically.
>
> But it's not what I try to show you with the 'round' method.
>
> You're currently displaying the 3 most significant segments which I
> still find too verbose IMO when the 2 last segments are too small
> compare to the first one.
>

How do I turn this animated elapsed time feature off? Its cool but very
distracting.

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

Richard Riley <rileyrg@googlemail.com> writes:

> Francis Moreau <francis.moro@gmail.com> writes:
>
>> Lars Ingebrigtsen <larsi@gnus.org> writes:
>>
>>> Francis Moreau <francis.moro@gmail.com> writes:
>>>
>>>> Could have something like
>>>>
>>>>   (setq gnus-date-lapsed-format 'default)
>>>>   (setq gnus-date-lapsed-format 'round)
>>>
>>> I don't really see the point in making this customisable.  At present
>>> it shows the three most significant segments, so this all works
>>> automatically.
>>
>> But it's not what I try to show you with the 'round' method.
>>
>> You're currently displaying the 3 most significant segments which I
>> still find too verbose IMO when the 2 last segments are too small
>> compare to the first one.
>>
>
> How do I turn this animated elapsed time feature off? Its cool but very
> distracting.
>

(setq gnus-article-update-lapsed-header nil)

I think
-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[Steinar Bang]

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

> What kind of IMAP server are you talking to?  dovecot?  Gmail?
> Exchange?  Something else?

> I'm using an IMAP server,

"I'm using a dovecot IMAP server", I mean to say...

> and apart from a few scares in the early days, my experience with the
> rewrite have been positive.

Still holds.

Re: Proposed default change: Show the lapsed time instead of the date

[Eric Abrahamsen]

On Fri, Jan 28 2011, Francis Moreau wrote:

> Richard Riley <rileyrg@googlemail.com> writes:
>
>> Francis Moreau <francis.moro@gmail.com> writes:
>>
>>> Lars Ingebrigtsen <larsi@gnus.org> writes:
>>>
>>>> Francis Moreau <francis.moro@gmail.com> writes:
>>>>
>>>>> Could have something like
>>>>>
>>>>>   (setq gnus-date-lapsed-format 'default)
>>>>>   (setq gnus-date-lapsed-format 'round)
>>>>
>>>> I don't really see the point in making this customisable.  At present
>>>> it shows the three most significant segments, so this all works
>>>> automatically.
>>>
>>> But it's not what I try to show you with the 'round' method.
>>>
>>> You're currently displaying the 3 most significant segments which I
>>> still find too verbose IMO when the 2 last segments are too small
>>> compare to the first one.
>>>
>>
>> How do I turn this animated elapsed time feature off? Its cool but very
>> distracting.
>>
>
> (setq gnus-article-update-lapsed-header nil)

This seems not to be respected…

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:

> On Fri, Jan 28 2011, Francis Moreau wrote:
>
>> Richard Riley <rileyrg@googlemail.com> writes:
>>
>>> Francis Moreau <francis.moro@gmail.com> writes:
>>>
>>>> Lars Ingebrigtsen <larsi@gnus.org> writes:
>>>>
>>>>> Francis Moreau <francis.moro@gmail.com> writes:
>>>>>
>>>>>> Could have something like
>>>>>>
>>>>>>   (setq gnus-date-lapsed-format 'default)
>>>>>>   (setq gnus-date-lapsed-format 'round)
>>>>>
>>>>> I don't really see the point in making this customisable.  At present
>>>>> it shows the three most significant segments, so this all works
>>>>> automatically.
>>>>
>>>> But it's not what I try to show you with the 'round' method.
>>>>
>>>> You're currently displaying the 3 most significant segments which I
>>>> still find too verbose IMO when the 2 last segments are too small
>>>> compare to the first one.
>>>>
>>>
>>> How do I turn this animated elapsed time feature off? Its cool but very
>>> distracting.
>>>
>>
>> (setq gnus-article-update-lapsed-header nil)
>
> This seems not to be respected…

This is all I have:

 (setq gnus-treat-date-combined-lapsed 'head)
 (setq gnus-article-update-lapsed-header nil)

-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[Philipp Haselwarter]

just make it reasonable by default (I find the
$actual_date $three_most_relevant_elapsed
scheme perfect as default)
and have an option that takes a `format-time-string'-like parameter for
those who want to fiddle. I don't think you could cover all the cases,
so just offer some sensible options and let the user hack if he wants
something special...

-- 
Philipp Haselwarter

Re: Proposed default change: Show the lapsed time instead of the date

[Wes Hardaker]

>>>>> On Fri, 28 Jan 2011 09:16:21 +0100, Steinar Bang <sb@dod.no> said:

SB> What kind of IMAP server are you talking to?  dovecot?  Gmail?
SB> Exchange?  Something else?

dovecot...  It was the "losing marks" that really scared me though.

[it's just time; I'm swamped at the moment till next Wednesday; I've
actually been eager to try it.]

SB> When you do upgrade: the new search functionality is nice and something
SB> to look forward to (if your IMAP server supports search, that is).

Good to know.  I've been using the existing search, which is quiet good,
but ended up needing to write a special defun to mark multiple related
groups together to do searching.  (IE box, old.box and archive.box).



-- 
Wes Hardaker                                     
My Pictures:  http://capturedonearth.com/
My Thoughts:  http://pontifications.hardakers.net/

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

Steinar Bang <sb@dod.no> writes:

>>>>>> Wes Hardaker <wes@hardakers.net>:
>
>> Slick.  I may have to do a 'git pull' yet.  I'm still pre-imap-rewrite
>> until my mail becomes less critical (probably in the next few weeks).
>
> What kind of IMAP server are you talking to?  dovecot?  Gmail?
> Exchange?  Something else?
>
> I'm using an IMAP server, and apart from a few scares in the early days,
> my experience with the rewrite have been positive.
>
> But I do my mail splitting outside of emacs.  It seems like that's the
> place where people have had most of their issues...?
>
> When you do upgrade: the new search functionality is nice and something
> to look forward to (if your IMAP server supports search, that is).

Ah... interesting, could you point out the revelant part of IMAP search
in the documentation ?

Thanks
-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[Steinar Bang]

>>>>> Francis Moreau <francis.moro@gmail.com>:

> Ah... interesting, could you point out the revelant part of IMAP
> search in the documentation ?

(gnus) Top > Searching > nnir

I got there with `C-h i gnus RET i nnir RET', but you should get there
with `C-h i gnus RET searching RET nnir RET.

The stuff in the "Basic Usage" section should tell you what you need.

Re: Proposed default change: Show the lapsed time instead of the date

[Steinar Bang]

>>>>> Wes Hardaker <wes@hardakers.net>:

> dovecot...  It was the "losing marks" that really scared me though.

That was mainly a scare I think.  I haven't actually lost any marks.

SB> When you do upgrade: the new search functionality is nice and something
SB> to look forward to (if your IMAP server supports search, that is).

> Good to know.  I've been using the existing search, which is quiet
> good, but ended up needing to write a special defun to mark multiple
> related groups together to do searching.  (IE box, old.box and
> archive.box).

Grouping together groups in that way is something I would still like to
see, but I don't know how well it will play with the `A T' functionality
(which warps the search item back to the actual group, where you can
tick it, and reply to it, with the appropriate posting-style)...?

Anyway, it's so fast now that doing a `L' to make the archive group
visible, and then a new `G G' isn't much of a big deal.

The improved nnir search works extremely well on gmane NNTP groups as
well.

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

Steinar Bang <sb@dod.no> writes:

>>>>>> Francis Moreau <francis.moro@gmail.com>:
>
>> Ah... interesting, could you point out the revelant part of IMAP
>> search in the documentation ?
>
> (gnus) Top > Searching > nnir
>
> I got there with `C-h i gnus RET i nnir RET', but you should get there
> with `C-h i gnus RET searching RET nnir RET.

This what I can read:

     FIXME: As a first step, convert the commentary of `nnir' to texi.

-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[Francis Moreau]

Steinar Bang <sb@dod.no> writes:

[...]

>
> The improved nnir search works extremely well on gmane NNTP groups as
> well.

Are you talking about doing 'G W' in the group buffer and selecting
gmane as a search engine ?

Thanks
-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[Steinar Bang]

>>>>> Francis Moreau <francis.moro@gmail.com>:

> Steinar Bang <sb@dod.no> writes:
>>>>>>> Francis Moreau <francis.moro@gmail.com>:
>> 
>>> Ah... interesting, could you point out the revelant part of IMAP
>>> search in the documentation ?
>> 
>> (gnus) Top > Searching > nnir
>> 
>> I got there with `C-h i gnus RET i nnir RET', but you should get there
>> with `C-h i gnus RET searching RET nnir RET.

> This what I can read:

>      FIXME: As a first step, convert the commentary of `nnir' to texi.

Right.  You will need to get git gnus,
 http://gnus.org/distribution.html
and build at least the texi, and get it into the Info-directory-list.

E.g. like this (sets load-path an Info-directory-list for a git gnus
found in ~/git/gnus ):

(defvar git-workspace (expand-file-name "~/git"))

;; Git version of Gnus:
(let ((git-gnus-directory
       (if running-xemacs
	   (concat git-workspace "/gnus.xemacs")
	 (concat git-workspace "/gnus"))))
  (push (concat git-gnus-directory "/lisp") load-path)
  (let ((git-gnus-info-dir (concat git-gnus-directory "/texi")))
    (if (boundp 'Info-directory-list)
	(push git-gnus-info-dir Info-directory-list)
      (setq Info-directory-list (append
				 (list git-gnus-info-dir)
				 Info-default-directory-list)))))


It would be nice if an HTML version of the current git gnus manual could
be found on http://gnus.org/

Re: Proposed default change: Show the lapsed time instead of the date

[Steinar Bang]

>>>>> Francis Moreau <francis.moro@gmail.com>:

> Steinar Bang <sb@dod.no> writes:
> [...]

>> The improved nnir search works extremely well on gmane NNTP groups as
>> well.

> Are you talking about doing 'G W' in the group buffer and selecting
> gmane as a search engine ?

Nope.  `G G' on a gmane group.

IMO superiour to `G W' since `G G' lets you warp the results to the
actual gmane group, where you can tick the articles and have
gnus-posting-styles set up right for the group, and you can get
responded to and forwarded marks set.

It's fast too.  Both the search itself, and doing `A T' to warp to the
thread in the original group.

nnir search (was Re: Proposed default change: Show the lapsed time instead of the date)

[Francis Moreau]

Steinar Bang <sb@dod.no> writes:

>>>>>> Francis Moreau <francis.moro@gmail.com>:
>
>> Steinar Bang <sb@dod.no> writes:
>> [...]
>
>>> The improved nnir search works extremely well on gmane NNTP groups as
>>> well.
>
>> Are you talking about doing 'G W' in the group buffer and selecting
>> gmane as a search engine ?
>
> Nope.  `G G' on a gmane group.
>
> IMO superiour to `G W' since `G G' lets you warp the results to the
> actual gmane group, where you can tick the articles and have
> gnus-posting-styles set up right for the group, and you can get
> responded to and forwarded marks set.

Ok so after tried it, I think I'm simply going to forget `G W'.

A couple of question though, hope you don't mind.

Doing `G G' on topic is really cool to perform the search on a set of
groups. But it doesn't work if the topic contains sub topics. Groups
belonging to the sub topics won't be searched. For instance:

   [ A ]
      10: nnimap+gmail:INBOX 
       *: nnimap+gmail:Trash 
       *: nnimap+gmail:Sent 
       *: nnimap+gmail:Drafts 
       *: nnimap+gmail:Queue 
     [ B ]
        *: nnimap+gmail:gnus

if I do `G G' on topic A, gnus group won't be searched.

Is this expected.

Is it possible to doing a limited search on to a specific part of the
message (like wih the imap search engine) with the gmane engine ?

For local stuff, I'm using nnmairix. Why isn't it a possible search
backend of nnir ?

Thanks
-- 
Francis

Re: nnir search (was Re: Proposed default change: Show the lapsed time instead of the date)

[Andrew Cohen]

>>>>> "Francis" == Francis Moreau <francis.moro@gmail.com> writes:

    Francis> Steinar Bang <sb@dod.no> writes:
    >>>>>>> Francis Moreau <francis.moro@gmail.com>:
    >> 
    >>> Steinar Bang <sb@dod.no> writes: [...]
    >> 
    >>>> The improved nnir search works extremely well on gmane NNTP
    >>>> groups as well.
    >> 

[...]

    Francis> A couple of question though, hope you don't mind.

    Francis> Doing `G G' on topic is really cool to perform the search
    Francis> on a set of groups. But it doesn't work if the topic
    Francis> contains sub topics. 

[...]

    Francis> Is this expected.

Erh, uhm, maybe :) I never thought about recursing into sub-topics. I
can see about adding this. 

    Francis> Is it possible to doing a limited search on to a specific
    Francis> part of the message (like wih the imap search engine) with
    Francis> the gmane engine ?

Not really. The gmane search interface (documented at search.gmane.org)
just searches the whole message (aside from restricting by group and
author). 

    Francis> For local stuff, I'm using nnmairix. Why isn't it a
    Francis> possible search backend of nnir ?

It's on my todo list. But nnmairix provides some functionality that
isn't available through the nnir interface, so if you use any of these
features you should stick with nnmairix anyway.

Re: nnir search

[Francis Moreau]

Andrew Cohen <cohen@andy.bu.edu> writes:

>>>>>> "Francis" == Francis Moreau <francis.moro@gmail.com> writes:
>
>     Francis> Steinar Bang <sb@dod.no> writes:
>     >>>>>>> Francis Moreau <francis.moro@gmail.com>:
>     >> 
>     >>> Steinar Bang <sb@dod.no> writes: [...]
>     >> 
>     >>>> The improved nnir search works extremely well on gmane NNTP
>     >>>> groups as well.
>     >> 
>
> [...]
>
>     Francis> A couple of question though, hope you don't mind.
>
>     Francis> Doing `G G' on topic is really cool to perform the search
>     Francis> on a set of groups. But it doesn't work if the topic
>     Francis> contains sub topics. 
>
> [...]
>
>     Francis> Is this expected.
>
> Erh, uhm, maybe :) I never thought about recursing into sub-topics. I
> can see about adding this. 
>

Cool !

Please let me know when you're done, I'll do some testings.

>
>     Francis> Is it possible to doing a limited search on to a specific
>     Francis> part of the message (like wih the imap search engine) with
>     Francis> the gmane engine ?
>
> Not really. The gmane search interface (documented at search.gmane.org)
> just searches the whole message (aside from restricting by group and
> author). 
>

Well I guess I can just do a seach and once in the nnir group, filter
articles by subject.

But the same can be done with filtering by author.

I dunno, I would have thought that filtering by subject was more
common...

>
>     Francis> For local stuff, I'm using nnmairix. Why isn't it a
>     Francis> possible search backend of nnir ?
>
> It's on my todo list. But nnmairix provides some functionality that
> isn't available through the nnir interface, so if you use any of these
> features you should stick with nnmairix anyway.

I don't see which features you're thinking. That probably means I'm not
in this case. Basically I use nnmairix to do basic search by subject,
from... and of course update my database a couple of time in the week.

Thanks
-- 
Francis

Re: Proposed default change: Show the lapsed time instead of the date

[Detlev Zundel]

Hi,

> On Wed, Jan 26 2011, Lars Ingebrigtsen wrote:
>
>> Philipp Haselwarter <philipp.haselwarter@gmx.de> writes:
>>
>>> RS> | Date: 2011-01-25 21:18:22 GMT (10 hours and 17 minutes ago)
>>>
>>> I like this. It is informative, readable and it makes sense. *want*
>>
>> I've now pushed out a variant of this.  Probably needs some
>> bikeshedding before it's perfect.  :-)
>
> Ahhhh!!!!!! No seconds, no seconds!!!! I love it, except please for the
> love of god don't make my new messages tick at me…

Yes, seconded.  Actually after updating emacs from git, this nuisance
made me look up this thread and state my clear aversion of this
irritating change.  Please stick to a static mail display.

Cheers
  Detlev

-- 
Der Kluge tue gleich anfangs, was der Dumme erst am Ende.
                                    --- Baltasar Gracian

Re: Proposed default change: Show the lapsed time instead of the date

[Philipp Haselwarter]

Steinar Bang <sb@dod.no> writes:

>>>>>> Francis Moreau <francis.moro@gmail.com>:
>
>> Steinar Bang <sb@dod.no> writes:
>>>>>>>> Francis Moreau <francis.moro@gmail.com>:
>>> 
>>>> Ah... interesting, could you point out the revelant part of IMAP
>>>> search in the documentation ?
>>> 
>>> (gnus) Top > Searching > nnir
>>> 
>>> I got there with `C-h i gnus RET i nnir RET', but you should get there
>>> with `C-h i gnus RET searching RET nnir RET.
>
>> This what I can read:
>
>>      FIXME: As a first step, convert the commentary of `nnir' to texi.
>
> Right.  You will need to get git gnus,
>  http://gnus.org/distribution.html
> and build at least the texi, and get it into the Info-directory-list.
>
> E.g. like this (sets load-path an Info-directory-list for a git gnus
> found in ~/git/gnus ):
>
> (defvar git-workspace (expand-file-name "~/git"))
>
> ;; Git version of Gnus:
> (let ((git-gnus-directory
>        (if running-xemacs
> 	   (concat git-workspace "/gnus.xemacs")
> 	 (concat git-workspace "/gnus"))))
>   (push (concat git-gnus-directory "/lisp") load-path)
>   (let ((git-gnus-info-dir (concat git-gnus-directory "/texi")))
>     (if (boundp 'Info-directory-list)
> 	(push git-gnus-info-dir Info-directory-list)
>       (setq Info-directory-list (append
> 				 (list git-gnus-info-dir)
> 				 Info-default-directory-list)))))
>
>
> It would be nice if an HTML version of the current git gnus manual could
> be found on http://gnus.org/

How do you actually add the icons (well, the stuff in "/share/emacs/etc")?

-- 
Philipp Haselwarter

up to date manual on gnus.org (was: Proposed default change: Show the lapsed time instead of the date)

[Peter Münster]

Steinar Bang <sb@dod.no> writes:

> It would be nice if an HTML version of the current git gnus manual could
> be found on http://gnus.org/

And/or PDF version.

-- 
           Peter

Re: up to date manual on gnus.org

[Steinar Bang]

>>>>> pmlists@free.fr (Peter Münster):

> Steinar Bang <sb@dod.no> writes:
>> It would be nice if an HTML version of the current git gnus manual could
>> be found on http://gnus.org/

> And/or PDF version.

For my original wish, that would be "and", because the nice thing with
an HTML version is that I in this case could have linked directly into
the relevant section of the current git HEAD version of the gnus manual.

Re: up to date manual on gnus.org

[Lars Ingebrigtsen]

pmlists@free.fr (Peter Münster) writes:

>> It would be nice if an HTML version of the current git gnus manual could
>> be found on http://gnus.org/
>
> And/or PDF version.

I think it makes more sense to have the version of the manual on
gnus.org be the released version.  People who use git Gnus can jolly
well read the Gnus Info manual.

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

Re: Proposed default change: Show the lapsed time instead of the date

[Richard Riley]

Steinar Bang <sb@dod.no> writes:

> It would be nice if an HTML version of the current git gnus manual could
> be found on http://gnus.org/

I think that would confuse people who are using released Emacs
versions. Google will be leading people to the wrong places (not for the
first time ;).

git Gnus changes so rapidly it might also create another unwanted
workload. 

Re: Proposed default change: Show the lapsed time instead of the date

[Antoine Levitt]

01/02/11 13:34, Detlev Zundel
> Hi,
>
>> On Wed, Jan 26 2011, Lars Ingebrigtsen wrote:
>>
>>> Philipp Haselwarter <philipp.haselwarter@gmx.de> writes:
>>>
>>>> RS> | Date: 2011-01-25 21:18:22 GMT (10 hours and 17 minutes ago)
>>>>
>>>> I like this. It is informative, readable and it makes sense. *want*
>>>
>>> I've now pushed out a variant of this.  Probably needs some
>>> bikeshedding before it's perfect.  :-)
>>
>> Ahhhh!!!!!! No seconds, no seconds!!!! I love it, except please for the
>> love of god don't make my new messages tick at me…
>
> Yes, seconded.  Actually after updating emacs from git, this nuisance
> made me look up this thread and state my clear aversion of this
> irritating change.  Please stick to a static mail display.
>
> Cheers
>   Detlev

Thirded. This is a great change, but it needs better defaults. It seems
to me a sensible automagical default would be to display only one
segment, and never display seconds. In other words:

- Elapsed time is >1h : display "n hours ago"

- Elapsed time is >1m, <1h : display "m minutes ago"

- Elapsed time is <1m : display "less than one minute ago", or "just
 now", or something like that.

AFAIK, this is close to what gmail does. Of course, this being emacs, it
should be customizable. What do you think?

Also, 1s is too short for emacs to run a timer. Shouldn't
gnus-article-update-date-headers be higher?

Antoine

Re: up to date manual on gnus.org

[Steinar Bang]

>>>>> Lars Ingebrigtsen <larsi@gnus.org>:

> I think it makes more sense to have the version of the manual on
> gnus.org be the released version.  People who use git Gnus can jolly
> well read the Gnus Info manual.

I'm not suggesting you should replace that.  Not on the one linked to
from the front page, or nearly there.

I'm thinking in addition to, under the development version info
somewhere.

Or... the development version could be listed under the released version
on the web page...?

That's how they've done it for the subversion book:
 http://svnbook.red-bean.com/index.en.html
 

Re: up to date manual on gnus.org

[Peter Münster]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> I think it makes more sense to have the version of the manual on
> gnus.org be the released version.  People who use git Gnus can jolly
> well read the Gnus Info manual.

I prefer html or pdf over info. But it seems, that I can build the pdf
locally with "make pdf" in the texi directory. So it does not need to be
on-line.

Just one little error:

./gnustmp.texi:8652: Undefined control sequence.
�->@guilletright 
                 
@tclose ...on @rawbackslash @plainfrenchspacing #1
                                                  }@null 
@samp ...{@setupmarkupstyle {samp}@lq @tclose {#1}
                                                  @rq @null }
l.8652 like.  For instance, @samp{�}
                                     is tranlated into @samp{>>}, and so
                                                  on.

-- 
           Peter

Re: up to date manual on gnus.org

[Steinar Bang]

> Lars Ingebrigtsen <larsi@gnus.org> writes:

>> I think it makes more sense to have the version of the manual on
>> gnus.org be the released version.  People who use git Gnus can jolly
>> well read the Gnus Info manual.

The use case here was showing someone not already on git gnus what the
new search functionality in git gnus was like.

Not having an HTML file for people who can't type `C-h i gnus RET'.

Re: up to date manual on gnus.org

[Lars Ingebrigtsen]

Steinar Bang <sb@dod.no> writes:

> The use case here was showing someone not already on git gnus what the
> new search functionality in git gnus was like.
>
> Not having an HTML file for people who can't type `C-h i gnus RET'.

Can't you just quote the relevant bits from the Info manual?  :-)

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

Re: up to date manual on gnus.org

[Lars Ingebrigtsen]

pmlists@free.fr (Peter Münster) writes:

> I prefer html or pdf over info. But it seems, that I can build the pdf
> locally with "make pdf" in the texi directory. So it does not need to be
> on-line.
>
> Just one little error:
>
> ./gnustmp.texi:8652: Undefined control sequence.
> �->@guilletright 

I'm unable to reproduce this error.  "make pdf" works for me.

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

Re: up to date manual on gnus.org

[Steinar Bang]

>>>>> Lars Ingebrigtsen <larsi@gnus.org>:

> Steinar Bang <sb@dod.no> writes:
>> The use case here was showing someone not already on git gnus what the
>> new search functionality in git gnus was like.
>> 
>> Not having an HTML file for people who can't type `C-h i gnus RET'.

> Can't you just quote the relevant bits from the Info manual?  :-)

That particular bit was a bit on the big side.  It just becomes noise
when added to a message.

Of course I could have found my own old "try this out" message, but I
figured Andy's stuff in the manual would be more up to date.

Re: up to date manual on gnus.org

[Peter Münster]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> pmlists@free.fr (Peter Münster) writes:
>
>> Just one little error:
>>
>> ./gnustmp.texi:8652: Undefined control sequence.
>> �->@guilletright 
>
> I'm unable to reproduce this error.  "make pdf" works for me.

There was a bug in my texinfo installation, in the texinfo.tex file:
  \gdef^^bb{\guilletright}
Must be: \guillemetright

-- 
           Peter

Re: up to date manual on gnus.org

[Ted Zlatanov]

On Tue, 01 Feb 2011 05:10:45 -0800 Lars Ingebrigtsen <larsi@gnus.org> wrote: 

LI> pmlists@free.fr (Peter Münster) writes:
>>> It would be nice if an HTML version of the current git gnus manual could
>>> be found on http://gnus.org/
>> 
>> And/or PDF version.

LI> I think it makes more sense to have the version of the manual on
LI> gnus.org be the released version.  People who use git Gnus can jolly
LI> well read the Gnus Info manual.

I'll second Steinar's suggestion to have space on the webserver for the
master Git branch's version on the manual.  I've been planning to set
that up for a while now (compared to my pace, glaciers are lively).  It
would be nice to have and I can set it up on the post-commit hook.  But
we should make sure robots.txt says it's not to be indexed so Google
searches don't come up with the Git version's manual.  Is all that OK?

Ted

Re: up to date manual on gnus.org

[Lars Ingebrigtsen]

Ted Zlatanov <tzz@lifelogs.com> writes:

> But we should make sure robots.txt says it's not to be indexed so
> Google searches don't come up with the Git version's manual.  Is all
> that OK?

Sounds good.  And the pages should be marked clearly (in the title, or
something) saying that it's for the development version.

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

Re: up to date manual on gnus.org

[Ted Zlatanov]

On Thu, 03 Feb 2011 11:59:53 -0800 Lars Ingebrigtsen <larsi@gnus.org> wrote: 

LI> Ted Zlatanov <tzz@lifelogs.com> writes:
>> But we should make sure robots.txt says it's not to be indexed so
>> Google searches don't come up with the Git version's manual.  Is all
>> that OK?

LI> Sounds good.  And the pages should be marked clearly (in the title, or
LI> something) saying that it's for the development version.

Is there some LaTeX/Texinfo magic to set the header accordingly (so it
will show up in the HTML and in the PDF outputs)?

Ted

Re: up to date manual on gnus.org

[Lars Ingebrigtsen]

Ted Zlatanov <tzz@lifelogs.com> writes:

> Is there some LaTeX/Texinfo magic to set the header accordingly (so it
> will show up in the HTML and in the PDF outputs)?

Uhm...  I don't recall.  It's probably easier just to do it with sed.  :-)

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

Re: up to date manual on gnus.org

[Andreas Schwab]

Ted Zlatanov <tzz@lifelogs.com> writes:

> Is there some LaTeX/Texinfo magic to set the header accordingly (so it
> will show up in the HTML and in the PDF outputs)?

@ifset DEVEL
@title Gnus Manual (development version)
@else
@title Gnus Manual
@end ifset

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: up to date manual on gnus.org

[Lars Ingebrigtsen]

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

> @ifset DEVEL
> @title Gnus Manual (development version)
> @else
> @title Gnus Manual
> @end ifset

Great.  I've now pushed this.  And is there some, er, like, -DDEVEL thing
that one can add to the HTML generation thing?

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

Re: up to date manual on gnus.org

[Ted Zlatanov]

On Fri, 04 Feb 2011 11:08:07 -0800 Lars Ingebrigtsen <larsi@gnus.org> wrote: 

LI> Andreas Schwab <schwab@linux-m68k.org> writes:
>> @ifset DEVEL
>> @title Gnus Manual (development version)
>> @else
>> @title Gnus Manual
>> @end ifset

LI> Great.  I've now pushed this.  And is there some, er, like, -DDEVEL thing
LI> that one can add to the HTML generation thing?

I pushed a different implementation that modifies the titles of all the
manuals if they are built that way.  It's all controlled via
overrides.texi (see `make webhack' and `make nowebhack') since I don't
know how to do the -DWEBHACKDEVEL either :)

@ifset...@end ifset and @ifclear...@end ifclear worked for me while
@else didn't.  I think it's better to use them.

Thanks
Ted

Re: up to date manual on gnus.org

[Michael Albinus]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Andreas Schwab <schwab@linux-m68k.org> writes:
>
>> @ifset DEVEL
>> @title Gnus Manual (development version)
>> @else
>> @title Gnus Manual
>> @end ifset
>
> Great.  I've now pushed this.  And is there some, er, like, -DDEVEL thing
> that one can add to the HTML generation thing?

$(MAKEINFO) -D DEVEL ...

$(MAKEINFO) --html -D DEVEL ...

Re: up to date manual on gnus.org

[Ted Zlatanov]

On Fri, 04 Feb 2011 20:30:46 +0100 Michael Albinus <michael.albinus@gmx.de> wrote: 

MA> Lars Ingebrigtsen <larsi@gnus.org> writes:
>> Andreas Schwab <schwab@linux-m68k.org> writes:
>> 
>>> @ifset DEVEL
>>> @title Gnus Manual (development version)
>>> @else
>>> @title Gnus Manual
>>> @end ifset
>> 
>> Great.  I've now pushed this.  And is there some, er, like, -DDEVEL thing
>> that one can add to the HTML generation thing?

MA> $(MAKEINFO) -D DEVEL ...

MA> $(MAKEINFO) --html -D DEVEL ...

Lars, can you or someone else DTRT in texi/Makefile.in?  It would be
nice if we keep overrides.texi as a global way to control manual
settings for all the Gnus manuals IMO.  But it's not a strong opinion
and I'll go with whatever you prefer.

Ted

Re: up to date manual on gnus.org

[Andreas Schwab]

Ted Zlatanov <tzz@lifelogs.com> writes:

> @else didn't.

Sorry, I made that up without checking.

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: up to date manual on gnus.org

[Ted Zlatanov]

On Thu, 03 Feb 2011 11:59:53 -0800 Lars Ingebrigtsen <larsi@gnus.org> wrote: 

LI> Ted Zlatanov <tzz@lifelogs.com> writes:
>> But we should make sure robots.txt says it's not to be indexed so
>> Google searches don't come up with the Git version's manual.  Is all
>> that OK?

LI> Sounds good.  And the pages should be marked clearly (in the title, or
LI> something) saying that it's for the development version.

This thread is what I was referring to earlier about linking the build
of the Gnus manual with the HTML deployment, and setting
gnus-overrides.texi.

Ted