SOP or other forward functions get reverse order

SOP or other forward functions get reverse order

[reader]

Anyone else noticing that any of the forwarding commands result in a
message where the entries are in reverse order of how they were
process marked?

I'm pretty sure that used to put the messages in order of process marking.

Re: SOP or other forward functions get reverse order

[Steven E. Harris]

reader@newsguy.com writes:

> Anyone else noticing that any of the forwarding commands result in a
> message where the entries are in reverse order of how they were
> process marked?

Yes, I've complained about this bug before. Well, first I complained
that using `S O p' at the root of a thread would put the messages in
seemingly random order. It was probably deterministic, but not obvious
what sort criteria it used.

The advice was to process mark the messages, on the grounds that
`S O p' was not advertised to choose a meaningful order on its own.

> I'm pretty sure that used to put the messages in order of process
> marking.

Recently I tried forwarding a set of messages, and had to redo it five
or six times because the order seemed backward, despite my having
process-marked the messages in the desired order. In my frustration I
thought that maybe I had misremembered the intended behavior, marked
the messages backward, and finally got the digest sent.

In summary: Yes.

-- 
Steven E. Harris

Re: SOP or other forward functions get reverse order

[Alan Shutko]

[-- Attachment #1: Type: text/plain, Size: 464 bytes --]

"Steven E. Harris" <seh@panix.com> writes:

> Recently I tried forwarding a set of messages, and had to redo it five
> or six times because the order seemed backward, despite my having
> process-marked the messages in the desired order.

Yes, it's broken.  I don't know why, cvs annotate isn't working on
that file for me, but here's the fix I've been using.

-- 
Alan Shutko <ats@acm.org> - I am the rocks.
Balance the budget. Declare politicians a game species!

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: gnus-uu.diff --]
[-- Type: text/x-patch, Size: 532 bytes --]

Index: lisp/gnus-uu.el
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/gnus-uu.el,v
retrieving revision 7.20
diff -u -r7.20 gnus-uu.el
--- lisp/gnus-uu.el	25 Sep 2005 21:27:49 -0000	7.20
+++ lisp/gnus-uu.el	21 Dec 2005 22:57:34 -0000
@@ -1127,7 +1127,7 @@
 	    (setq n (1- n))))
 	(nreverse articles)))
      (gnus-newsgroup-processable
-      (reverse gnus-newsgroup-processable))
+      gnus-newsgroup-processable)
      (t
       (gnus-uu-find-articles-matching)))))
 

Re: SOP or other forward functions get reverse order

[Katsumi Yamaoka]

>>>>> In <87ek4683ca.fsf@vera.springies.com>
>>>>>	Alan Shutko <ats@acm.org> wrote:

> but here's the fix I've been using.

> --- lisp/gnus-uu.el	25 Sep 2005 21:27:49 -0000	7.20
> +++ lisp/gnus-uu.el	21 Dec 2005 22:57:34 -0000
> @@ -1127,7 +1127,7 @@
>  	    (setq n (1- n))))
>  	(nreverse articles)))
>       (gnus-newsgroup-processable
> -      (reverse gnus-newsgroup-processable))
> +      gnus-newsgroup-processable)
>       (t
>        (gnus-uu-find-articles-matching)))))

It will probably be a wrong approach.  All the clauses of that
`cond' form seem to have to return a reversed list of article
numbers.

Re: SOP or other forward functions get reverse order

[Katsumi Yamaoka]

>>>>> In <q947j9ysaml.fsf@chlorine.gnostech.com> Steven E. Harris wrote:

> reader@newsguy.com writes:

>> Anyone else noticing that any of the forwarding commands result in a
>> message where the entries are in reverse order of how they were
>> process marked?

> Yes, I've complained about this bug before. Well, first I complained
> that using `S O p' at the root of a thread would put the messages in
> seemingly random order. It was probably deterministic, but not obvious
> what sort criteria it used.

Ouch!  I realized that that was my fault right now.

2004-09-24  Katsumi Yamaoka  <yamaoka@jpl.org>

	* gnus-uu.el (gnus-uu-digest-mail-forward): Obey the process/prefix
	convention fully; don't miss the root article of a thread; make
	the X-Draft-From header with correct article numbers.

When I changed it, I didn't know the gnus-newsgroup-processable
variable has to have articles in reverse order.  Fixed in the
CVS trunk.

Sorry to all.

Re: SOP or other forward functions get reverse order

[Katsumi Yamaoka]

>>>>> In <b4mbqz9k2xa.fsf@jpl.org> Katsumi Yamaoka wrote:

> Ouch!  I realized that that was my fault right now.
[...]
> Fixed in the CVS trunk.

Please don't check out the current version.  I'll fix it soon.

> Sorry to all.

Re: SOP or other forward functions get reverse order

[Katsumi Yamaoka]

>>>>> In <b4mr785zi34.fsf@jpl.org> Katsumi Yamaoka wrote:

> Please don't check out the current version.  I'll fix it soon.

I've fixed it so as to work even if neither a number of articles
nor a region is specified.  However, it might not work correctly
if the thread is sparse.  So, when you send a digest of articles,
you'd better always specify marks or a region.

Please check it out now.

Re: SOP or other forward functions get reverse order

[reader]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

>>>>>> In <b4mr785zi34.fsf@jpl.org> Katsumi Yamaoka wrote:
>
>> Please don't check out the current version.  I'll fix it soon.
>
> I've fixed it so as to work even if neither a number of articles
> nor a region is specified.  However, it might not work correctly
> if the thread is sparse.  So, when you send a digest of articles,
> you'd better always specify marks or a region.
>

Thanks for your prompt action Katsumi.

Re: SOP or other forward functions get reverse order

[Steven E. Harris]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> So, when you send a digest of articles, you'd better always specify
> marks or a region.

When designating a region, will the articles be ordered from top to
bottom? That seems like the most obvious behavior, but I just want to
make sure.

-- 
Steven E. Harris

Re: SOP or other forward functions get reverse order

[Katsumi Yamaoka]

>>>>> In <83d5jj4oiz.fsf@torus.sehlabs.com>
>>>>>	"Steven E. Harris" <seh@panix.com> wrote:

> Katsumi Yamaoka <yamaoka@jpl.org> writes:

>> So, when you send a digest of articles, you'd better always specify
>> marks or a region.

> When designating a region, will the articles be ordered from top to
> bottom? That seems like the most obvious behavior, but I just want to
> make sure.

Maybe yes, though I'm not quite sure of it.  In that case, the
functions `gnus-summary-work-articles' and `gnus-summary-find-next'
determine the order of articles, which reflects the order
displayed in the summary buffer.

I'm sorry for the late response.