pan-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Pan-devel] Patch to restore "Followup and Reply" and "Go/Top of Thr


From: Charles Kerr
Subject: Re: [Pan-devel] Patch to restore "Followup and Reply" and "Go/Top of Thread"
Date: Mon, 18 Aug 2003 09:15:39 -0700
User-agent: Mutt/1.5.4i

On Sun, Aug 17, 2003 at 10:00:14PM +0200, Torstein Sunde wrote:
> Charles Kerr wrote:
> 
> > [...] I wish you wouldn't apply your patches to the Mandake RPMs.
> > That's what happened with the whole GNet mess.
> 
> Sorry to say so, but I'm not interested in making a set of RPMs I won't
> use in any case.
> 
> As far as I can see, and unlike the "GNet mess", these small patches
> don't involve any extra library prerequisites, only functions that
> Pan is using anyway.

First, to be clear: I appreciate the time you've taken reporting Pan bugs,
I appreciate the time you've taken rolling RPMs, and I appreciate the time
you've taken on this patch.

However, your patch is not a good idea for 0.14.1:

(1) I don't like the Message-ID filter because nobody will use it.
    It only makes sense in the context of "watch subthread" and
    "ignore subthread", so I'd rather make that context explicit by
    adding those two buttons to the article menu instead.

(2) I *hate* SCREEN_SIZE_ABOVE_SVGA.  A compile-time flag for screen size?
    The space eaten used up by the unneeded Message-ID filter makes the
    dialog too big, and rather than deal with this problem you've punted
    it to the RPM builders, who won't know about the compile-time flag,
    and if they did, would have to be clairvoyant to know what screen size
    their users have.

(3) I don't like your running roughshod over the development cycle by
    applying patches to your RPMs.  You're very helpful, and I don't
    want to run you off, but IMO Chris and I are the only ones with
    enough sweat equity to have final say of what goes into Pan.
    The GPL gives you a legal right to add patches, but I won't mirror them.

(4) Your new strings will not be translated because they're not in CVS.
    It's not in CVS for the reasons above, and for the fact that we're
    in a string freeze.

Here's what I'm going to do:

* I'll put "watch subthread" and "ignore subthread" buttons
  into CVS after 0.14.1 comes out this week.

* I'll put up on rebelbase.com any patch-free RPMs you submit for 0.14.1
  and beyond.

* I'll accept a patch if you want to get "watch subthread" and
  "ignore subthread" into CVS sooner rather than later.
  It can be done with trivial modifications to article-actions.c:
  article_action_something_thread() is a function that
  article_action_{watch|ignore}_thread() wrap around.
  It has this two-clause code:

     section_name = score_create_section_str(article);
     add_items[0].on = TRUE;
     add_items[0].negate = FALSE;
     add_items[0].key = HEADER_MESSAGE_ID;
     add_items[0].value = filter_phrase_create_regex (thread_message_id, 
PHRASE_MATCH_IS);
     add_items[1].on = TRUE;
     add_items[1].negate = FALSE;
     add_items[1].key = HEADER_REFERENCES;
     add_items[1].value = filter_phrase_create_regex (thread_message_id, 
PHRASE_MATCH_STARTS_WITH);
     score_add (score_name, section_name, do_list[action].score, TRUE, 31, 
AGGREGATE_TYPE_OR, add_items, 2, TRUE);

  All that needs to be done is to add
  article_action_{watch|ignore}_subthread functions that also wrap
  around something(), and add a second argument to something that takes
  a FilterAggregateType so that we can set the second clause's
  match type to either STARTS_WITH or CONTAINS.
  (STARTS_WITH is faster, so we should keep it for watch/ignore thread).

-- 
cheers,
Charles




reply via email to

[Prev in Thread] Current Thread [Next in Thread]