Naohiro Aota | 2 Mar 04:38 2010
Picon

[emacs-w3m:11165] Re: scroll image-buffer

Naohiro Aota <naota <at> elisp.net> writes:

> Katsumi Yamaoka <yamaoka <at> jpl.org> writes:
>
>>> 画像バッファのスクロールに image-mode を使う patch を書いてみました。画像
>>> のサイズがバッファ幅より大きい場合のスクロールが滑らかになるのではないか
>>> と思います。
>>
>> ちょっと試してみました。
>> Emacs 22.2 以上、(not XEmacs)、(not tty) という条件ですね。
>
> おっと… NEWS* は grep したのですが… 22.2 以上でしたか。

とりあえず commit してみました。 また、どんな buffer でも使えるように考え
てみます。

--
青田

David Engster | 3 Mar 22:48 2010
Picon

[emacs-w3m:11166] Re: sb-atom and sb-rss fetching already deleted articles

Katsumi Yamaoka <yamaoka <at> jpl.org> writes:
>>>>>> In [emacs-w3m : No.11162] David Engster wrote:
>> I've implemented the approach I quoted above and attached a patch
>> against current CVS.  The advantage of this approach is that it works
>> more generally, without the need for an extra parameter.
>
> Oh, I understood what you meant at last.  It picks articles of
> which the date is newer than the ones having already been fetched,
> and saves articles of which the date cannot be identified.  That's
> an excellent idea!

I had to change this a bit. Turns out there are feeds out there which
happily introduce new items with *older* dates (the 'topthemen' from the
sueddeutsche-de shimbun, for instance). To make those items show up, I
now compare with the date of the *oldest* existing article in the
group. This should still make sure that the article wasn't expired and
is actually a new one.

I also made sure that the returned items are sorted by ascending
date. This is to make sure that newer articles get higher article
numbers in Gnus, since in the expiry process, it is the article with the
highest article number which is kept and never deleted.

It's still not perfect, though. Imagine a feed introduces one new item
with an old date; it will get the highest article number in Gnus. If you
now completely expire the group, only this article will remain. When you
now refresh the group, all items with a newer date will again show
up. However, I think this is a very rare case - it would need a feed
which introduces new items with old dates and a pretty short expiry
time. If this really turns out to be problem, it could still be fixed in
(Continue reading)

jidanni | 3 Mar 20:41 2010

[emacs-w3m:11167] Please stop rendering "昆明街 二 段" exactly the same as "昆明街二段"

Sorry to mention it again, but I wish (emacs-)w3m would act like lynx
and Firefox and please stop rendering "昆明街 二 段" exactly the same as
"昆明街二段". It is unfair to UTF-8 users. Test file:

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
"http://www.w3.org/TR/html4/strict.dtd">

<html lang="zh-tw">
<head>
 <meta http-equiv="Content-Type" content=
 "text/html; charset=utf-8">
 <meta http-equiv="Content-Language" content="zh-tw">

 <title>Test</title>
</head>

<body>
 <p>Please stop rendering "昆明街 二 段" exactly the same as "昆明街二段"!</p>
</body>
</html>

Katsumi Yamaoka | 4 Mar 03:14 2010
X-Face

[emacs-w3m:11168] Re: Please stop rendering "昆明街 二 段" exactly the same as "昆明街二段"

>>>>> In [emacs-w3m : No.11167] jidanni <at> jidanni.org wrote:
> Sorry to mention it again, but I wish (emacs-)w3m would act like lynx
> and Firefox and please stop rendering "昆明街 二 段" exactly the same as
> "昆明街二段". It is unfair to UTF-8 users. Test file:

Unfortunately there seems no chance to improve emacs-w3m so as to
let those SPCs remain since it is what w3m does.  This message is
CC'd to the w3m list.

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
"http://www.w3.org/TR/html4/strict.dtd">

<html lang="zh-tw">
<head>
 <meta http-equiv="Content-Type" content=
 "text/html; charset=utf-8">
 <meta http-equiv="Content-Language" content="zh-tw">

 <title>Test</title>
</head>

<body>
 <p>Please stop rendering "昆明街 二 段" exactly the same as "昆明街二段"!</p>
</body>
</html>

Katsumi Yamaoka | 4 Mar 04:14 2010
X-Face

[emacs-w3m:11169] Re: sb-atom and sb-rss fetching already deleted articles

>>>>> In [emacs-w3m : No.11166] David Engster wrote:
> I had to change this a bit. Turns out there are feeds out there which
> happily introduce new items with *older* dates (the 'topthemen' from the
> sueddeutsche-de shimbun, for instance). To make those items show up, I
> now compare with the date of the *oldest* existing article in the
> group. This should still make sure that the article wasn't expired and
> is actually a new one.

> I also made sure that the returned items are sorted by ascending
> date. This is to make sure that newer articles get higher article
> numbers in Gnus, since in the expiry process, it is the article with the
> highest article number which is kept and never deleted.

I see.  The beauty of it is sorting headers in the ascending order
before comparing with existing old headers.

> It's still not perfect, though. Imagine a feed introduces one new item
> with an old date; it will get the highest article number in Gnus. If you
> now completely expire the group, only this article will remain. When you
> now refresh the group, all items with a newer date will again show
> up. However, I think this is a very rare case - it would need a feed
> which introduces new items with old dates and a pretty short expiry
> time. If this really turns out to be problem, it could still be fixed in
> nnshimbun.el.

Ok.  I can imagine it would not be a trouble even if read articles
appear on rare occasions.

[...]

(Continue reading)

Allan Gottlieb | 4 Mar 05:05 2010
Picon

[emacs-w3m:11170] w3m-reload-this-page moves to beginning

I use gentoo and have the following versions of
w3m and emacs-w3m

app-emacs/emacs-w3m
Installed versions:  1.4.379_pre20100125

www-client/w3m
Installed versions:  0.5.2-r2

When I issue w3m-reload-this-page
the page reloads fine but focus moves to the beginning of the page.
I believe this is new behavior.  I believe focus used to stay
where it was before the command.

As a work around I now use

  (let ((old-point (point)))
    (w3m-reload-this-page)
    (sit-for 3 'no-disp)
    (goto-char old-point))

Is there an option to w3m and/or emacs-w3m that would restore the old
behavior in which focus stayed at point without my current
workaround?

Thank you for w3m and emacs-w3m, which work very well.

allan

(Continue reading)

Katsumi Yamaoka | 4 Mar 05:46 2010
X-Face

[emacs-w3m:11171] Re: w3m-reload-this-page moves to beginning

>>>>> In [emacs-w3m : No.11170] Allan Gottlieb wrote:
> I use gentoo and have the following versions of
> w3m and emacs-w3m

> app-emacs/emacs-w3m
> Installed versions:  1.4.379_pre20100125

Could you try 1.4.380 or later?  I've fixed the bug you pointed
out.  You can get the latest emacs-w3m from:

http://cvs.namazu.org/emacs-w3m.tar.gz

Thanks.

jidanni | 4 Mar 18:01 2010

[emacs-w3m:11172] Re: Please stop rendering "昆明街 二 段" exactly the same as "昆明街二段"

>>>>> "KY" == Katsumi Yamaoka <yamaoka <at> jpl.org> writes:

KY> Unfortunately there seems no chance to improve emacs-w3m so as to
KY> let those SPCs remain since it is what w3m does.  This message is
KY> CC'd to the w3m list.

Thanks. It drives one nuts:

$ for browser in w3m lynx; do echo ==== $browser:; $browser -dump http://pots.tw/node/4415|grep -A 5
對談人; done
==== w3m:
日期對談人主題
3/19 平路我與「她」—穿越虛實的小說與人生
3/26 伊苞鼓動—聽見生命的聲音
4/2 韓良露生命可以如此豐盛
4/9 賀照緹選擇另一種生活,並且活在其中
4/16 吳碧容三十而立,舞動人生
==== lynx:
   日期 對談人 主題
   3/19 平 路 我與「她」—穿越虛實的小說與人生
   3/26 伊 苞 鼓動—聽見生命的聲音
   4/2 韓良露 生命可以如此豐盛
   4/9 賀照緹 選擇另一種生活,並且活在其中
   4/16 吳碧容 三十而立,舞動人生

jidanni | 4 Mar 18:33 2010

[emacs-w3m:11173] ffap vs. URLs in UTF-8

Just curious, how do you emacs-w3m pros "click" on these links that are
not linked on the following page. E.g.,
$ w3m -dump -cols 999 http://pots.tw/node/4415|perl -nwle '/\S(http\S+)/&&print $1'
http://tw.myblog.yahoo.com/taipeigayinn/
http://coswas.org/01coswas/966
http://www.psygarden.com.tw/images/activities/MjAxMDAyMDgxMzE2NDI=/MjAxMDAyMDgxMzE2NDI=.03redingclub-signup.doc
http://spreadsheets.google.com/viewform?formkey=dGEzVG95Vl9Mc3lCZWVVdG5jem0zX2c6MA
http://tinyurl.com/ybfjulc
http://0rz.tw/XGtTg下載報名表,填寫完後email到CLPENG <at> hotline.org.tw,報名時間至3/21截止。
http://woman1.pixnet.net/blog)
http://spreadsheets.google.com/viewform?formkey=dE5IYWtzSlQwUl9kMGZEMVRULXhDbUE6MA
http://msmtaiwan.blogspot.
http://www.wretch.cc/blog/moongirl1225/21391679
http://www.wretch.cc/blog/taboo126/21681160

ffap fails to detect them because its matcher is too strict, so C-x C-f
won't work. One ends up needing to use C-k, then C-x C-f ... C-y. Yuck.

My point of this message is I hope someone will change emacs' ffap
regexp, as posting to bug-gnu-emacs didn't get any change ever.

Katsumi Yamaoka | 5 Mar 02:19 2010
X-Face

[emacs-w3m:11175] Re: Meta charset differs from Content-Type charset

>>>>> In [emacs-w3m : No.11174] 山岡 wrote:
> 前回のものは shimbun が決めた charset で meta tag の内容を書き換
> えてしまうものでしたが、逆に meta tag の指定を優先するものを CVS
> に入れました。変な meta tag や Emacs が知らない charset は無視し
> ます。その charset が正しいことを、同名の coding system があるか
> どうかで判断するという Gnus の悪習を持ち込みました。

すみません、ちょっと修正しました。このメールを見る前にチェックア
ウトしてしまった人はやり直してくださいませ。
--

-- 
山岡