We should convert article bodies into UTF-8 at the point where they
are brought into Pan (in this case, nntp_body()) rather than in a
piecemeal fashion (works in text.c because it uf8izes before calling
update_body_pane, but fails in Reply because message-window.c doesn't
utf8ize).

This is another instance of the same problem described in 80002.

For discussion: should the .msg files in ~/.pan/cache/ be UTF-8?
It will make coding Pan simpler; do most editors handle UTF-8?

if pan would simple to utf8ize cache, I agree to do that.
simple is right.
we'll have native locale(non-utf8) article by 'save article'.

at this time gedit2 is absolutely support utf8.
vim6; I think I remember I've heard about utf8 support, not sure.

*** Bug 80697 has been marked as a duplicate of this bug. ***

Kang, does this work in 0.11.93 now?

kang: ping

punting NEEDINFO reports from 0.12.0 to 0.12.1.

If the bug reporters come through with more information,
they can be moved back into 0.12.0.

kang: I believe this is working correctly in 0.12.0.  Can you confirm
this ?

yeah, it seems to work with 0.12.1 that I've used.
(sorry to be late. my ISP drop the newsgroup service.. :(

somebody complained me about filter phrase not work,
thou article subject and filter phrase equivalent on display.
I think this is native<->utf8 problem inside,
like this bug 80086, and bug 80004.

tried a little dive from the beginning of phrase entry,
and lost where the filter phrase working to match subjects.

wanted not open another bug. so reopen this close one.

What release was this person using ?



Pan 0.12.94 *does* convert the subject to UTF-8 when it gets the header. The subject filter works fine on my end, even for 8bit characters.



So, could you 

1) validate that this is happening with 0.12.94 ?

2) provide a group + message_id + search criteria that is failing ?



Thanks.

1. I tried and failed with 0.12.94 (pan --version said me)

2. you mean you're giving a try to filter with a phrase?
a. server: news.nuri.net
b. group: han.test
c. phrase: '연습' (means 'trial' in Korean)

item c needs your attention a bit.
A. do you have Korean XIM client, ami installed?
if you have RedHat distro, ami provided may.
B. then, you turn on the im-xim module for gtk2, to allow ami to work
with gtk2.
C. launch pan and go to the specific group.
D. give focus filter entry and type this sequence. 'dustmq'. and
you'll see the Korean character building on entry.
E. press the enter and nothing left on filter.

3. I'm afraid that I misunderstood your order.
how can I send you the raw message that pan transferred?

I'll have a look to see if I can reproduce this using your steps 1&2.

Regarding 3: no need to send any raw messages. Just give us the
Message-ID of the message that you're trying to filter.

(you can see the article's Message-ID by looking at the headers -
shortcut 'h').

From what I can tell, utf8 is working now.
Kang, what is the state of this bug? :)

<WP6r9.58501$v85.3761400@news.bora.net>
is a sample thing on han.binaries.photo.

and I think pan does not convert locale to utf8;
tried save article body and found the file is ok with full text.
but pan's article pane does not show me any text for them.
I suppose pan does not convert the text and gtk2 make the text of
native locale invisible.. :)

and, if it's possible, why don't you convert the filename
of save as.. by G_BROKEN_FILENAME=boolean?

Created attachment 11600 [details]
Kang's message

This bug report has changed contents so many times I can't work out
anymore what it is we're supposed to fix. :)

According to your last entry, you're having trouble displaying that
message. I can read it on my system. Here's what I did:

1. Set LANG=ko_KR.euckr
2. Start Pan
3. Set default charset of han.binaries.photo to 'Korean (euc-kr)'
4. Retrieve all headers for that group
5. Read message with Message-ID you mentioned.
6. Look very puzzled at strange Korean characters. :)

I'll post a screenshot, just so you can see it.

Created attachment 11603 [details]
screenshot

uhm.. yeah, it's nice shot with most ugly Korean font. :)
but I cannot reproduce your method.
there's no character visible, even broken.
I did GDK_USE_XFT=0|1 both.

or, this may be a ridiculous simple problem of mine.
but I cannot figure out what the problem is..
sorry for lack of info.

Haven't been able to reproduce this, but there's been lots of fixes 
in this area since 0.11.91. Could you:

1. upgrade to 0.13.92
2. verify the group's charset is right.
3. download all headers in the group.
4. see if the problem persists?

Thanks.

Closing stale bugreports. If you still see this on 0.14.2 or later,
feel free to reopen and provide more info.