no support for unicode e-mails messages

[Nadav]

When receiving non-English messages, gibberish is shown instead of the message text. It appears to be base64, but I'm not sure.

[Alexander Hungenberg (alexander-hungenberg-gmail)]

could you please add a screenshot?

[Nadav (nadavz)]

I sent myself an e-mail in Hebrew, and this is what happened. Iv'e also confirmed that it is indeed showing the base64 encoding of my message.

[Alexander Hungenberg (alexander-hungenberg-gmail)]

okay, this already was fixed some time ago in trunk...

[Ryou Ohsawa (xr0038)]

hello, I use GMail Notifier v0.9 on Ubuntu 9.10.
this app is very useful. I am pleased to met this notification system. Thanks!

Then, I have the same problem since I receive mails in Japanese.
I have made some modifications on gm-notify.py and get the app to display Japanese charactors

add one module,
++ from email.header import decode_header

and input some lines before cutting the long title.
        # aggregate the titles of the messages... cut the string if longer than 20 chars
        titles = ""
        for title in newmail:
            self.addIndicator(title)
++ if title.find("=?UTF-8?B?")==0:
++ title = decode_header(title).pop()[0]
++ title = title.decode('utf-8')
            if len(title) > 20:
                title = title[0:20] + "..."

at present, it works well.

[Ryou Ohsawa (xr0038)]

sorry, I find a serious bug. my modification work well only in UTF-8.
the code below might work well for other codings.

 # aggregate the titles of the messages... cut the string if longer than 20 chars
 titles = ""
 for title in newmail:
  self.addIndicator(title)
++ if title.find("=?")==0 & title.find("?=")!=-1:
++ title_str = decode_header(title).pop()[0]
++ title_code = decode_header(title).pop()[1]
++ title = title_str.decode(title_code)
  if len(title) > 20:
   title = title[0:20] + "..."

[Alexander Hungenberg (alexander-hungenberg-gmail)]

Hi Ryou,
a fix for this bug is already commited to current trunk because it features a completely new backend. Could you please test this version to see if the bug is fixed?