[luc-mobile]

I have an OS 4.5.0.110 Curve and my BIS plan gives me 11 mailboxes: an instantaneous one, <user@carrier.blackberry.com>, and 10 mailboxes that are polled every so many minutes. Of course, the instantaneous is great, that's genuine push, but it can't be integrated with my own site domain.

That's never been a problem. I've had, for quite a few years, a mail server and a catch-all system with a powerful script that identifies my most important address in the To: field of any message and sends a copy of the message to my <user@carrier.blackberry.com> address whenever there is a match. That arrangement worked fine for two years.

The problem began a few days or maybe about a month ago. That copy that my script sends to the BB now may reach my phone with the subject line all messed up whenever there are accented characters in the Subject line. Accented characters are very common in my native language. Example:

Subject: Re: Isto é acentuação 8

becomes...

Subject: =?Windows-1252?B?UmU6IElzdG8g6SBhY2VudHVh5+NvIDg=?=

Interesting that the Subject line reaches my BlackBerry intact the first time. The problem begins AFTER I REPLY TO IT FROM THE BLACKBERRY. People will get my reply and I suppose it looks good, because no one has complained and it does look good on the reply tests I have sent myself. But whenever they reply on my reply (or I reply to my reply when I'm testing), I get the ugly Subject message. Confusing? Let me explain it another way:

- Someone writes to me.
- Message reaches my BlackBerry and my desktop and looks correct on both.
- I reply from my BlackBerry.
- The other person receives my reply. I assume the Subject line still looks correct. It looks OK on my tests.
- I receive the person's reply. The Subject line is ruined.

I tried to be astute: I changed the redirecting script to detect and decipher the Subject line. It actually decodes the line and replaces it with an iso-8859-1 transcription. It didn't work.

Then I noticed this in the replies written on the BlackBerry:
Content-Type: text/plain; charset="Windows-1252"

So I went further: I had the script rewrite the Subject line and replace charset="Windows-1252 with charset="iso-8859-1". Still no go. Apparently, RIM has a diehard predilection for the Windows-1252 charset.

What bugs me even more is that I use Linux. My e-mail client isn't adding that Windows-1252 charset. It could happen on other people's machines, but not on mine. RIM is adding that junk, and their own device can't read the Windows-1252 encoded Subject lines correctly.

Does anyone have any suggestion for me to fix it? I don't think I should have to upgrade the OS. The problem began about 1 month ago, and I've been running OS 4.5.0.110 since February or March.