Encoding problems in Brutus-Evolution

[Mikael Karl Karlsson]

Hello,

I have been trying out Brutus-Evolution on Ubuntu 8.04.1. The problem is
that something happen when I send mails with swedish characters.
The reciever gets something like this "dGVzdCDDpcOkw7YNCg==".
In the calendar the swedish characters end up as "@".

When I set up an IMAP account using the brutus-evolution client it works
well, so it must be something with the brutus plugin?

The faulty mails seem to have the right charset and everything:
Content-Type: text/plain; charset=UTF-8
Mime-Version: 1.0

There's no problem with recieving mails with swedish characters. The
only anoying thing with that is that even if I have set UTF-8 as default
encoding both in the mail and in the composer settings I have to go to
View->Character encoding and choose UTF-8 instead of Default to get it
right.

Anyone having the same problem or have any good suggestion? Please tell
if you need more information!

Regards

Mikael Karlsson

[Jules Colding]

Hi Mikael,

Post by Mikael Karl Karlsson
I have been trying out Brutus-Evolution on Ubuntu 8.04.1. The
problem is
that something happen when I send mails with swedish characters.
The reciever gets something like this "dGVzdCDDpcOkw7YNCg==".
When I set up an IMAP account using the brutus-evolution client it works
well, so it must be something with the brutus plugin?

If the encoding fails for Exchange mails but succeeds for IMAP mails
then yes.

Post by Mikael Karl Karlsson
Content-Type: text/plain; charset=UTF-8
Mime-Version: 1.0

So it only looks bad for the receiver, not in your "Sent" folder?

Post by Mikael Karl Karlsson
There's no problem with recieving mails with swedish characters. The
only anoying thing with that is that even if I have set UTF-8 as default
encoding both in the mail and in the composer settings I have to go to
View->Character encoding and choose UTF-8 instead of Default to get it
right.

Yes, I vaguely remember this being a problem in Evolution. Nothing e-b
can do about it I'm afraid.


Best regards,
jules

[Mikael Karlsson]

Post by Jules Colding
Hi Mikael,

Post by Mikael Karl Karlsson
I have been trying out Brutus-Evolution on Ubuntu 8.04.1. The
problem is
that something happen when I send mails with swedish characters.
The reciever gets something like this "dGVzdCDDpcOkw7YNCg==".
When I set up an IMAP account using the brutus-evolution client it works
well, so it must be something with the brutus plugin?

If the encoding fails for Exchange mails but succeeds for IMAP mails
then yes.

Another behavior is that default encoding with imap mails looks good, while
the swedish characters disapears if I choose UTF-8 (weird)... When using
the brutus plugin,
I have to manually set it to UTF-8 on View -> Encoding.. instead of
using default
(which is UTF-8..).

Since Brutus is developed in denmark, shouldn't you have the same problem?

Post by Jules Colding

Post by Mikael Karl Karlsson
Content-Type: text/plain; charset=UTF-8
Mime-Version: 1.0

So it only looks bad for the receiver, not in your "Sent" folder?

It looks bad in the "Sent" folder also. I forgot to check that I think
before writing this mail.

Post by Jules Colding

Post by Mikael Karl Karlsson
There's no problem with recieving mails with swedish characters. The
only anoying thing with that is that even if I have set UTF-8 as default
encoding both in the mail and in the composer settings I have to go to
View->Character encoding and choose UTF-8 instead of Default to get it
right.

Yes, I vaguely remember this being a problem in Evolution. Nothing e-b
can do about it I'm afraid.

Let us hope it will be fixed sometime in 2015 then.. ;)

Post by Jules Colding
Best regards,
jules
_______________________________________________
brutus mailing list
http://www.42tools.com/mailman/listinfo/brutus

[Mikael Karlsson]

Hello again,

In camel/camel-brutus-folder.c the e_iconv_charset_name() is used, which
seem to be deprecated in
newly written code
(http://library.gnome.org/devel/libedataserver/unstable/libedataserver-e-iconv.html).

It seems like the body is examined to get the right charset, which
obviously doesn't work well. I've tried to
analyse the code to understand which functions that are used when
sending mail, but it's kinda hard.
Can I have some hints? It would be nice to know atleast which files to
look in.
What I do understand is that incoming mails are analysed more than
outgoing mails are.

I've also tried to hardcode the charset to utf-8 in the code, but it
didn't work (I'm not very surprised).

Regards

Mikael

[Jules Colding]

Hi Mikael,

Post by Mikael Karlsson
In camel/camel-brutus-folder.c the e_iconv_charset_name() is used, which
seem to be deprecated in
newly written code
(http://library.gnome.org/devel/libedataserver/unstable/libedataserver-e-iconv.html
).
It seems like the body is examined to get the right charset, which
obviously doesn't work well. I've tried to
analyse the code to understand which functions that are used when
sending mail, but it's kinda hard.
Can I have some hints? It would be nice to know atleast which files to
look in.

Work your way backwards starting from "camel/camel-brutus-transport.c"
line 282. Please yell at me if you find anything bad.

Post by Mikael Karlsson
What I do understand is that incoming mails are analysed more than
outgoing mails are.

I've tried to use UTF8 as the default everywhere, but I've recently
noticed that Outlook have a hard time supporting UTF8. Maybe this is
some of the reason?

Post by Mikael Karlsson
I've also tried to hardcode the charset to utf-8 in the code, but it
didn't work (I'm not very surprised).

Neither am I :-(

Thanks,
jules

[Mikael Karlsson]

It seems like there are some problems with case sensitive, ie UTF-8 and
utf-8 can
be handled different, somewhere, maybe..

Maybe put in som e_strdown(charset) here and there?

I'm not a C programmer, but I understand written code, I think...

In evolution-brutus/server/brutus-camel.c we have a
brutus_label_to_cpid("UTF-8") which I dont know
what it does, but anyway.. :)

So, what do you think about the e_strdown thingy?

/ Mikael

Post by Jules Colding
Hi Mikael,

Post by Mikael Karlsson
In camel/camel-brutus-folder.c the e_iconv_charset_name() is used, which
seem to be deprecated in
newly written code
(http://library.gnome.org/devel/libedataserver/unstable/libedataserver-e-iconv.html
).
It seems like the body is examined to get the right charset, which
obviously doesn't work well. I've tried to
analyse the code to understand which functions that are used when
sending mail, but it's kinda hard.
Can I have some hints? It would be nice to know atleast which files to
look in.

Work your way backwards starting from "camel/camel-brutus-transport.c"
line 282. Please yell at me if you find anything bad.

Post by Mikael Karlsson
What I do understand is that incoming mails are analysed more than
outgoing mails are.

I've tried to use UTF8 as the default everywhere, but I've recently
noticed that Outlook have a hard time supporting UTF8. Maybe this is
some of the reason?

Post by Mikael Karlsson
I've also tried to hardcode the charset to utf-8 in the code, but it
didn't work (I'm not very surprised).

Neither am I :-(
Thanks,
jules
_______________________________________________
brutus mailing list
http://www.42tools.com/mailman/listinfo/brutus

[Jules Colding]

Hi Mikael,

Post by Mikael Karlsson
It seems like there are some problems with case sensitive, ie UTF-8 and
utf-8 can
be handled different, somewhere, maybe..
Maybe put in som e_strdown(charset) here and there?
I'm not a C programmer, but I understand written code, I think...

e-b is really in need of other eyes than mine, so your effort are
nonetheless highly appreciated :-)

Post by Mikael Karlsson
In evolution-brutus/server/brutus-camel.c we have a
brutus_label_to_cpid("UTF-8") which I dont know
what it does, but anyway.. :)

It translates the encoding string (e.g. "UTF-8") into a numeric value
understood (or at least accepted) by Outlook.

Post by Mikael Karlsson
So, what do you think about the e_strdown thingy?

It won't have any effect, unfortunately. The encoding strings should
really look as they do.

Thanks,
jules

Post by Mikael Karlsson
/ Mikael

Post by Jules Colding
Hi Mikael,

Post by Mikael Karlsson
In camel/camel-brutus-folder.c the e_iconv_charset_name() is used, which
seem to be deprecated in
newly written code
(http://library.gnome.org/devel/libedataserver/unstable/libedataserver-e-iconv.html
).
It seems like the body is examined to get the right charset, which
obviously doesn't work well. I've tried to
analyse the code to understand which functions that are used when
sending mail, but it's kinda hard.
Can I have some hints? It would be nice to know atleast which files to
look in.

Work your way backwards starting from "camel/camel-brutus-
transport.c"
line 282. Please yell at me if you find anything bad.

Post by Mikael Karlsson
What I do understand is that incoming mails are analysed more than
outgoing mails are.

I've tried to use UTF8 as the default everywhere, but I've recently
noticed that Outlook have a hard time supporting UTF8. Maybe this is
some of the reason?

Post by Mikael Karlsson
I've also tried to hardcode the charset to utf-8 in the code, but it
didn't work (I'm not very surprised).

Neither am I :-(
Thanks,
jules
_______________________________________________
brutus mailing list
http://www.42tools.com/mailman/listinfo/brutus

_______________________________________________
brutus mailing list
http://www.42tools.com/mailman/listinfo/brutus

[Mikael Karlsson]

Could this possibly be a bug in Orbit and not E-B?

Post by Jules Colding
Hi Mikael,

Post by Mikael Karlsson
It seems like there are some problems with case sensitive, ie UTF-8 and
utf-8 can
be handled different, somewhere, maybe..
Maybe put in som e_strdown(charset) here and there?
I'm not a C programmer, but I understand written code, I think...

e-b is really in need of other eyes than mine, so your effort are
nonetheless highly appreciated :-)

Post by Mikael Karlsson
In evolution-brutus/server/brutus-camel.c we have a
brutus_label_to_cpid("UTF-8") which I dont know
what it does, but anyway.. :)

It translates the encoding string (e.g. "UTF-8") into a numeric value
understood (or at least accepted) by Outlook.

Post by Mikael Karlsson
So, what do you think about the e_strdown thingy?

It won't have any effect, unfortunately. The encoding strings should
really look as they do.
Thanks,
jules

Post by Mikael Karlsson
/ Mikael

Post by Jules Colding
Hi Mikael,

Post by Mikael Karlsson
In camel/camel-brutus-folder.c the e_iconv_charset_name() is used, which
seem to be deprecated in
newly written code
(http://library.gnome.org/devel/libedataserver/unstable/libedataserver-e-iconv.html
).
It seems like the body is examined to get the right charset, which
obviously doesn't work well. I've tried to
analyse the code to understand which functions that are used when
sending mail, but it's kinda hard.
Can I have some hints? It would be nice to know atleast which files to
look in.

Work your way backwards starting from "camel/camel-brutus-
transport.c"
line 282. Please yell at me if you find anything bad.

Post by Mikael Karlsson
What I do understand is that incoming mails are analysed more than
outgoing mails are.

I've tried to use UTF8 as the default everywhere, but I've recently
noticed that Outlook have a hard time supporting UTF8. Maybe this is
some of the reason?

Post by Mikael Karlsson
I've also tried to hardcode the charset to utf-8 in the code, but it
didn't work (I'm not very surprised).

Neither am I :-(
Thanks,
jules
_______________________________________________
brutus mailing list
http://www.42tools.com/mailman/listinfo/brutus

_______________________________________________
brutus mailing list
http://www.42tools.com/mailman/listinfo/brutus

_______________________________________________
brutus mailing list
http://www.42tools.com/mailman/listinfo/brutus

[Jules Colding]

Post by Mikael Karlsson
Could this possibly be a bug in Orbit and not E-B?

I doubt the bug is in ORBit2. It is far more likely to be something in
e-d-s or glib. There are so many components playing together here that
the bug has many places to hide.


Best regards,
jules

Post by Mikael Karlsson

Post by Jules Colding
Hi Mikael,

Post by Mikael Karlsson
It seems like there are some problems with case sensitive, ie UTF-8 and
utf-8 can
be handled different, somewhere, maybe..
Maybe put in som e_strdown(charset) here and there?
I'm not a C programmer, but I understand written code, I think...

e-b is really in need of other eyes than mine, so your effort are
nonetheless highly appreciated :-)

Post by Mikael Karlsson
In evolution-brutus/server/brutus-camel.c we have a
brutus_label_to_cpid("UTF-8") which I dont know
what it does, but anyway.. :)

It translates the encoding string (e.g. "UTF-8") into a numeric value
understood (or at least accepted) by Outlook.

Post by Mikael Karlsson
So, what do you think about the e_strdown thingy?

It won't have any effect, unfortunately. The encoding strings should
really look as they do.
Thanks,
jules

Post by Mikael Karlsson
/ Mikael

Post by Jules Colding
Hi Mikael,

Post by Mikael Karlsson
In camel/camel-brutus-folder.c the e_iconv_charset_name() is used, which
seem to be deprecated in
newly written code
(http://library.gnome.org/devel/libedataserver/unstable/libedataserver-e-iconv.html
).
It seems like the body is examined to get the right charset, which
obviously doesn't work well. I've tried to
analyse the code to understand which functions that are used when
sending mail, but it's kinda hard.
Can I have some hints? It would be nice to know atleast which files to
look in.

Work your way backwards starting from "camel/camel-brutus-
transport.c"
line 282. Please yell at me if you find anything bad.

Post by Mikael Karlsson
What I do understand is that incoming mails are analysed more than
outgoing mails are.

I've tried to use UTF8 as the default everywhere, but I've recently
noticed that Outlook have a hard time supporting UTF8. Maybe this is
some of the reason?

Post by Mikael Karlsson
I've also tried to hardcode the charset to utf-8 in the code, but it
didn't work (I'm not very surprised).

Neither am I :-(
Thanks,
jules
_______________________________________________
brutus mailing list
http://www.42tools.com/mailman/listinfo/brutus

_______________________________________________
brutus mailing list
http://www.42tools.com/mailman/listinfo/brutus

_______________________________________________
brutus mailing list
http://www.42tools.com/mailman/listinfo/brutus

_______________________________________________
brutus mailing list
http://www.42tools.com/mailman/listinfo/brutus

[Luis Correia]

Hi all

Post by Jules Colding

Post by Mikael Karlsson
Could this possibly be a bug in Orbit and not E-B?

I doubt the bug is in ORBit2. It is far more likely to be something in
e-d-s or glib. There are so many components playing together here that
the bug has many places to hide.
From my large windows, outlook and exchange, and being a participant

in many mailing lists from the OpenSource world, I know for sure that
Outlook has some issues about the UTF-8 encoding.

While this may not be feasible at all, using ISO-8851* is probably
better, but then again, this may be just me :)

Luis Correia

[Mikael Karlsson]

Post by Luis Correia
Hi all

Post by Jules Colding

Post by Mikael Karlsson
Could this possibly be a bug in Orbit and not E-B?

I doubt the bug is in ORBit2. It is far more likely to be something in
e-d-s or glib. There are so many components playing together here that
the bug has many places to hide.
From my large windows, outlook and exchange, and being a participant

in many mailing lists from the OpenSource world, I know for sure that
Outlook has some issues about the UTF-8 encoding.
While this may not be feasible at all, using ISO-8851* is probably
better, but then again, this may be just me :)

Unfortunatly ISO-8851* doesn't work either.. Same error with a different
look.

If UTF-8 gives EWWERWEQQW=
ISO-8851* gives something like DFASFAS=

:)

/ Mikael

[Jules Colding]

Post by Luis Correia
Hi all

Post by Jules Colding

Post by Mikael Karlsson
Could this possibly be a bug in Orbit and not E-B?

I doubt the bug is in ORBit2. It is far more likely to be something in
e-d-s or glib. There are so many components playing together here that
the bug has many places to hide.
From my large windows, outlook and exchange, and being a participant

in many mailing lists from the OpenSource world, I know for sure that
Outlook has some issues about the UTF-8 encoding.
While this may not be feasible at all, using ISO-8851*

Don't you mean ISO-8859-1?

Post by Luis Correia
is probably
better, but then again, this may be just me :)

but:

http://groups.google.com/group/comp.soft-sys.ace/browse_thread/thread/290c3d2d04b19fce

If only latin-1 could solve it...

Thanks,
jules

[Mikael Karlsson]

Post by Jules Colding

Post by Luis Correia
Hi all

Post by Jules Colding

Post by Mikael Karlsson
Could this possibly be a bug in Orbit and not E-B?

I doubt the bug is in ORBit2. It is far more likely to be something in
e-d-s or glib. There are so many components playing together here that
the bug has many places to hide.

I think I'm totally lost trying to find this problem since I don't know
how the code in E-B is
all put together in all those files. :( Maybe if I watched it for a week
or so.

Post by Jules Colding

Post by Luis Correia

Post by Jules Colding
From my large windows, outlook and exchange, and being a participant

in many mailing lists from the OpenSource world, I know for sure that
Outlook has some issues about the UTF-8 encoding.
While this may not be feasible at all, using ISO-8851*

Don't you mean ISO-8859-1?

I actually did.. Sorry, I just didn't thought about it..

Post by Jules Colding

Post by Luis Correia
is probably
better, but then again, this may be just me :)

http://groups.google.com/group/comp.soft-sys.ace/browse_thread/thread/290c3d2d04b19fce
If only latin-1 could solve it...
Thanks,
jules

[Jules Colding]

Post by Mikael Karlsson

Post by Luis Correia
Hi all

Post by Jules Colding

Post by Mikael Karlsson
Could this possibly be a bug in Orbit and not E-B?

I doubt the bug is in ORBit2. It is far more likely to be something in
e-d-s or glib. There are so many components playing together here that
the bug has many places to hide.

I think I'm totally lost trying to find this problem since I don't know
how the code in E-B is
all put together in all those files. :( Maybe if I watched it for a week
or so.

I know. It's a lot of code, but you'll get used to it sooner or
later ;-)

Best regards,
jules

[Mikael Karlsson]

Post by Jules Colding

Post by Mikael Karlsson

Post by Luis Correia
Hi all

Post by Jules Colding

Post by Mikael Karlsson
Could this possibly be a bug in Orbit and not E-B?

I doubt the bug is in ORBit2. It is far more likely to be something in
e-d-s or glib. There are so many components playing together here that
the bug has many places to hide.

I think I'm totally lost trying to find this problem since I don't know
how the code in E-B is
all put together in all those files. :( Maybe if I watched it for a week
or so.

I know. It's a lot of code, but you'll get used to it sooner or
later ;-)

Unfortunatly I'm not allowed to spend time looking through the code. We
would appreciate
if the encoding problem could be fixed, but I will not continue to be
active with this project
until this problem get higher priority.
For some motivation to solve it, at least 200 people here would be happy
to use Brutus-Evolution in the future.. ;)


Regards

Mikael