(The following is based on 0.9.4 and Linux)
As a test on the mstor-files, I copied a mstor-file to the var/spool/mail directory where the pop3 deamon would open it. Then I connected to the POP3 Server and tried to load messages from there.
The Pop3 Server did not provide the messages out of the mstor file and complained about the From_ lines. After digging around I found:
- mstor saves the From_ lines with a date formatted in the current locale. ("Do Jun 9 10:11:12 2005") (German locale)
- when manually changing the Dates in the mstor-files, the Linux-Pop3-Server did not complain. ("Do" ->"Thu")
- Ive applied the following changes to the sources:
in MboxFile.java and in MetaDateFormat.java:
from:
SimpleDateFormat(<DATE_FORMAT>)
to:
SimpleDateFormat(<DATE_FORMAT>, Locale.US)
- after those changes, mstor saves the received-Date in the From_ line in the US-Format ("Thu Jun 9 10:11:12 2005")
- now I can copy the mstor file to the POP3 Server and it wont complain.
Another thing Id like to mention (but the pop Server doesnt complain about it):
If I open an mstor File on Linux with vi, all lines except the From_ lines end with "^M". (I cannot tell, which format would be right, I just notized it...)
my 2 cents
Thomas
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thanks for the feedback, I've now updated the date formats to be US-specific as you suggested.
With the "^M" characters, these are actually carriage return (CR) characters present in Windows-centric ascii files. I think their presence in the mbox file may be a symptom of the file encoding I am using: "ISO-8859-1". I had changed this to use the default file encoding for the operating system, however this seemed to cause problems for some Linux users so I changed it back.
I'm sure there must be some file encoding that is most appropriate for an mbox file (UTF8 perhaps?) but I haven't come across any definitive answer yet.
If you want to experiment with different file encodings you can specify the system property:
mstor.mbox.encoding=<some_encoding>
regards,
ben
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Regarding the ^M:
It usually comes up, when a file from Windows is copied over to a Linux box and doesnt get converted. (Win uses CRLF as End of Line, where Linux/Unix just uses LF - so the unwanted CR aka ^M ist displayed).
What happens inside mstor I currently dont know. But when the need arises, we could still dig into it...
Thomas
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello Ben!
First of all, many thanks for mstor!
(The following is based on 0.9.4 and Linux)
As a test on the mstor-files, I copied a mstor-file to the var/spool/mail directory where the pop3 deamon would open it. Then I connected to the POP3 Server and tried to load messages from there.
The Pop3 Server did not provide the messages out of the mstor file and complained about the From_ lines. After digging around I found:
- mstor saves the From_ lines with a date formatted in the current locale. ("Do Jun 9 10:11:12 2005") (German locale)
- when manually changing the Dates in the mstor-files, the Linux-Pop3-Server did not complain. ("Do" ->"Thu")
- Ive applied the following changes to the sources:
in MboxFile.java and in MetaDateFormat.java:
from:
SimpleDateFormat(<DATE_FORMAT>)
to:
SimpleDateFormat(<DATE_FORMAT>, Locale.US)
- after those changes, mstor saves the received-Date in the From_ line in the US-Format ("Thu Jun 9 10:11:12 2005")
- now I can copy the mstor file to the POP3 Server and it wont complain.
Another thing Id like to mention (but the pop Server doesnt complain about it):
If I open an mstor File on Linux with vi, all lines except the From_ lines end with "^M". (I cannot tell, which format would be right, I just notized it...)
my 2 cents
Thomas
Hi Thomas,
Thanks for the feedback, I've now updated the date formats to be US-specific as you suggested.
With the "^M" characters, these are actually carriage return (CR) characters present in Windows-centric ascii files. I think their presence in the mbox file may be a symptom of the file encoding I am using: "ISO-8859-1". I had changed this to use the default file encoding for the operating system, however this seemed to cause problems for some Linux users so I changed it back.
I'm sure there must be some file encoding that is most appropriate for an mbox file (UTF8 perhaps?) but I haven't come across any definitive answer yet.
If you want to experiment with different file encodings you can specify the system property:
mstor.mbox.encoding=<some_encoding>
regards,
ben
HI Ben,
Many thanks, that was a quick reply.
Regarding the ^M:
It usually comes up, when a file from Windows is copied over to a Linux box and doesnt get converted. (Win uses CRLF as End of Line, where Linux/Unix just uses LF - so the unwanted CR aka ^M ist displayed).
What happens inside mstor I currently dont know. But when the need arises, we could still dig into it...
Thomas