Hi,
Thanks for interesting tool. This is helping me to move my wife's old
Windows mail to a diifferent platform.
I realized that old windows (95,98,Me) versions used non-utf-8 encodings
for different countries. So when one extracts the content of pst file,
filenames are encoded in traditional non-utf-8 format (Names like Inbox
are translated into each language in pst).
readpst just writes its content to the filename as is. In case of Japan,
it was Shift-JIS. But my sid/etch box is utf-8.
You need to convert filenames to UTF-8 encoding on modrn Linux system
and all other new platforms (Mac, NT, ...).
So having capability to pipe filenames created by readpst with iconv
comannd will be good thing to do if options are given.
-f shift_jis
-t utf-8
(Just use iconv or equivalent libray calls, please)
Also, the file contents seems to be in the traditional non-utf-8 format,
i.e., shift-jis for Japan. This issue needs a bit more investigation.
Simple run of iconv will error out. -c option helps processig them but
makes some mail with "mojibake", i.e., decoding error. Funny thing is
even if original mail is 7 bit encoded iso-2022-jp file, stored content
seems to be in shift-jis thus readable on utf-8 console running vim.
(Unix mbox usually keeps original 7 bit encoding while storing it. MS
may be doing some shortcut here since pst is their proprietary format.)
So far just running iconv to convert entire generated file does the
decent job.
For multipart plain/text contents, I needed to change the stratig part:
to be
for mutt to read them OK. (Non-extensive test. vim was always able to
read it since it does not care these mailbox specific encoding
directives.)
For mutt on Linux, I may have other ways to read it but this was problem
moving file to Mac. This kind of rewrite is best done in readpst.
Once you adress the first issue, please downgrade this to the wish list.
"Lacks capability to adress encoding issue for the mail content".
I will post more detailed report soon.
I think addressing issues here properly will address non-ASCII people
(eastern europe, Asia, ... even ISO-8859-1).