I've noticed recently that I'm getting some mojibake in messages from
a few sources. Both examples I have handy have a quoted-printable UTF-8
encoded text/html part, and one also has a quoted-printable UTF-8
encoded text/plain part.
The one which is HTML only happens also to be in German, and
Just out of curiousity ... did you produce this message with nmh? I ask
because it didn't actually have proper MIME encoding. This can happen
when you use nmh but don't run mhbuild on the message, for example (my
plans are to make this automatically happen for 1.6).
I've noticed recently that
The munged character in your fist example looks like it's
supposed to be c3 bc c3, but instead is 83 c2 bc, if I did
that right. It takes more than one step to get from here to
there, such as losing bits and wrong endian?
Actually, I think Joel was trying to say für, which has the middle
letter
Hi Joel,
The one which is HTML only happens also to be in German, and what's
getting munged are the umlauted vowels
nmh doesn't interpret HTML, so what command is being run to do that and
with what options, e.g. lynx. `ps xfww' in another terminal whilst the
corruption is being viewed might
Thus spake Ken Hornstein:
Just out of curiousity ... did you produce this message with nmh? I ask
because it didn't actually have proper MIME encoding. This can happen
when you use nmh but don't run mhbuild on the message, for example (my
plans are to make this automatically happen for 1.6).
Hi Joel,
I think I found something related to the cause: I have this line in my
mhn.defaults:
mhshow-show-text/html: %p/usr/bin/lynx -force_html -dump '%f' | less
I have a similar single line.
mhshow-show-text/html: lynx -dump -width `tput cols` '%F' |
expand | sed 's/ *$//'
It's just as you said---I didn't run mhbuild. Having that run
automatically wouldn't be a bad idea. Is there a way to have that run
automatically in 1.5?
Yeah, if you put:
automimeproc: 1
in your mh_profile, that will do it.
That has some side effects; any line beginning with a '#' will be
Thus spake Ralph Corderoy:
Hi Joel,
I think I found something related to the cause: I have this line in my
mhn.defaults:
mhshow-show-text/html: %p/usr/bin/lynx -force_html -dump '%f' | less
I have a similar single line.
mhshow-show-text/html: lynx -dump -width `tput cols`
Does one of the MIME RFCs require an encoding declaration when the
charset is already given in the MIME header? If so, then virtually
every email in my inbox which has a text/html part is wrong:
It's sort of a corner case, but RFC 2854 says that the charset MIME
parameter indicates the character
Hi Joel,
I find it gives the behaviour you describe if the HTML file contains a
charset declaration that's incorrect for the content of the HTML, e.g.
ISO-8859-1 when mhstore shows the glyph is UTF-8 encoded. Lynx obeys
the charset in the file. Might be worth searching for `charset' in
Thus spake Ralph Corderoy:
Perhaps see if you can get `lynx -dump' as specified in your
~/.mh_profile to give correct output on a HTML file you create, feeding
the output to cat one time and less another.
If I add -assume_charset=UTF-8 as an option for lynx, both the HTML
written by mhstore
The problem with that is that the HTML being sent to lynx might not be
in a UTF-8 compatible encoding.
Right, that's what I was talking about in my message. That doesn't
currently exist. %a puts all of the MIME parameters out there, but that's
inconvenient to deal with. That should be easy
Thus spake Ken Hornstein:
The problem with that is that the HTML being sent to lynx might not be
in a UTF-8 compatible encoding.
Right, that's what I was talking about in my message. That doesn't
currently exist. %a puts all of the MIME parameters out there, but that's
inconvenient to
I didn't know about %a. It turns out that's sufficient, if kludgy:
mhshow-show-text/html: %p/usr/bin/lynx -assume_charset=`echo %a | sed
's/.*charset=\([^]\+\).*/\1/'` -force_html -dump '%f' | less
The only problem with that is if charset doesn't exist as a parameter,
you don't get the default
Challenge accepted:
mhshow-show-text/html: %p/usr/bin/lynx `echo %a | sed -n
's/.*charset=\([^]\+\).*/-assume_charset=\1/p'` -force_html -dump '%f' |
less
Well played :-)
Although ... will that end up with a zero-length parameter to lynx? Or
will that be eaten by the shell? I think you
Thus spake Ken Hornstein:
Challenge accepted:
mhshow-show-text/html: %p/usr/bin/lynx `echo %a | sed -n
's/.*charset=\([^]\+\).*/-assume_charset=\1/p'` -force_html -dump '%f' |
less
Well played :-)
Although ... will that end up with a zero-length parameter to lynx? Or
will that be
Thus spake Ken Hornstein:
So, I just found out about this:
http://offlineimap.org
It seems like it's close to what people are interested in. The big wrinkle
is that right now the local store is Maildir; it occurs to me that it
should be straightforward to add nmh folder support to
It seems to me there are two IMAP-related things people have wanted:
1) Have nmh commands act on IMAP-stored messages.
2) Expose an nmh folder via IMAP.
#1 would let you use nmh against arbitrary IMAP accounts. #2 would
let you access your existing nmh storage over IMAP. I think both would
On Oct 24, 2013, at 11:35 AM, Joel Uckelman uckel...@nomic.net wrote:
I looked this over a bit and wasn't able to satisfy myself as to what
OfflineIMAP would do.
For many years now my primary email engine has been IMAP. The driving force
behind this is that I need to access my mail folders
But that describes a simple 1:1 mapping case. Nothing there says the
proof extends to the 1:n mapping case (i.e. multiple clients).
Fair enough ... I'm just trying to imagine exactly a) what conflicting
clients would be doing, exactly, to conflict, and b) what SHOULD happen
when a conflict
On 24 Oct 2013, at 18:17, Ken Hornstein k...@pobox.com wrote:
But that describes a simple 1:1 mapping case. Nothing there says the
proof extends to the 1:n mapping case (i.e. multiple clients).
Fair enough ... I'm just trying to imagine exactly a) what conflicting
clients would be doing,
On Oct 24, 2013, at 6:17 PM, Ken Hornstein k...@pobox.com wrote:
But that describes a simple 1:1 mapping case. Nothing there says the
proof extends to the 1:n mapping case (i.e. multiple clients).
Fair enough ... I'm just trying to imagine exactly a) what conflicting
clients would be
Wouldn't a FUSE IMAP layer largely solve the problem of conflicts
by working on the live data store? Perhaps a customization of
something like:
http://imapfs.sourceforge.net/
http://www.sr71.net/projects/gmailfs/
___
Nmh-workers mailing list
On Oct 24, 2013, at 7:22 PM, Jerrad Pierce belg4...@pthbb.org wrote:
Wouldn't a FUSE IMAP layer largely solve the problem of conflicts
by working on the live data store? Perhaps a customization of
something like:
No. nmh works in many places FUSE will never enjoin.
It's mostly about refile and delete, in the IMAP-MH case. Client
A deletes message 1. Client B moves it to folder foo. Who wins?
Especially when B syncs after A, thus message 1 is no longer in place on
the server. (These are *very* simple examples of what you have to deal
with ...)
It looks
On Oct 24, 2013, at 7:24 PM, Ken Hornstein k...@pobox.com wrote:
Of course, the question really should be: what SHOULD happen in that
case?
Like I said, it gets complicated.
The internet calendaring folks are still having ulcers over this.
___
Wouldn't a FUSE IMAP layer largely solve the problem of conflicts
by working on the live data store? Perhaps a customization of
something like:
Well, I note this from the FUSE IMAP web page:
Note: My project is old and has a lot of unresolved issues and design
problems. Please don't try to
have to manage the mapping between MH message numbers and IMAP messages,
which involves a synchronization process. Also, it just seems like to
Yeah, if I were doing it I'd probably not support all of MH's numbering
and sequence goodness... for simplicity/sanity's sake. Just allow basic
sortm to
have to manage the mapping between MH message numbers and IMAP messages,
which involves a synchronization process. Also, it just seems like to
Yeah, if I were doing it I'd probably not support all of MH's numbering
and sequence goodness... for simplicity/sanity's sake. Just allow basic
sortm to
For anyone who's looking to implement something new, I think that
git's internal architecture might be a good starting place, written
in C. In slightly related news, I looked at notmuch recently, and
it made me miss MH.
I've heard that before, but I don't see how looking at git helps
anything
30 matches
Mail list logo