Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Emacs 23

0 views
Skip to first unread message

Uday S Reddy

unread,
Nov 20, 2009, 9:57:23 AM11/20/09
to
Dear all,

Emacs 23 was released only in July 2009 and VM hasn't been maintained
since November 2008. Moreover, Emacs 23 seems to be a significant
revision over Emacs 22 with all character coding mechanisms changed.
So, I think it will be some time before we can "certify" that VM works
correctly under Emacs 23.

Meanwhile, please feel free to *test* out VM with Emacs 23 and file bug
reports for the problems you find. But the keyword is *test*. Don't
depend on VM with Emacs 23 for your daily work.

I myself still use Emacs 22.2 for daily work.

Cheers,
Uday

blueman

unread,
Nov 22, 2009, 4:18:41 AM11/22/09
to

So far, Emacs 23 has been working ok for me under vm...

Uday S Reddy

unread,
Nov 22, 2009, 5:34:14 AM11/22/09
to
blueman wrote:

> So far, Emacs 23 has been working ok for me under vm...

That is good to know. Ulrich has also written to me that Emacs 23 has
been working fine for him.

When VM is working correctly, it only writes ASCII characters into the
mail folders. These should work fine under any version of Emacs. You
can verify this by going to the folder buffer (not the presentation
buffer), and checking the variable buffer-file-coding-system. If it is
undecided or raw-text, then your VM folder is as it should be. If it
responds with some other character set, then it is possible that the
folder is not portable across different Emacsen.

Cheers,
Uday

Julian Bradfield

unread,
Nov 22, 2009, 6:47:48 AM11/22/09
to
On 2009-11-22, Uday S Reddy <uDOTsD...@cs.bham.ac.uk> wrote:
> When VM is working correctly, it only writes ASCII characters into the
> mail folders.

Nonsense. My mailfolders are full of high-bit characters. Nowadays,
mostly as part of UTF-8.

Which is compatible with ...

> buffer), and checking the variable buffer-file-coding-system. If it is
> undecided or raw-text, then your VM folder is as it should be. If it

since raw-text allows binary characters.

Uday S Reddy

unread,
Nov 22, 2009, 8:26:37 AM11/22/09
to
Julian Bradfield wrote:
> On 2009-11-22, Uday S Reddy <uDOTsD...@cs.bham.ac.uk> wrote:
>> When VM is working correctly, it only writes ASCII characters into the
>> mail folders.
>
> Nonsense. My mailfolders are full of high-bit characters. Nowadays,
> mostly as part of UTF-8.

Do you mean to say the mail system now guarantees 8 bit transport? Is
there any RFC on it? Or, is your VM converting the messages to 8-bit
before storing them?

Looking through my mail folder, I found several messages that claimed to
be "binary" or "8bit" UTF-8, but they really had only ASCII characters.
All messages that had any real 8-bit content were MIME encoded.

> since raw-text allows binary characters.

Ok, but how would Emacs & VM know what to do with such characters?

Cheers,
Uday

Julian Bradfield

unread,
Nov 22, 2009, 10:38:38 AM11/22/09
to
On 2009-11-22, Uday S Reddy <uDOTsD...@cs.bham.ac.uk> wrote:
> Julian Bradfield wrote:
>> Nonsense. My mailfolders are full of high-bit characters. Nowadays,
>> mostly as part of UTF-8.

> Do you mean to say the mail system now guarantees 8 bit transport? Is
> there any RFC on it? Or, is your VM converting the messages to 8-bit
> before storing them?

Unless your MTA is antique, it most probably supports the SMTP
8BITMIME (RFC1652) extension. Whether it chooses to deliver mail to
the local inbox in binary is up to the administrator, I guess.

> Ok, but how would Emacs & VM know what to do with such characters?

Emacs mustn't do anything - that's why it's vital that all mail files
are read in binary (or raw-text), per the patches of mine you just
incorporated. VM uses the MIME headers to know what to do with them.
(Of course, the MIME headers don't apply to the headers, so it's
always wrong to store non-ASCII characters in headers.)

mikko_h

unread,
Nov 24, 2009, 1:07:10 PM11/24/09
to
On Nov 22, 12:34 pm, Uday S Reddy <uDOTsDOTre...@cs.bham.ac.uk> wrote:
> blueman wrote:
> > So far, Emacs 23 has been working ok for me under vm...
>
> That is good to know.  Ulrich has also written to me that Emacs 23 has
> been working fine for him.

I've been using VM 8.0.x with Emacs 23 since the 23 pretest releases.
I've had problems with the inbox character encoding and with messages
staying labeled 'unread' when they have been read, or not getting
deleted when I delete them. Also, some of the menu entries are
invisible on 23 in X, as I've reported in the bug tracker. It's buggy,
but I haven't lost any data (although to get around the character
encoding problems I've edited my largish inbox by hand, which is not
exactly a good practice).

I'm on Fedora and I've used the emacs-vm packages from the distro. I
just built a RPM of the current bzr revision of 8.1 and upgraded to
that. I'll see how it goes.

Uday S Reddy

unread,
Dec 14, 2009, 6:01:46 PM12/14/09
to
Richard Sharman filed a bug report on Launchpad saying that the
"Display using External Viewer" context menu item is not working for
MIME attachments in Emacs 23. I have tried it with the current
version of VM and I think the problem is there.

Have other people run into this problem as well? Anybody who has
managed to get around it?

Cheers,
Uday

Uday S Reddy

unread,
Dec 15, 2009, 12:32:28 PM12/15/09
to
Uday S Reddy wrote:

> Richard Sharman filed a bug report on Launchpad saying that the
> "Display using External Viewer" context menu item is not working for
> MIME attachments in Emacs 23. I have tried it with the current
> version of VM and I think the problem is there.

Sorry, false alarm. The problem is *not* present in the current devo
version.

(I had some customization code that was getting fired and behaving
like the old code.)

But, more generally, the problem here was caused by a call to
set-buffer-multibyte (which converts a unibyte buffer to multibyte or
vice versa). The use of this function is questionable and dependent
on the low-level character representation used inside Emacs. Doing a
grep showed 13 calls to this function in the current devo version. It
will take us some time to verify that all the 13 calls are all working
correctly. Meanwhile, please be vigilant and fall back to Emacs 22
when necessary.

Cheers,
Uday

0 new messages