want to go to the 2345th byte, not character

want to go to the 2345th byte, not character

[Dan Jacobson]

I'm sitting in a unicode buffer and I want to go to the 2345th byte.
ESC 2 3 4 5 C-f will take me to the 2345th character, wide and narrow
included, but not the 2345th byte.  Therefore there should be an
additional command that operates byte wise.  No I don't want to leave
unicode mode.

So a C-f analog that counts in bytes would be good.
"byte-forward-char" sounds good but must be an unrelated variable.

P.S.
   C-f runs the command forward-char
   Move point right N characters (left if N is negative).

Perhaps that right and left stuff should be reworded in view of BIDI
(Arabic, Hebrew) languages? Just say forward and backward?

Re: want to go to the 2345th byte, not character

[Eli Zaretskii]

> From: Dan Jacobson <jidanni@jidanni.org>
> Date: Thu, 23 Jun 2005 01:20:07 +0800
> 
> I'm sitting in a unicode buffer and I want to go to the 2345th byte.

There are no original bytes in the Emacs buffer.  Emacs decodes the
file when it reads it; the decoding process removes the original byte
sequences that stood for the encoded characters in the file, and in
their stead inserts the characters in the Emacs internal
representation, which is very different from what's in the file.

> ESC 2 3 4 5 C-f will take me to the 2345th character, wide and narrow
> included, but not the 2345th byte.  Therefore there should be an
> additional command that operates byte wise.

Can you tell why you need that?

Anyway, if you want to treat the file as a stream of bytes, you need a
byte editing mode; use hexl-find-file for that.

>    C-f runs the command forward-char
>    Move point right N characters (left if N is negative).
> 
> Perhaps that right and left stuff should be reworded in view of BIDI
> (Arabic, Hebrew) languages?

You assume that C-f in a right-to-left language will still move in the
logical (as opposed to visual) direction.  That is not a given; when
we implement bidi support for Emacs, we will have to figure out what
is the best behavior here.  Until then, there's no need to use
abstractions where none are required.

Re: want to go to the 2345th byte, not character

[Dan Jacobson]

Eli> Can you tell why you need that?

iconv said there was a problem at the 2345th character in that file
and I wanted to see what the file looked like around there.

Eli> use hexl-find-file for that.

But then I can't see it in the native language I bet.

OK, next time I'll try dd(1).

Re: want to go to the 2345th byte, not character

[Kevin Rodgers]

Dan Jacobson wrote:
 > Eli> Can you tell why you need that?
 >
 > iconv said there was a problem at the 2345th character in that file
 > and I wanted to see what the file looked like around there.

C-x RET c raw-text RET C-x C-f my_file.txt RET

 > Eli> use hexl-find-file for that.
 >
 > But then I can't see it in the native language I bet.

Hexl mode displays each byte's corresponding character as well as its
2-digit hexadecimal value.

 > OK, next time I'll try dd(1).

Yuck.

-- 
Kevin Rodgers

Re: want to go to the 2345th byte, not character

[Eli Zaretskii]

> From: Dan Jacobson <jidanni@jidanni.org>
> Date: Thu, 23 Jun 2005 23:56:38 +0800
> 
> Eli> Can you tell why you need that?
> 
> iconv said there was a problem at the 2345th character in that file
> and I wanted to see what the file looked like around there.

I hope you filed a bug report with the iconv maintainers, asking them
to report the character number as well as the byte number.  I don't
see why Emacs should fix the entire world out there.

> Eli> use hexl-find-file for that.
> 
> But then I can't see it in the native language I bet.

If hexl doesn't show decoded non-ASCII characters alongside the hex
byte values (I don't remember whether it should), then perhaps you
should request a feature for it to do so.

hexl vs. non-ASCII

[Dan Jacobson]

Eli> If hexl doesn't show decoded non-ASCII characters alongside the
Eli> hex byte values (I don't remember whether it should), then
Eli> perhaps you should request a feature for it to do so.

OK, hexl is apparently pure ASCII when it could also show wide
characters, so I hereby request a feature for it to do so, if you
insist :-)

Maybe it could show all three side by side, 4 bit, 8 bit, 16 bit &
more bit.  Morbid.