reading binary, non-unix file

reading binary, non-unix file

[Mickey Ferguson]

[-- Attachment #1.1: Type: text/plain, Size: 1307 bytes --]

I'm using emacs 21.1.1 on Windows 2000 SP4, and I'm having trouble
reading a file.  The file is a Dr. Watson crash dump.  Apparently it's
got some binary characters in the file, causing emacs to think it's a
Unix file, when it isn't.  When the file loads into Emacs, the mode line
at the bottom says "(Unix)", and it displays a bunch of control
character kinds of stuff.  If I just open the file in ordinary notepad,
it views fine.
 
Question 1:  How do I load this file into emacs so that it's viewable,
just like it is in notepad?  Of course I could open it in notepad, copy
the entire buffer contents, and then paste it into an emacs buffer, but
that's clunky.  I just want to be able to use ordinary find-file, or
start up emacs directly with the file.  Alternatively, if loading it
directly won't work, at least have some kind of function I can use so I
can load it and edit.
 
Question 2:  How do I prevent emacs from ever loading in a file in this
(unix) manner?  I'm never going to be loading any unix-created files on
my system, and just want them loaded just the same as if I had loaded in
notepad.
 
Please reply directly to me at MFerguson (AT) peinc (DOT) com, since I
have no access to any usenet news host and thus can't read any responses
in the newsgroups.  Thanks!

[-- Attachment #1.2: Type: text/html, Size: 2372 bytes --]

[-- Attachment #2: Type: text/plain, Size: 152 bytes --]

_______________________________________________
Help-gnu-emacs mailing list
Help-gnu-emacs@gnu.org
http://lists.gnu.org/mailman/listinfo/help-gnu-emacs

Re: reading binary, non-unix file

[J. David Boyd]

try
 M-x find-file-literally

Dave

Re: reading binary, non-unix file

[Mickey Ferguson]

"J. David Boyd" <dave@adboyd.com> wrote in message
news:mailman.4677.1098473798.2017.help-gnu-emacs@gnu.org...
>
> try
>  M-x find-file-literally

Nope.  It loads the same (bad) way.

(P.S.  I found a free news server - quimby.gnus.org - that hosts
gnu.emacs.help and allows posting, so I can now follow the thread that way.)

Re: reading binary, non-unix file

[Mathias Dahl]

"Mickey Ferguson" <MFerguson@peinc.com> writes:

> "J. David Boyd" <dave@adboyd.com> wrote in message
> news:mailman.4677.1098473798.2017.help-gnu-emacs@gnu.org...
> >
> > try
> >  M-x find-file-literally
> 
> Nope.  It loads the same (bad) way.
> 
> (P.S.  I found a free news server - quimby.gnus.org - that hosts
> gnu.emacs.help and allows posting, so I can now follow the thread that way.)

Can you paste some part of the file here? My guess is that
what you are looking at in emacs is a UTF-16 encoded file.

I am not sure at all, but I know that when I export stuff
from the registry to a file and just drag and drop that into
emacs it displays a lots of "^@" (ASCII 0, Ctrl-0) between
each "real" character. This is because it is UTF-16 encoded.

If I guess correctly you should be able to do this:

C-x <return> c u t f - 1 6 <return> C-x C-f m y _ f i l e . t x t <return>

where my_file.txt is the file you want to look at, and Emacs
will parse the file as it should.

/Mathias

Re: reading binary, non-unix file

[Mickey Ferguson]

We're almost there.  I don't have a single UTF-16 coding choice.  When I
tried what you provided, I saw:

Possible completions are:
utf-16-be                          utf-16-be-dos
utf-16-be-mac                      utf-16-be-unix
utf-16-le                          utf-16-le-dos
utf-16-le-mac                      utf-16-le-unix

I chose utf-16-le and it seemed to do it properly.  I just don't know if
that was the right choice - I don't fully understand what each of these
provides.

Second, after I determine which one of the above to use, can anyone help me
to write a function so that I can then map a key combination (similar to C-X
C-F uses Find-File), that will load in the proper coding and then find the
file?  I'm lisp-impaired, so any help would be appreciated.  I'm capable of
taking an interactive function that's been defined and mapping it to a
keystroke, but that's about it.

"Mathias Dahl" <brakjoller@gmail.com> wrote in message
news:ud5z7xpur.fsf@gmail.com...
> "Mickey Ferguson" <MFerguson@peinc.com> writes:
>
> Can you paste some part of the file here? My guess is that
> what you are looking at in emacs is a UTF-16 encoded file.
>
> I am not sure at all, but I know that when I export stuff
> from the registry to a file and just drag and drop that into
> emacs it displays a lots of "^@" (ASCII 0, Ctrl-0) between
> each "real" character. This is because it is UTF-16 encoded.
>
> If I guess correctly you should be able to do this:
>
> C-x <return> c u t f - 1 6 <return> C-x C-f m y _ f i l e . t x t <return>
>
> where my_file.txt is the file you want to look at, and Emacs
> will parse the file as it should.
>
> /Mathias

Re: reading binary, non-unix file

[Daniel Pittman]

On 26 Oct 2004, Mickey Ferguson wrote:
> We're almost there. I don't have a single UTF-16 coding choice. When I
> tried what you provided, I saw:
>
> Possible completions are:
> utf-16-be                          utf-16-be-dos
> utf-16-be-mac                      utf-16-be-unix
> utf-16-le                          utf-16-le-dos
> utf-16-le-mac                      utf-16-le-unix
>
> I chose utf-16-le and it seemed to do it properly. I just don't know if
> that was the right choice - I don't fully understand what each of these
> provides.

They can be read as three section:  'utf-16', which is the 16 bit
version of Unicode, 'be' or 'le', which stand for big-endian or
little-endian, and 'mac', 'dos', or 'unix', which indicate line ending
conversion.

UTF-16 is the only coding system to posses the be/le split, and that is
because Microsoft unilaterally declared that they would be implementing
UTF-16-le in their OS, regardless of what the IETF and Unicode people
decided as an endian encoding.

The 'line ending' stuff is, basically, what line endings to expect/use
when you hit return.  -unix is LF, -mac is CR, and -dos is CRLF, through
quirks of historic accident, mostly.

> Second, after I determine which one of the above to use, can anyone help me
> to write a function so that I can then map a key combination (similar to C-X
> C-F uses Find-File), that will load in the proper coding and then find the
> file?  I'm lisp-impaired, so any help would be appreciated.  I'm capable of
> taking an interactive function that's been defined and mapping it to a
> keystroke, but that's about it.

    C-x RET c <coding system> <whatever command>

C-x RET c runs `universal-coding-system-argument', which allows you to
specify the encoding for the next command.

Alternately, you can from `file-coding-system-alist', which maps regular
expressions to coding systems automatically.

You can use that to specify that whatever filename you want is loaded as
whatever coding system you want.

Regards,
        Daniel
-- 
Estne Tibi Forte Magna Feles Fulva Et Planissima?

Re: reading binary, non-unix file

[Mathias Dahl]

"Mickey Ferguson" <MFerguson@peinc.com> writes:

> We're almost there.  I don't have a single UTF-16 coding choice.
> When I tried what you provided, I saw:
 
> Possible completions are:
> utf-16-be                          utf-16-be-dos
> utf-16-be-mac                      utf-16-be-unix
> utf-16-le                          utf-16-le-dos
> utf-16-le-mac                      utf-16-le-unix
 
> I chose utf-16-le and it seemed to do it properly.  I just don't
> know if that was the right choice - I don't fully understand what
> each of these provides.

I wont comment on -le and -be but I have noticed that "utf-16" also
works (at least I have it in my CVS Emacs). utf-16-le is what
Microsoft uses to encode Unicode data in files so that is the one you
want to use.

> Second, after I determine which one of the above to use, can anyone
> help me to write a function so that I can then map a key combination
> (similar to C-X C-F uses Find-File), that will load in the proper
> coding and then find the file?  I'm lisp-impaired, so any help would
> be appreciated.  I'm capable of taking an interactive function
> that's been defined and mapping it to a keystroke, but that's about
> it.

I had the same problem and instead of searching the net or asking here
I just created a keyboard macro which I then named, put in my .emacs
and which I then bound a key to.

You start and stop a keyboard macro with C-x ( and C-x )
respectively. You can then name it with M-x name-last-kbd-macro <name>
and then insert it so that it acts as a function (sort of) in your
.emacs with M-x insert-kbd-macro <name>.

Hope this helps!

/Mathias

Re: reading binary, non-unix file

[Mickey Ferguson]

"Mathias Dahl" <brakjoller@gmail.com> wrote in message
news:uis8x6e8i.fsf@gmail.com...
> "Mickey Ferguson" <MFerguson@peinc.com> writes:
> > Second, after I determine which one of the above to use, can anyone
> > help me to write a function so that I can then map a key combination
> > (similar to C-X C-F uses Find-File), that will load in the proper
> > coding and then find the file?  I'm lisp-impaired, so any help would
> > be appreciated.  I'm capable of taking an interactive function
> > that's been defined and mapping it to a keystroke, but that's about
> > it.
>
> You start and stop a keyboard macro with C-x ( and C-x )
> respectively. You can then name it with M-x name-last-kbd-macro <name>
> and then insert it so that it acts as a function (sort of) in your
> .emacs with M-x insert-kbd-macro <name>.

One thing I omitted in my requirements specification was that I cannot just
bind this utf-16-le coding to a particular file type (e.g. *.log).  I need
to be able to do it on a file-by-file basis.

Given the above information, I tried your suggestion above, and I've
definitely got something wrong.  I tried the following, for which the fset
is the result of using the C-x ( and C-x ):

;;; define function to load in a unicode file (utf-16-le encoding)
(fset 'find-unicode-file
   [?\C-x return ?c ?u ?t ?f ?- ?1 ?6 ?- ?l ?e])

This results in a very large minibuffer error message, which is pretty much
incomprehensible to me.

OK, I tried a different tack.  I tried this function definition, taking a
completely wild stab in the dark, based on what one of the other helpful
responders suggested for the underlying functions.  It's probably completely
messed up, but I bet one of the lisp gurus out there can figure out how to
fix it.  Basically, what I want this function to do is to perform the
equivalent of having performed a C-x <enter> c, specifying the utf-16-le
encoding, followed by a regular find-file operation.  In general, I like
this approach much better than the previous approach, but I also prefer to
use something that actually works, instead of failing :-), so whatever we
can get to solve the problem is good for me!

Here's the (completely wrong) code:

;;; define function to load in a unicode file (utf-16-le encoding)
(defun find-unicode-file (filename)
  "Run find-file on a unicode (utf-16-le encoding) file."
  (interactive "FFind unicode file: \np")
  (universal-coding-system-argument "utf-16-le")
  (find-file filename)
  )

Re: reading binary, non-unix file

[Mathias Dahl]

"Mickey Ferguson" <MFerguson@peinc.com> writes:

> > You start and stop a keyboard macro with C-x ( and C-x )
> > respectively. You can then name it with M-x name-last-kbd-macro <name>
> > and then insert it so that it acts as a function (sort of) in your
> > .emacs with M-x insert-kbd-macro <name>.
> 
> ...
> Given the above information, I tried your suggestion above, and I've
> definitely got something wrong.  I tried the following, for which the fset
> is the result of using the C-x ( and C-x ):
> 
> ;;; define function to load in a unicode file (utf-16-le encoding)
> (fset 'find-unicode-file
>    [?\C-x return ?c ?u ?t ?f ?- ?1 ?6 ?- ?l ?e])

You are right, that did not work. When I used a keyboard for doing
this a while back it was for *saving* files. It seems that it does not
work with C-x C-f, leaving out the file name.

I looked at how universal-coding-system-argument works and hacked
together the following, which seems to do what you want:

(defun find-unicode-file ()
  "Run find-file on a unicode (utf-16-le encoding) file."
  (interactive)
  (let* ((coding-system 'utf-16-le)
         (coding-system-for-read coding-system)
         (coding-system-require-warning t))
    (call-interactively 'find-file)))

Hope it works for you.

/Mathias

Re: reading binary, non-unix file

[Mickey Ferguson]

Indeed, it does work.  Thanks, Mathias!

Can someone explain the operation of the 'let*' operation?  My lisp
understanding is about pre-school level, maybe not even that.  I've seen the
'let' operation, but never a 'let*' one.  And it's not in my (copyright
1981! - from college days) LISP book.

"Mathias Dahl" <brakjoller@gmail.com> wrote in message
news:u654w61tf.fsf@gmail.com...
> "Mickey Ferguson" <MFerguson@peinc.com> writes:
>
> I looked at how universal-coding-system-argument works and hacked
> together the following, which seems to do what you want:
>
> (defun find-unicode-file ()
>   "Run find-file on a unicode (utf-16-le encoding) file."
>   (interactive)
>   (let* ((coding-system 'utf-16-le)
>          (coding-system-for-read coding-system)
>          (coding-system-require-warning t))
>     (call-interactively 'find-file)))

Re: reading binary, non-unix file

[Reiner Steib]

On Wed, Oct 27 2004, Mickey Ferguson wrote:

> Can someone explain the operation of the 'let*' operation?  

(info "(emacs-lisp-intro)fwd-para let")
(info "(elisp)Local Variables")

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

RE: reading binary, non-unix file

[Drew Adams]

Like "let", but the bindings are sequential and interdependent (not parallel
and independent). You can use variables bound previously in subsequent
variable bindings of the same "let*".

E.g.

(let* ((foo 2)          ; define foo
       (bar (* 3 foo))) ; use foo to define bar
  (format "foo: %s, bar: %s" foo bar))

=>

foo: 2, bar: 6

-----Original Message-----
Can someone explain the operation of the 'let*' operation?

Re: reading binary, non-unix file

[Mickey Ferguson]

We're almost there.  I don't have a single UTF-16 coding choice.  When I
tried what you provided, I saw:

Possible completions are:
utf-16-be                          utf-16-be-dos
utf-16-be-mac                      utf-16-be-unix
utf-16-le                          utf-16-le-dos
utf-16-le-mac                      utf-16-le-unix

I chose utf-16-le and it seemed to do it properly.  I just don't know if
that was the right choice - I don't fully understand what each of these
provides.

Second, after I determine which one of the above to use, can anyone help me
to write a function so that I can then map a key combination (similar to C-X
C-F uses Find-File), that will load in the proper coding and then find the
file?  I'm lisp-impaired, so any help would be appreciated.  I'm capable of
taking an interactive function that's been defined and mapping it to a
keystroke, but that's about it.

"Mathias Dahl" <brakjoller@gmail.com> wrote in message
news:ud5z7xpur.fsf@gmail.com...
> "Mickey Ferguson" <MFerguson@peinc.com> writes:
>
> Can you paste some part of the file here? My guess is that
> what you are looking at in emacs is a UTF-16 encoded file.
>
> I am not sure at all, but I know that when I export stuff
> from the registry to a file and just drag and drop that into
> emacs it displays a lots of "^@" (ASCII 0, Ctrl-0) between
> each "real" character. This is because it is UTF-16 encoded.
>
> If I guess correctly you should be able to do this:
>
> C-x <return> c u t f - 1 6 <return> C-x C-f m y _ f i l e . t x t <return>
>
> where my_file.txt is the file you want to look at, and Emacs
> will parse the file as it should.

Re: reading binary, non-unix file

[Kevin Rodgers]

[Please don't top-post.]

Mickey Ferguson wrote:
 > We're almost there.  I don't have a single UTF-16 coding choice.  When I
 > tried what you provided, I saw:
 >
 > Possible completions are:
 > utf-16-be                          utf-16-be-dos
 > utf-16-be-mac                      utf-16-be-unix
 > utf-16-le                          utf-16-le-dos
 > utf-16-le-mac                      utf-16-le-unix
 >
 > I chose utf-16-le and it seemed to do it properly.  I just don't know if
 > that was the right choice - I don't fully understand what each of these
 > provides.

LE = Little Endian and BE = Big Endian.  Emacs should provide an
encoding in which the endianness is determined from the BOM, but since
it doesn't, utf-16-le is right for x86 processors.  See
http://en.wikipedia.org/wiki/UTF-16

 > Second, after I determine which one of the above to use, can anyone 
help me
 > to write a function so that I can then map a key combination (similar 
to C-X
 > C-F uses Find-File), that will load in the proper coding and then 
find the
 > file?  I'm lisp-impaired, so any help would be appreciated.  I'm 
capable of
 > taking an interactive function that's been defined and mapping it to a
 > keystroke, but that's about it.

What's wrong with `C-x RET c utf-16-le RET C-x C-f' :-)

If all the files are UTF-16 on your system, perhaps you can just put
this in your .emacs file:

(setq-default buffer-file-coding-system 'utf-16-le)

-- 
Kevin Rodgers

Re: reading binary, non-unix file

[Daniel Pittman]

On 23 Oct 2004, Mickey Ferguson wrote:
> I'm using emacs 21.1.1 on Windows 2000 SP4, and I'm having trouble reading a
> file.  The file is a Dr. Watson crash dump.  Apparently it's got some binary
> characters in the file, causing emacs to think it's a Unix file, when it
> isn't.  When the file loads into Emacs, the mode line at the bottom says
> "(Unix)", and it displays a bunch of control character kinds of stuff.  If I
> just open the file in ordinary notepad, it views fine.
>
> Question 1: How do I load this file into emacs so that it's viewable, just
> like it is in notepad?  Of course I could open it in notepad, copy the
> entire buffer contents, and then paste it into an emacs buffer, but that's
> clunky.  I just want to be able to use ordinary find-file, or start up emacs
> directly with the file.  Alternatively, if loading it directly won't work,
> at least have some kind of function I can use so I can load it and
> edit.

The issue is one of character set detection, which you need to override
in this case.

You can do this by using the 'universal coding system' prefix to the
find-file command, by entering the following sequence:

   C-x RET c <coding system> RET C-x C-f <file>

Basically, C-x RET c runs `universal-coding-system-argument', which
prompts for a coding system, then uses that for the `find-file'
operation.

Picking something like latin-1-dos should give you what you want there,
by forcing the end-of-line convention as well, rather than relying on
the automatic detection.

> Question 2: How do I prevent emacs from ever loading in a file in this
> (unix) manner?  I'm never going to be loading any unix-created files on my
> system, and just want them loaded just the same as if I had loaded in
> notepad.

Modify the variable `file-coding-system-alist' to use the -dos variants,
rather than the generic (autodetecting) variants of the coding systems.

Regards,
        Daniel
-- 
The aim of psychoanalysis is to relieve people of their neurotic unhappiness
so that they can be normally unhappy.
        -- Sigmund Freud