mule : multibyte to 7 bit conversions

mule : multibyte to 7 bit conversions

[nospam55]

Hi

since I passed to emacs version 21.2.1 on Redhat 8  I noticed that in Man mode 
(or at least after M-x man RET bash RET), looking for hyphens with C-s didn't work; 
same problem even  when saving that man page on a file.

After some attempt I saw that with 

    (toggle-enable-multibyte-characters nil)

an innocent  -c  became visible as \234\364\370\262c  , and so on ; 
Mysterious utf8 proposals arose from attempts to save the manpage...

How can I bring this multibyte manpage into ordinary 7 bit ascii form ?
How to have manpages under emacs 7 bit by default ?

   Thank you :-)

Re: mule : multibyte to 7 bit conversions

[Kevin Rodgers]

nospam55 wrote:

> Hi
> 
> since I passed to emacs version 21.2.1 on Redhat 8  I noticed that in Man mode 
> (or at least after M-x man RET bash RET), looking for hyphens with C-s didn't work; 
> same problem even  when saving that man page on a file.
> 
> After some attempt I saw that with 
> 
>     (toggle-enable-multibyte-characters nil)
> 
> an innocent  -c  became visible as \234\364\370\262c  , and so on ; 
> Mysterious utf8 proposals arose from attempts to save the manpage...
> 
> How can I bring this multibyte manpage into ordinary 7 bit ascii form ?
> How to have manpages under emacs 7 bit by default ?

There was a discussion about this on the bug-gnu-emacs or emacs-devel list,
but I can't find it now to give you a reference.  You might try wrapping
LC_ALL=C around the call to the Unix man command, or (setenv "LC_ALL" "C")
around the call to the Emacs man command.

-- 
Kevin Rodgers

Re: mule : multibyte to 7 bit conversions

[nospam55]

I live already under 

  export LC_COLLATE=C

your suggestion is important, it suggests me that 
the problem's origin is not emacs-related;
in fact if I say from bash

 $ man bash > read.tmp

then read.tmp is hardly readable because of plenty of embedded ctrl chars such as ^H;

I will  try  man man

Re: mule : multibyte to 7 bit conversions

[nospam55]

Unfortunately man man didn't help. 
I will post a non-emacs-related query. :-(

Re: mule : multibyte to 7 bit conversions

[Kevin Rodgers]

nospam55 wrote:

> I live already under 
> 
>   export LC_COLLATE=C
> 
> your suggestion is important, it suggests me that 
> the problem's origin is not emacs-related;


Correct.

> in fact if I say from bash
> 
>  $ man bash > read.tmp
> 
> then read.tmp is hardly readable because of plenty of embedded ctrl chars such as ^H;


That is a separate problem, which Emacs handles in the Man-fontify-manpage
function (C-h v).  You can even invoke it directly via M-x.

What happens if you run this command:

LC_ALL=C man bash > bash.man

Are there still non-ASCII characters in the bash.man file?


> I will  try  man man

The `man foo` command is basically equivalent to:


nroff -man /usr/man/man?/foo.? | ${PAGER:more}


It works because your terminal interprets the backspace and various escape
sequences that nroff outputs.

-- 
Kevin Rodgers

Re: mule : multibyte to 7 bit conversions

[nospam55]

> What happens if you run this command:
> 
> LC_ALL=C man bash > bash.man
> 
> Are there still non-ASCII characters in the bash.man file?

unfortunately with 


  $ LC_ALL=C man bash > bash.man

I get in bash.man the same funny ^H etc chars as before; moreover, I have from many weeks 

  $ echo  ${!LC*}
  LC_COLLATE
  $ declare -p LC_COLLATE
  declare -x LC_COLLATE="C"



> The `man foo` command is basically equivalent to:
> 
> 
> nroff -man /usr/man/man?/foo.? | ${PAGER:more}
> 
> 
> It works because your terminal interprets the backspace and various escape
> sequences that nroff outputs.`

This suggestion seems to be important, I will try using it and will let you know 

    nospam55

Re: mule : multibyte to 7 bit conversions

[nospam55]

After struggling around man nroff etc , I saw that passing in /etc/man.conf from 
  NROFF         /usr/bin/nroff -c -mandoc
to
  NROFF         /usr/bin/nroff -c -mandoc -Tascii
made emacs doing manpages in apparent ascii encoding, :-)

Unfortunately however , from the shell, with 

  $ man bash > bash.tmp

I get pages wich now are in ascii (no funny utf8 stuff) , but still a plenty of ctrl chars
to control a terminal; how could I fix this and get only-printable ascii encoding 
from the shell  ?