Re: funny characters

["Robert G. Brown" <bob@xxxxxxx>]

>hi all stateside
>
>anybody got an idea why:-
>
>i open under zindoze   notepad
>
>
>eh voila   there is  three little charcters  black filled squares
>i can cut and copy   etc but if i delete them  save and reopen the file
>they're back
>so i look at it with  xedit under linux
>it shows the follwing    -----    ^M
>delte it it, then save and reopen    voila    gone!!!
>   what is it ????? 
>sorry if this question is a bit thick
>but 7 yrs of windows  has made a blind man of me when it comes
>to this sort of thing    thanx   
>oh yeh    "we won the cup"

DOS-based text files are defined (by convention) to have records, or rows, 
seperated by the 2 character ASCII sequence CR-LF (or Carriage-Return,
Linefeed). UNIX and UNIX-like systems (or which Linux is one), seperate
records with only the Linefeed character. 

If you move a DOS file to a UNIX system (without doing some sort of 
conversion on it), the "extra" character (the CR) is often represented 
in UNIX editors by the characters ^M, or Control-M, another way of pro-
ducing a CR. Go to a Linux console or an xterm shell, hold down the 
control (or Ctrl) key, and press m, and you'll see it produces a car-
riage return. 

On moving a file from a UNIXX to a DOS-based system, there are no
records! That is, since the CR-LF sequence appears nowhere, DOS 
applications have no way to figure out where records start. Some
graphical editors in these systems will take all the non-printable
characters (LF is one), and place a black bullet character (I don't
know the value offhand) to show that SOMETHING is there, but it can't
be displayed.

The solution is to run conversion programs in each direction, in going
DOS->UNIX, you need a program that removes that CR, and while going 
UNIX->DOS, you need one that puts it back. explicitly these programs
exist (I've written a pair, but I do weird things with files!), but
most of the time you don't need them. When you use an ftp client program
to move data, it will perform this removal or insertion when you're in
ASCII mode, and you can always run the ftp client/server components on
 your own machine, even when not connected to a network (ftp localhost).

A final point: there are some packages in the DOS/Windows environment
(one that comes to mind is Wordperfect) that does not follow the DOS 
conventions in a strict sense, as it recognizes LF-only files as having 
distinct records. It's possible to load a "DOS native" file that came
from a UNIX system into Wordperfect, then save it as a "DOS native file",
and have the CR's inserted. I don't know if this behavior holds of Unix
releases of WP, but I suspect it does (others, like Adam, can comment on 
this).  I also expect that WP is not unique in this regard.

                                                        ---> RGB <---