Cfxtrjtrk

1. I created a text file by copying its different parts from different
   sources (webpages, other text files, pdf files) into gedit and
   saving it to the file. I guess that is the reason that I have
   multiple encodings in the text file, but I am not sure. How can I
   avoid creating a text file with mixed encodings by copying its
   different parts from different sources into gedit?


2. 
   

   Whenever I open the file in gedit, gedit can always show or decode
   every part of the text correctly. It seems that gedit can handle a
   text file with mixed encodings, but I am not sure.

   
   
   

   But when I open the file in emacs, there will be characters that
   can't be shown correctly. (I am not sure why emacs can't do that.)
   So I would like to convert the file from mixed encodings to a single
   encoding such as utf-8.

   
   
   

   Since I think gedit can detect the correct encodings for different parts of the text file, and I don't know if there are other applications that can do so, would it be possible to ask gedit to convert the file to
   utf-8, or at least tell me what encoding it finds for which part of the file?

   
   

Thanks.

Hmmm... the concept of a file with various encoding is somewhat wobbly, to be honest. If you have a bit of time, this article (and this one) are worth reading.

For Linux a file is a sequence of bytes. If you ask a program to interpret it as a text file, it will do it using a mapping between bytes and characters; this mapping is the encoding. Almost all the text editor I know (not word processors!) just understand the concept of one encoding for one file.

I am not expert on gedit; maybe it is doing some magic like trying to autodetect the encoding line by line or text block by text block... if it is the case you can try to do the same using enca(1):

 while read line; do echo $line | enconv -L none -x utf8; done < text.mixed > text.utf8

...but it depends on how good is enca in guessing you encoding (works almost well with Eastern European, but not with Latin1, for example).

(1) It's in the repos, just install it with sudo apt-get enca.

I had the same problem and solved it with Emacs. The solution is quoted from here:

Another possible solution is to mark each region appearing with Chinese characters and recode it with M-x recode-region, giving "Text was really in" as utf-16-le and "But was interpreted as" as utf-16-be.

Another one is to split the the parts which have different encodings, paste them into different files, convert the encoding of the one and add it to the other. In my case this worked with Atom, but not with Notepad++ (utf16-le/be).