Emacs smart umlaut conversion for TeX



Smart umlaut conversion for TeX.

License GPLv3


Transform LaTeX encoded non-ASCII characters to and from their visible (utf-8) representations when visiting a file and preserve their original encoding when saving the buffer.

TeX/LaTeX documents often contain special characters that are not available in ASCII code in a special command form. For example, the German umlaut ä can be written in LaTeX as {"a}, "{a}, "a or even "a. Of course, nowadays one should use an encoding like utf-8, which contain a natural representation of these characters. Unfortunately, old documents may still use the old encodings or non-ASCII encodings are not allowed for other reasons. Emacs can already automatically encode and decode such characters from and to their LaTeX representation using the iso-cvt package: during editing the document the characters are utf-8 encoded but when saved to disk the characters are transformed to a LaTeX representation.

Unfortunately, this an all-or-nothing approach: either all characters are transformed or none. But when working on a document with several authors this may be problematic. This may particularly true if the document is stored in a revision control system. In these cases each author may use its own editor, each editor may have its own setting and encodes umlauts in a special preferred way. Each time one author does a small change, all umlauts in the whole document may get transformed. This may also lead to huge diffs that consist of many hunks that only change the encoding of some characters.

The purpose of tex-smart-umlauts is to automatically encode and decode umlauts from and to their tex representation, while preserving the original encoding. In other words, when a LaTeX file is visited, the original encoding of each character is saved and the character is transformed to its visible (utf-8) representation. When the document is saved again, each character that has been present when the document has been loaded is saved in its original encoding. Only newly inserted non-ASCII characters get a new encoding that depends on the user-options of tex-smart-umlauts. This way, a small change in a document will not reencode all non-ASCII charactes as iso-cvt would do and only the modified parts of the document will really be modified on disk.


Simply add

(add-hook 'LaTeX-mode-hook #'tex-smart-umlauts-mode)

to your .emacs file. This will convert all LaTeX-encoded characters to their resp. visible encoding and store their original encodings. This will also automatically register the encoding so that the original encodings are restored when the buffer is saved.

Change Log


  • Minor changes.


  • Use a true minor-mode.
  • Handle reverting a buffer.


  • Add encoding for ß.


  • Convert string is iso-cvt table to multibyte
  • Use inhibit-read-only instead of buffer-read-only

Function Documentation

(tex-smart-umlauts-reencode-all &optional FROM TO)

Reencode all charactes in region. Only characters between the buffer positions FROM and TO are decoded. If FROM and TO are nil the whole buffer is encoded. The original encodings of all characters in the region is dropped and replaced by a new encoding according to the rules of tex-smart-umlauts-encodings and tex-smart-umlauts-encode.


Show encodings of all umlauts in buffer.


Hide encodings of all umlauts in buffer.