Prettifying LaTeX Buffers

Contents

• prettify-symbols-mode
  • Going further
• Going too far: code folding
• Conclusion

A friend recently confided in me that, after years of using Emacs, he was only now getting into using prettify-symbols-mode for editing LaTeX buffers! After overwhelming him with more customisations related to this, I realised that the topic—while seemingly easy on the surface—actually has a lot more depth to it than one would first imagine.

prettify-symbols-mode§

If you’ve never come into contact with prettify-symbols-mode before, I encourage you to just issue an M-x prettify-symbols-mode RET in a LaTeX-mode buffer. The default prettify-symbols-alist that e.g. AUCTeX provides already includes a lot of symbols. For the uninitiated, it would display a buffer like this⊕

as the already much more readable

Basically, prettify-symbols-mode is a home-cooked ligature system for Emacs, because of course that exists.

However, the above already has some problems: \otimes is quite small, and thus barely discernible from \bullet, and of course user-defined macros—like \defeq or \cat—don’t get picked up on at all. We can refine the displayed symbols by adding our own definitions to the (buffer-local) prettify-symbols-alist variable:

Executing that hook makes the above situation even prettier.

Going further§

Naturally—this being Emacs—substituting one symbol for another is not all that prettify-symbols-mode has to offer. Digging into the documentation of prettify-symbols-alist, the second entry of the cons pair can be⊕

[…] a character, a string or a vector or list of integers and rules.

If it is a character, it is an alternate character to display instead of the text in the region.

If it is a string, the elements are alternate characters. In this case, TAB element has a special meaning. If the first character is TAB, the glyphs are displayed with left padding space so that no pixel overlaps with the previous column. If the last character is TAB, the glyphs are displayed with right padding space so that no pixel overlaps with the following column.

If it is a vector or list, it is a sequence of alternate characters and composition rules, where (2N)th elements are characters and (2N+1)th elements are composition rules to specify how to compose (2N+2)th elements with previously composed N glyphs.

A composition rule is a cons of global and new glyph reference point symbols. See the documentation of reference-point-alist for more details.

We’ve only used a plain character as the replacement so far, but in fact we can supply a whole list of them, with some rules on how everything should be glued together. The gluing of each pair of glyphs is done according to the following diagram, taken from the documentation of reference-point-alist:

A gluing instruction either comprises a cons pair of two locations—e.g., (cr . cl) to glue position 9 of the left glyph to position 11 of the right—or a list with some additional x and y offsets, like '(cr cl 0 10).

As an easy example, consider \defeq from the above LaTeX snippet, which is a better-looking version of the definitional equality symbol :=. Sadly, the appropriate Unicode symbol ≔ does not look particularly nice in my font—simply fusing : and = together works quite well, however! Defining

yields a decent approximation of what I would want the Unicode symbol to look like. Here are some more examples of this in action:

Going too far: code folding§

⊕

The only thing that prettify-symbols-mode lacks is being able to supply a function as a replacement, in order to manipulate the matched symbol.

Enter code folding, which does not have this restriction. Most often, it is used to hide parts of a definition or section, in order to make the respective buffer more easily scannable for a human. However, AUCTeX’s machinery is set up in such a way that one can supply an arbitrary function that transforms the matched text somehow. This is extremely convenient for things like mathematical alphabets, as then one does not have to hard-code all replacements.

The main variables to change here are TeX-fold-math-spec-list and TeX-fold-macro-spec-list. The parsing of tex-fold.el is fairly featureful; the documentation for the latter variable tells the whole story.

List of replacement specifiers and macros to fold.

The first element of each item can be a string, an integer or a function symbol. The second element is a list of macros to fold without the leading backslash. If the first element is a string, it will be used as a display replacement for the whole macro. Numbers in braces, brackets, parens or angle brackets will be replaced by the respective macro argument. For example “{1}” will be replaced by the first mandatory argument of the macro. One can also define alternatives within the specifier which are used if an argument is not found. Alternatives are separated by “||”. They are most useful with optional arguments. As an example, the default specifier for

is “[1]:||*” which means that if there is an optional argument, its value is shown followed by a colon. If there is no optional argument, only an asterisk is used as the display string.

If the first element is an integer, the macro will be replaced by the respective macro argument.

If the first element is a function symbol, the function will be called with all mandatory arguments of the macro and the result of the function call will be used as a replacement for the macro.

Setting this variable does not take effect immediately. Use Customize or reset the mode.

Quite the functionality! The simplest example is probably hiding the \emph macro, which indicates emphasised text.⊕

This ends up looking like the following:

As with prettify-symbols-mode, things unprettify when the point is on them.

Using these techniques, we can define replacements for calligraphic letters (or any other alphabet, like Fraktur or blackboard bold) in a much more straightforward fashion. First, we create a function to shift each codepoint in a string by a certain offset.

Now, all that’s left to do is to define some folds for mathematical alphabets, and give them the right starting points:⊕

Note that this again has the advantage of making user-defined macros much easier to integrate into the whole system. Above we have met the \cat macro, which should be displayed in much the same way as \mathcal is. With code folding, all this entails is to amend the respective line in TeX-fold-math-spec-list:

Conclusion§

Using everything in this article together—see here—one can transform a boring old LaTeX buffer

into something like this:

Nice.