Understanding Stata's Comment Hierarchy

[Mauricio Caceres]

I was hoping to get feedback in understanding Stata comments. The documentation was helpful but I wonder if I'm missing anything:

- Multi-line comments can be nested. This means that every "/*" within a multi-line comment must be closed.

- Stata ignores everything within "/*" and "*/" across lines unless the first "/*" appears in a line that starts with "//" or after " //".

- Stata ignores everything after in-line comments: "//" at the start of a line or after " //" at the end of a line (given they is not preceded by "/*").

- Stata ignores code after "*" at the start of a line or "*" directly after a multi-line comment. However, line continuations (" ///"), multi-line comments ("/*" to "*/"), and in-line comment rules (" //") apply after a "*".

If you run the following:

* /* This will be a multi-line comment
disp "Not printed"
*/

* // /* Ignored due to inline comment
disp "Printed 1"

// /* Also ignored due to inline comment
disp "Printed 2"

*// /* This is not an inline comment, so this is multi-line again
disp "Not printed"
*/

* ///
disp "Not printed. Line continuation applies"

// /// Line continuation ignored due to inline comment
disp "Printed 3"

/*
/* Nested */
disp "Not printed"
*/* disp "Not printed"

(Note that if "*/*" was replaced with "*///" we would get an error). If I am missing something in terms of comment hierarchy, I would be really thankful for input. Cheers!

[Sergiy Radyakin]

Mauricio, thing get even more fun once you bring in the delimiter ;

See this thread:
https://www.stata.com/statalist/arch.../msg00040.html

Best, Sergiy Radyakin

[Mauricio Caceres]

Indeed. "#delim ;" makes this quite interesting. But I think I figured it out:

1. "//" means ignore everything until the next line break (carriage return).

2. "///" means ignore everything until the next carriage return, including the line break (carriage return).

3. "*" means ignore all code (but, crucially, not all comments) until the next delimiter, which is carriage return or ";"

#delimit ;

// This line is ignored, but the line break is not
di "Printed 1";

// Same for multi-line /*
di "Printed 2";

* Line continuations do apply
di "Not printed";

* Same for multi-line /*
di "Hello one"  */;

Moreover:

#delimit cr

* Line continuation ///
// Breaks line continuation ///
di "Printed 1"

disp "Line continuation" ///
// Breaks line continuation ///
di "Printed 2"

#delimit ;

* Line continuation
// Does not break line continuation
di "Not printed";

disp "Line start"  
 // This is ignored, and does not give error
"; Line end" ;

disp "Line start"  
// This does give an error
"; Line end" ;

The last one is annoying because there is a trailing space, but it seems Stata strips trailing spaces when joining lines. These also work:

local ninety 90
disp `ni/*9*/nety'
disp `ni/*

*/nety'
disp `ni ///
nety'

The last one displays nothing, however, since it's effectively calling `ni nety'.

[Kyle Barron]

FYI, this is now fixed in the syntax highlighter for Atom (except for the corner cases with #delimit ; )