dust_dds_gen: Cannot have trailing whitespace or comments after #endif

[Notgnoshi]

In C/C++, I habitually leave trailing comments on #endifs, namespaces, and nested classes.

It's not strictly necessary to do so in IDL, because it'd be quite strange to have nested preprocessor statements, but I thought I'd report that dust_dds_gen can't parse the following:

#ifndef FOO
#define FOO

// ...

#endif // FOO

This fails with:

thread 'main' (1723130) panicked at build.rs:7:52:
called `Result::unwrap()` on an `Err` value: "  --> 6:1\n   |\n6 | #endif // FOO\n   | ^---\n   |\n   = expected line"
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

I had hoped this might be an easy fix by adding

block_comment = @{ "/*" ~ (!"*/" ~ ANY)* ~ "*/" }
line_comment = @{ "//" ~ (!NEWLINE ~ ANY)* }
COMMENT = _{ block_comment | line_comment }

to the Pest grammar, but I think the issue arises from

ifdef_directive   =  { "#ifdef" ~ identifier ~ NEWLINE ~ line* ~ "#endif" }

requiring the line ends with strictly #endif\n (I get the same kind of failure if I add any trailing whitespace to the #endif).

But I don't think

ifdef_directive   =  { "#ifdef" ~ identifier ~ NEWLINE ~ line* ~ "#endif" ~ ANY }

is correct either, because then that'd accept

#endif GARBAGE HERE

So maybe

ifdef_directive   =  { "#ifdef" ~ identifier ~ NEWLINE ~ line* ~ "#endif" ~ WHITESPACE* ~ COMMENT? }

Or maybe it's the definition of line that doesn't leave room for trailing whitespace or comments?

I consider this a very low priority issue, but I thought I'd report it anyways.

[jrebelo]

Thanks for reporting this. It is indeed not too common in IDL to have this type of comments but it is not so great that it doesn't work. When we made this we intended only to support a very small subset of the C preprocessor but as usual these things tend to grow out of their boundaries quickly so if we have the time we will have a look into this.