First of all, that's not what I recommended. You still naming the type struct _host which makes no sense.

As for putting an uppercase "Host" in the middle of a function name, why would would do something silly like that. Or do you mean Host_push_back akin to Perl's Host::push_back? Definitely a gain since there's no equivalent to ::. Different things should look different.

And you didn't say what you'd use instead of Host host;. What would the following look like in your system?

Host host = Host_new();
[download]

”… why…do something silly like that”

It’s not me, the library works like this:

struct _host {
  int status;
  str name;
};

typedef struct _host host;

void host_free(host *);
host host_copy(host *);

#define T host // <- custom type
#include <deq.h>

// functions to implement
void host_free(host *h) { str_free(&h->name); }

host host_copy(host *h) {
  host _ = *h;
  _.name = str_copy(&h->name);
  return _;
}
[download]

The library then provides:

container_type_push_back
container_type_pop_back   
container_type_push_front 
container_type_pop_front
[download]

…and some more, depending on the container.

If the container is a deq (dequeue) and its type named fubar this would result in deq_fubar_push_back and so on. Instead of a simple push in a well know post-modern language. One may regret this but it is like it is. And I don't know how else to do it. Apart from the fact that I can't write such macros.

Other libraries, like m*lib work similiar - m*lib does this excessively.

«The Crux of the Biscuit is the Apostrophe»

Maybe what ikegami meant was that the underscore is not needed. Struct names live in a different namespace, so you can write

typedef struct host host;
[download]

Of course, I recommend much longer struct names since C namespace is so limiting. I usually name my stuff with a project identifier prefix, like struct myproj_host. I find CamelCase and lowercase equally appealing for the prefix and object names, but tend to stick to one pattern per project. I often put a suffix of _t on my typedefs, and _p on pointer typedefs, but that would also make the CTL library generate ugly code. Whoops I just looked it up, and "_t" are reserved. So I guess I won't do that anymore. Sometimes I don't declare any typedefs and just live with 'struct' declared everywhere. It isn't pretty, but C is never really pretty and it helps document the code for people less familiar with it. I suppose the typedefs are required for that library you're using though, and with it putting prefixes on your type names for the container name that is only going to look OK with lowercase...

Maybe if the CTL library were written in cpppp it would be easier to parameterize the output :-)

(that vector class isn't finished; but maybe CTL would make a good model to base my eventual set of standard templates on)

To the other points: I haven't come to any conclusions about how I generally handle this. BTW, the idiom _struct_name type_name uses "C Programming - A Modern Approach" (King 2008) all the time. Like everything else, I didn't come up with it.

«The Crux of the Biscuit is the Apostrophe»

Well, that book contradicts the C99 standard which says that identifiers with leading underscore are reserved for file scope. (in other words, if it starts with underscore and lowercase letter, it should only be seen within one file, not shared between files via a header) I also found this stack overflow post with a summary from several different versions of the standard.