Another BoostBook/Doxygen bug?

Is it possible to get the BoostBook/Doxygen documentation system to include my commentary? This is the header file #include <boost/child/detail/pipe_mode.hpp> namespace boost { namespace child { /// A stateless class providing constants and typenames. struct pipes_base { /// A type used to specify how to communciate with the child process. typedef detail::pipe_mode pipe_mode; /// Write to the child's standard input stream. static const pipe_mode child_stdin = pipe_mode(1); /// Read from the child's standard output stream. static const pipe_mode child_stdout = pipe_mode(2); /// Read from the child's standard error stream. static const pipe_mode child_stderr = pipe_mode(4); /// Redirect the child's stderr to its stdout. static const pipe_mode redirect_child_stderr = pipe_mode(8); /// No pipes are to be used. static const pipe_mode no_pipes = pipe_mode(16); }; } // namespace child } // namespace boost And below is the generated documentation. See how all the descriptions of the constants have dissappeared. Also see how the typedef has been 'modified'. Regards, Angus Struct pipes_base boost::child::pipes_base ? A stateless class providing constants and typenames. Synopsis struct pipes_base { // types typedef unspecified pipe_mode; // A type used to specify how to communciate with the child process. static const pipe_mode child_stdin; static const pipe_mode child_stdout; static const pipe_mode child_stderr; static const pipe_mode redirect_child_stderr; static const pipe_mode no_pipes; };

Angus Leeming wrote: And one more... This class: class process_instance : noncopyable, public pipes_base { public: process_instance(process_data const & data) : data_(data) {} private: void operator=(process_instance const &); process_data data_; }; Generates this documentation. class process_instance : public boost::child::pipes_base { public: // construct/copy/destruct process_instance(process_data const &); process_instance& operator=(process_instance const &); // public member functions // private member functions }; Assignment operator has become public. (Same bug as yesterday?) The class no longer derives from noncopyable. Are these reports useful? Should I keep them coming? Regards, Angus

On Fri, 18 Jun 2004, Angus Leeming wrote:
Assignment operator has become public. (Same bug as yesterday?) The class no longer derives from noncopyable.
BoostBook really needs to start dealing with access control some time soon :)
Are these reports useful? Should I keep them coming?
Yes, they're useful. I've just recently moved and changed jobs, so BoostBook-related work has been rather slow. I'm batching everything up for a day of BoostBook hacking to fix all of these bugs :) Doug

Douglas Paul Gregor wrote:
Are these reports useful? Should I keep them coming?
Yes, they're useful. I've just recently moved and changed jobs, so BoostBook-related work has been rather slow. I'm batching everything up for a day of BoostBook hacking to fix all of these bugs :)
Ok, then here's one more for that wet and windy day of hacking ;-) namespace boost { namespace child { class pipes_base{}; class process_instance : noncopyable, public pipes_base {}; } /* namespace child */ } /* namespace boost */ Is documented as: boost::child::process_instance ? Synopsis class process_instance : public boost::child::pipes_base {}; Note how "pipes_base" has the full namespace signature, even though it is in the same namespace as "process_instance", which doesn't. Regards, Angus
participants (2)
-
Angus Leeming
-
Douglas Paul Gregor