- My first "goto" since … 1986? - 5 Updates
| "Alf P. Steinbach" <alf.p.steinbach+usenet@gmail.com>: Jan 10 08:37AM +0100 auto overflow( In_<int_type> char_code = traits_type::eof() ) -> int_type override { const Ptr_<const Char> p_after = pptr(); if( p_after != nullptr ) { const Ptr_<const Char> p_first = pbase(); if( p_after > p_first ) { String s( p_first, p_after ); if( char_code != traits_type::eof() ) { s << Char( char_code ); } console::write( s ); goto finished; } } if( char_code != traits_type::eof() ) { console::write_char( char_code ); } finished: init_put_area_pointers(); // return (success? traits_type::not_eof( char_code ) : traits_type::eof()); return traits_type::not_eof( char_code ); } The 1986 "goto" was in COBOL. If I remember this correctly. Cheers, - Alf |
| "Alf P. Steinbach" <alf.p.steinbach+usenet@gmail.com>: Jan 10 11:30AM +0100 On 1/10/2016 8:37 AM, Alf P. Steinbach wrote: > traits_type::eof()); > return traits_type::not_eof( char_code ); > } Well OK what else does one have to do on a Sunday's morning than posting to clc++? By guaranteeing a buffer to write to, the above code was simplified enough that the "goto" disappeared. This makes the code slightly less general, but there's not much point in writing potentially reusable code without some actual reuse visible ahead, so: auto overflow( In_<int_type> char_code = traits_type::eof() ) -> int_type override { const Ptr_<const Char> p_after = pptr(); const Ptr_<const Char> p_first = pbase(); if( p_after > p_first ) { String s( p_first, p_after ); if( char_code != traits_type::eof() ) { s << Char( char_code ); } console::write( s ); } else if( char_code != traits_type::eof() ) { console::write_char( char_code ); } set_put_area_pointers(); // return (success? traits_type::not_eof( char_code ) : traits_type::eof()); return traits_type::not_eof( char_code ); } If anyone should wonder, this is an override in a class derived from std::basic_streambuf<Char>. The interface is dirty, the naming of things is ungrokable, and different standard library implementations seem to have different opinions about how it should work... Cheers, - Alf |
| Mr Flibble <flibbleREMOVETHISBIT@i42.co.uk>: Jan 10 06:10PM On 10/01/2016 07:37, Alf P. Steinbach wrote: > } > The 1986 "goto" was in COBOL. > If I remember this correctly. No reason for that goto whatsoever: multiple return statements in a function is perfectly fine sausages. /Flibble |
| "Öö Tiib" <ootiib@hot.ee>: Jan 10 11:24AM -0800 On Sunday, 10 January 2016 12:30:54 UTC+2, Alf P. Steinbach wrote: > The interface is dirty, the naming of things is ungrokable, and > different standard library implementations seem to have different > opinions about how it should work... Attempts to have single return in C++ can become rather ugly with exceptions enabled. I have long dropped such attempts. If function fits to my monitor then I see all exits clearly (thanks to syntax highlighting). If it does not then ... that is defect ... one must be capable of writing shorter functions than 150 lines. Covering 150+ lines function with unit tests is often titanic task anyway. |
| Mr Flibble <flibbleREMOVETHISBIT@i42.co.uk>: Jan 10 08:08PM On 10/01/2016 19:24, Öö Tiib wrote: > highlighting). If it does not then ... that is defect ... one must > be capable of writing shorter functions than 150 lines. Covering > 150+ lines function with unit tests is often titanic task anyway. Enforcing single return statement makes sense in a language like C which doesn't have constructors, RAII and exceptions but is totally inappropriate in a language like C++ sausages. /Flibble |
| You received this digest because you're subscribed to updates for this group. You can change your settings on the group membership page. To unsubscribe from this group and stop receiving emails from it send an email to comp.lang.c+++unsubscribe@googlegroups.com. |
No comments:
Post a Comment