[2016-09-08 15:41] markus schnalke <[email protected]> > > part text/plain 2487 > [2016-09-08 15:12] Dmitry Bogatov <[email protected]> > > [2016-09-08 13:53] markus schnalke <[email protected]> > > > > > > The idea of your changes is good. On the first look, your > > > code looks good as well. Just one thing: Please use shorter > > > identifiers. > > > > Shorter function name? I can't invent shorter name, that > > implies same idea. What is wrong with three-words-name? > > Have a look at the following list of the number of mmh function > names grouped by their length: > > egrep -horI '^[a-zA-Z0-9_]+ *\(' config sbr uip | sort -u | awk '{sub(/ > *\(/,"");print length}' | sort -n | uni > q -c > 5 3 > 21 4 > 25 5 > 36 6 > 61 7 > 65 8 > 49 9 > 60 10 > 39 11 > 43 12 > 33 13 > 26 14 > 10 15 > 5 16 > 7 17 > 6 18 > 8 19 > 5 20 > 1 21 > 1 22 > 2 24 > 1 32 > > (Half of the function names are 9 chars or shorter. 90% of the > function names are 14 chars or shorter.) > > Here's a list of the function names: > > egrep -horI '^[a-zA-Z0-9_]+ *\(' config sbr uip | sort -u | awk '{sub(/ > *\(/,"");print length, $0}' | sort -n | > uniq > > The list ends with: > > 19 do_readonly_folders > 19 field_encode_quoted > 19 list_single_message > 19 output_content_file > 19 parse_format_string > 19 show_message_rfc822 > 19 show_multi_internal > 19 show_single_message > 20 field_encode_address > 20 get_attachmentheader > 20 get_folder_info_body > 20 init_decoded_content > 20 store_single_message > 21 output_content_folder > 22 BuildFolderListRecurse > 24 clean_up_temporary_files > 24 get_folder_info_callback > 32 make_mime_composition_file_entry > > Your function names are: > > 18 fill_mailname_type > 23 get_alternate_mailboxes > 26 append_alternate_mailboxes > > These are all at the upper end of the list (last 5%), ranking > #486, #508 and #511 out of 512 different function. Do you see > what I mean?
I see no meaning in calculating length of function names, > [...] > > If that would be helpful, I can propose shorter names when I > had a closer look at the code, as long as you can't prove, that it is possible to unambiguously fit same information in lesser count of symbols.
