I also expect that the situation will be fixed eventually, but in the meanwhile, can I issue a feature request? I see the Issues page is disabled on the dmd github page: https://github.com/D-Programming-Language/dmd Is there any reason for that?
On Sun, Nov 20, 2011 at 1:31 PM, Jonathan M Davis <[email protected]> wrote: > On Sunday, November 20, 2011 13:20:00 Nicolae Mihalache wrote: >> If this module contains only functions, it looks fine. But I see it's >> manually written. >> >> I would think it is possible to generate an index with all the functions, >> classes, structs, etc and with a short description of each and with a link >> to the full description. >> >> Currently all the top level functions, classes, structs, method of classes >> and structs appear in one list and it's difficult to say which is what. >> >> Maybe I'm too biased by javadoc. I will try to work a bit more with D and >> perhaps I will have a better idea. > > There's no reason why ddoc shouldn't be capable of generating a nice set of > links at the top that give proper structure. The problem is that the anchors > that it currently generates only hold the function name, so all hierarchical > information is lost. Once that's fixed, then it shouldn't be a problem to have > hierarchical links, but at the moment, it requires creating the links by hand, > and in the case where multiple functions in the same file have the same name > (including between user-defined types) results in a set of identical anchors, > so you can't link to them individually, so you _can't_ have proper links in > that case without editing the file after it's been generated - which > definitely > wouldn't work with the official docs and would obviously be a pain regardless. > > I expect that the situation will be fixed eventually, but until then, the > links > on ddoc are limited in how well-structured they can be. Modules of only free > functions can do fairly well as long the links are created by hand (e.g. > std.algorithm), but modules which define types are currently out of luck (e.g. > std.datetime). > > - Jonathan M Davis >
