Sender: rich AT phekda DOT freeserve DOT co DOT uk Message-ID: <3E428516.6E9BF039@phekda.freeserve.co.uk> Date: Thu, 06 Feb 2003 15:53:58 +0000 From: Richard Dawe X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.2.23 i586) X-Accept-Language: de,fr MIME-Version: 1.0 To: djgpp-workers AT delorie DOT com Subject: Re: Add @tindex for types in docs [PATCH] References: <3E3FDCB5 DOT D2875A7E AT phekda DOT freeserve DOT co DOT uk> <2561-Wed05Feb2003174006+0200-eliz AT is DOT elta DOT co DOT il> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Reply-To: djgpp-workers AT delorie DOT com Hello. Eli Zaretskii wrote: > > > Date: Wed, 5 Feb 2003 13:45:03 +0200 (EET) > > From: Esa A E Peuha > > > > I think that it's best not to have any duplicate definitions. > > Maybe it is, maybe it isn't. Think about a user who wants to use > `fstat', but is told that `struct stat' is documented on the `stat' > page. Such a user will need to jump between these two pages back and > forth all the time. Not many DJGPP users know enough secrets of the > Info reader that will allow them to do that and remain sane (e.g., > how many of those who read this list know that the stand-alone Info > reader can show two pages at the same time in a split-window display? > try "Ctrl-x 2" some day). I didn't know about Ctrl-x 2, but I unconciously use the Emacs keystrokes. I learnt Emacs after info, so perhaps I just don't expect info to be Emacs-like enough? > But I'm not going to veto any changes that delete multiple > descriptions of data types. Just keep this gotcha in mind before you > decide. [snip] It seems to me that the main concern is maintaining the descriptions of data types. E.g.: if someone adds a field, then you need to update all the nodes == tedious and error-prone. One solution would be to move the description into a separate file and @include that file. I can see a couple of problems with that solution: (a) makeinfo would need multiple -I options, so that the data type description files can be found. (b) Less seriously, what to call the files? We have to name them, so that mkdoc doesn't find them. We already use .txi and .txh. So .txd? .txp? Problem (a) could be solved by putting all the data-type definitions in one source directory. But that moves them away from the documentation they go with. Another solution to problem (a) is to get mkdoc to do the hard work. mkdoc could generate a file containing all the -I options. Then we could invoke makeinfo like this: makeinfo --no-split @makeinfo.opt libc.tex Bye, Rich =] -- Richard Dawe [ http://www.phekda.freeserve.co.uk/richdawe/ ]