> > The "adding a system call" roadmap for kprintf() does not > have a lot of explanation. But, much the problem is just > collecting a logical view or slice of an aspect, and once you > visit those files, a lot is just lists of numbers or tables > of function pointers and descriptions -- nothing really complex. > Well, except for the detail of each system call. Some are > trick; amny are trivial. kprintf() was easy. > > But just the raw navigation aid makes it easier to explore > just this one aspect of Solaris, in relative isolation from > the rest. It beats running cscope on 40,000+ files, and saying, > "now what?". > > I created a slice of the source code limited to just these > pathnames. It is useful. I thought I might create a similar > "live" annotated slice in a web browser. I like seeing a > slice that is narrowed down to a single useful logical aspect. > There are other aspects for which this could be useful. > > I work with plain text a great deal. So, I was doing the > slice with lynx in mind, no need for eye-candy -- just > hypertext navigation, with the accompanying line or two > of description. But I didn't get around to that, just cscope. > > Some of you may have tools you prefer for preparing and > viewing slices. I am sure there is much about this that can > be improved upon.
I still use links ( a browser like lynx ) all the time and have even been known to dive into emacs and not come out for days. This is all fantastic work on your part .. really stunning. What can we do to keep you around long term ? - Dennis Clarke
