On Thu, Aug 26, 2004 at 12:35:34PM +0200, Javier Fernández-Sanguino Peña wrote: > I agree, this is definitely something we need to implement since it has > been a longstanding issue. Many other distributions/OS provide online > manpages and those are a valuable thing, it should be possible to automate > extraction of manpages from binary packages, drop them in a per-package > hierarchy (to avoid conflicts) and generate browseable index (and > searchable index) for them. It's just a matter of dedicating some time to > write the appropiate code, write it, and run it over a mirror copy with the > output available in, for example, www.debian.org/doc/manpages/
I have played around a bit with the idea. Some preliminary results I put up at http://manpages.debian.net/ . Just very basic stuff so far. You can search for the exact name of a man page or package and read them in HTML, plain text and roff source. And there is a index available, sorted by man page names. Some limitations of this early version: - No real search engine yet, therefor currently no apropos or full text search. - Doesn't handle links (symbolic and .so) at all yet - My local mirror only has i386 and all, so no man pages from other architectures, yet. - No page layout, just enough HTML display the information - Man page conversion to HTML is done with man2html (available as Debian package), I already found some annoying bugs, e.g. in the link identifying heuristics. - More static indices to come. Comments welcome. (P.S.: The man pages can also be browsed in a (very deep) directory structure, http://manpages.debian.net/manpages/, main and contrib contain links to the real files in pool which are hashed by their md5sum) Gruesse, -- Frank Lichtenheld <[EMAIL PROTECTED]> www: http://www.djpig.de/
