[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: Updates to man-k.org
On Sat, May 13, 2017 at 5:44 AM, Robert Elz <kre%munnari.oz.au@localhost> wrote:
> Date: Sat, 13 May 2017 01:44:35 +0530
> From: Abhinav Upadhyay <er.abhinav.upadhyay%gmail.com@localhost>
> Message-ID: <CAHwRYJ=baNR3jsz4ov4zL6n=RFJLg4R-ZrOpYXbLgNuU_P0hJg%mail.gmail.com@localhost>
> | If you have any feedback or improvements in mind, please let me know :-)
> Have you considered (is it possible?) to do multi-platform searches.
> That is, in the "which release" menu, add "all", and ideally also:
> "all NetBSD" (all the NetBSD-* releases available)
> (similar for FreeBSD, OpenBSD, linux, posix)
> "all stable" (NetBSD 7.1, FreeBSD 11.0 OpenBSD..., Linux...)
> (probably also the earliest, maybe newest, posix)
> "all development" (*-current or whatever, and the most recent posix)
> In these cases the output would also need to identify (for each result
> returned) which release it came from. This would make it much easier
> to look for differences in the manual for xxx between NetBSD 6 and 7
> (for example) or between NetBSD and FreeBSD.
That's an interesting idea. I have different databases for each of the
operating systems' man pages, I would have to fire the query on each
of them and combine the results. It will probably be simpler to put it
on a new html page.
Another challenge will be the ranking of the results. I guess I could
take the top 5 results from each of the databases and combine them.
> If you can arrange to look in more than one release, and identify the
> output, adding the various combinations as search choices should be
> easy (you could even provide a checkbox interface, instead of a dropdown,
> and allow the user to pick whatever releases to search are desired.)
Yes, a check box based interface would be more appropriate.
> A way to compare man pages would be really nice ... (this would be hard,
> diff'ing the formatted versions would not be productive, and diff'ing
> the raw data would not produce something nice to look at, so it would
> need to diff the raw, then find the places in the formatted output where
> the differences occur, and show those.) This sounds like a very big
> project however...
mandoc(1) produces an AST of the man pages, possibly the AST could be
compared. Still a hard problem :)
> Also, I don't know where the posix-* data comes from, but posix-2016 ought
> be available now (that is, it is released, whether you can get at it or not
> for this purpose I have no idea.)
The posix man pages are from the Linux man-pages project. They have
been granted permission from IEEE and The Open Project to distribute
the posix man pages: https://lwn.net/Articles/581858/
I don't see any posix-2016 man pages available yet there:
Thank you for the feedback and the ideas. I will try to work on adding
search across releases :-)
Main Index |
Thread Index |