Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 3 Aug 1999 11:37:59 +0930
From:      Greg Lehey <grog@lemis.com>
To:        Mike Pritchard <mpp@mpp.pro-ns.net>
Cc:        Bruce Evans <bde@zeta.org.au>, rnordier@nordier.com, cvs-all@FreeBSD.org, cvs-committers@FreeBSD.org
Subject:   Re: cvs commit: src/sbin/disklabel disklabel.8
Message-ID:  <19990803113759.J62948@freebie.lemis.com>
In-Reply-To: <199908011141.GAA02125@mpp.pro-ns.net>; from Mike Pritchard on Sun, Aug 01, 1999 at 06:41:20AM -0500
References:  <199908010038.KAA16506@godzilla.zeta.org.au> <199908011141.GAA02125@mpp.pro-ns.net>

next in thread | previous in thread | raw e-mail | index | archive | help
On Sunday,  1 August 1999 at  6:41:20 -0500, Mike Pritchard wrote:
>>>>     sbin/disklabel       disklabel.8
>>>>   Log:
>>>>   Make intelligible:
>>>>     Describe the command formats in English.
>>>>     Add references to other programs (boot0cfg, fdisk).
>>>>     Remove some old cruft, including FUD about single-level bootstraps.
>>>>     Add example of output format.
>>>
>> I object too.  Apart from the points mentioned by Robert, it adds
>> hundreds of (mdoc) style bugs.  E.g.:
>>
>> ...[trimmed]
>> ! .if t ``a''
>> ! .if n "a"
>> ! partition.
>>   .Pp
>>
>
>> Here you take a paragraph written in normal mdoc style and change dubious
>> quoting to even more dubious quoting, and then reformat the paragraph to
>> abnormal/wrong style.
>>
>> Normal mdoc style: most lines are very short, even without markup; sentences
>> and even clauses begin in column 1.
>>
>> Dubious quoting: 'a' is a C character so it should probably be quoted using
>> C character quotes.
>>
>> More dubious quoting: ``a'' looks better than "a" even in nroff output.
>> Why not use .Dq?
>
> Arghhhh!  Man pages should never use direct *roff command for
> formatting.

Why not?  Because our forefathers did it that way?

> If for some reason you can't do it in mdoc, then either we should
> fix the particular mdoc macro or define a new one to do it.

There's ample scope for fixing mdoc.  There are some really weird
bugs^H^H^H^Hdependencies; the one Brian reported is only one of them.
I keep getting gratuitous empty lines in man pages, and I'm damned if
I know where they're coming from.  It's not raw troff directives,
anyway.

> Adding direct *roff directives usually winds up messing something
> else up in the end.  Using *roff commands directly can really
> mess things up in some cases when using mdoc.  I think we should
> be able to do what you want within the current mdoc package.

I've taken the .ifs out and replaced them with .Dq's.  My own feeling
is that ``'' looks silly in ASCII.  I note that mdoc(7) includes
explicit constructs such as \\*q to represent quotes, which is wrong
in my book, since they stay that way ("") in troff output.

> I would rather not have to add anything to mdoc, just so our man pages
> are compatible with the other *BSDs, but if we really do need to
> add something, we should.  Hopefully we can talk the other *BSDs
> into adopting it.
>
> I agree with Bruce here, just use .Dq, or .Ac, or .Aq (or one of
> ther other quoting macros).
>
> As for a comment in a follow-up to Bruce's message, if you learn mdoc,
> it is not that illegible.

My statement was that it was one of the most illegible macro
packages.  That doesn't change if you learn it, but learning does
help.

> I can write most mdoc man pages from scratch off the top of my head.

But only by avoiding fancy constructs.  Have you tried tables in mdoc?
I did in one of the vinum man pages, and it made such a mess of them
(and completely blew the pagination) that I removed them and did lists
instead.

> I've converted man pages from the old -man package to -mdoc by hand
> without ever having to refer to the man pages.  Yes, it is a little
> clumbsy at times.  But having the same text repeated twice just for
> quoting make it twice as likely that someone won't correct all of
> the text when making changes.

Well, that danger exists in a lot of cases.

> If no one objects, I'll fix disklabel.8 back into true mdoc format,
> which I would probably do anyways now that I know it is using *roff
> directives directly.

Don't bother, I'm doing it myself.

Greg
--
See complete headers for address, home page and phone numbers
finger grog@lemis.com for PGP public key


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990803113759.J62948>