Re: .IP without line args is not an ERROR

[Ingo Schwarze <schwarze@usta.de>]

Hi Joerg,

Joerg Sonnenberger wrote on Wed, Oct 27, 2010 at 02:43:37PM +0200:
> On Wed, Oct 27, 2010 at 12:22:24PM +0200, Ingo Schwarze wrote:

>> OK to commit this to bsd.lv, too?

> Are you sure that the handling is correct?

You mean, in man_term.c, (pre|post)_?P()?

I have not systematically tested all the cases, but i have read the
code, and it looks sane.  Besides, there are comments indicating
that Kristaps had this in mind when writing the code (e.g. in the
definition of struct mtermp).  I see no indication that anything
might be broken in that respect.  Those cases i looked at, work.

Besides, that would be a wholly seperate issue.  We are talking about
removing a non-fatal error in the parser.  The resulting AST has
always been rendered in some way.  Whether that rendering has issues
has nothing to do with the validity of the input.

> E.g. the "restore previous" behavior for missing arguments?

Actually, in that respect, man(7) is different than mdoc(7).
In mdoc, blocks can nest, but in man, if i understand correctly,
each block terminates the previous block, at least among the *P
blocks discussed here.  Thus, the values in mtermp need no
resetting at all, and those in termp just get reset to mt->offset
and to the maximum right margin, no matter what the arguments
of the block were.

Yours,
  Ingo
--
 To unsubscribe send an email to tech+unsubscribe@mdocml.bsd.lv