Bjarni Ingi Gislason <bjarn...@rhi.hi.is> wrote:
An: groff@gnu.org
Datum: 06.03.2015 22:57
Betreff: [Groff] Lack of professionalism, thinking, education, understanding,
wisdom, consequences, [references] ...
> A trigger was an output from my "man"-script with argument "mg"
>
> <standard input>:10: warning: unbalanced .el request
>
> (NB. Line number is wrong, because ...)
>
> Another script showed more:
>
> /usr/share/groff/1.22.3/tmac/mdoc/doc-common:690: backtrace: string `Dd'
> /usr/share/groff/1.22.3/tmac/andoc.tmac:49: backtrace: macro `Dd'
> <standard input>:4: warning: unbalanced .el request
>
> Input file is /usr/share/man/man1/mg.1.gz
>
> The part of the file around line number 690 is
>
> .de Dd
> .ds command-name
> .ie \n[.$] \{\
> .ie "\$1"$Mdocdate:" \
> .ds date-string \$2\~\$3, \$4
> .el .ie (\n[.$] == 3) \
> .ds date-string \$1\~\$2 \$3
> .el \{\
> .ds date-string "\*[date-\n[mo]]
> .as date-string \~\n[dy], \n[year]
> .\}
> .\}
> .el \
> .ds date-string Epoch
> ..
>
> I did not see at first anything wrong, there are two left curly
> brackets and two right ones. Each "ie" and "el" has a one line
> argument or a block, but the code is FLAT, there is NO STRUCTURE
> visible.
There is indentation in the macro Dd in doc-common.
And as far as I see there is no bug with .ie/.el in this macro.
It is ok to write
.ie ....
.el .ie ...
.el ...
For every .ie one .el is required and this is the case in this macro.
