Re: Regression in WoMan

[Sven Joachim <svenjoac@gmx.de>]

On 2008-01-07 00:11 +0100, Juri Linkov wrote:

>> Running WoMan on the attached file reveals a regression in
>> EMACS_22_BASE: press TAB and the cursor jumps to the "#include
>> <stdlib.h>" after the SYNOPSIS, press TAB again and it stays there.
>>
>> In Emacs 22.1, pressing TAB twice jumps to the "atof(3)" after the SEE
>> ALSO near the end, further TABs cycle through the references.
>
> This is caused by the 2007-06-04 change in man.el (Man-next-section)
> discussed in
> http://lists.gnu.org/archive/html/emacs-devel/2007-05/msg00243.html
> that ignores the last line of the man page as not belonging to
> any section.

I see.  The assumption that the last line always contains a footer seems
to be valid if the manpage is rendered by 'man', but 'woman' does not
insert such a footer (more precisely, it deliberately deletes the footer,
see the woman-man-buffer function).

>  Your file doesn't contain the last line in the format that
> contains the name of the current man page and a timestamp, so the last
> section "SEE ALSO" gets ignored by Man-highlight-references.
>
> It seems there was an invalid assumption that the last line should be
> ignored unconditionally.  Maybe some additional check should take care
> that the last line in the standard format (I can't find where it is
> documented).

There seem to be three possibilities to solve the issue: have woman
insert a footer like man does; differentiate the treatment of the last
line between man and woman ; or remove the footer in man buffers as
well.

What would be the best?

Sven