Re: mandoc HTML output: minor issues

[Jackson Pauls <jacksonpauls@gmail.com>]

Hi Anthony,

On 29 August 2017 at 09:47, Anthony J. Bentley <anthony@anjbe.name> wrote:
>> 4. Running v.Nu (https://validator.github.io/validator/) on a
>> collection of HTML files generated by mandoc picks up duplicate IDs,
>> and a bunch of other issues:
>> * unescaped characters in href attributes (%, "),
>> * unescaped characters in URL fragments (\, {, }, #, ^, [, ], <, |),
>> * <div>s appearing inside <pre>s, and
>> * mismatches between column count in <colgroup> and table rows.
>> I can gather a bunch of examples if of interest.
>
> Please do. mandoc(1) could use some improvement when it comes to
> producing valid HTML.

Ok, here goes:


Example of unescaped % in sudo.8:
 403 .It Li %H
becomes
375       [...]<a class="selflink" href="#%H">[...]
should be
<a class="selflink" href="#%25H">


Example of wrongly escaped double quote in sshd.8:
475 .It Cm command="command"
becomes
337   [...]<a class="selflink" href="#command=&quot;command&quot;">[...]
should be
<a class="selflink" href="#command=%22command%22">


Example of unescaped space character in URL path in sudoers.5:
3378 For more information see
3379 .Xr "GROUP PROVIDER PLUGINS" .
becomes
2305     For more information see <a class="Xr" title="Xr"
href="/1/GROUP PROVIDER
2306       PLUGINS">GROUP PROVIDER PLUGINS</a>.</dd>
the href should be "/1/GROUP%20PROVIDER%20PLUGINS"


Example of unescaped { and } characters in fragment in visudo.8:
347 .It Li Warning: unused {User,Runas,Host,Cmnd}_Alias
becomes
242 [...]<a class="selflink"
href="#Warning:_unused_{User,Runas,Host,Cmnd}_Alias">[...]
should be
<a class="selflink" href="#Warning:_unused_%7BUser,Runas,Host,Cmnd%7D_Alias">

There are other issues with unescaped characters in URL fragments,
which can only contain a limited set of characters:
https://stackoverflow.com/a/26119120


Example of a <div> inside a <pre> (cause by a missing </pre>) in rpld.conf.5:
 96 .Bd -literal -ffset indent
 97         foo = something;
 98 or
 99 .Bd -literal -ffset indent
100         bar;
101
102 .Ed
becomes
 52 <div class="Bd" style="margin-left: 0.00ex;">
 53 <pre class="Li">
 54         foo = something;
 55 or
 56
 57 <div class="Pp"></div>
 58 <div class="Bd" style="margin-left: 0.00ex;">
 59 <pre class="Li">
 60         bar;
 61
 62 </pre>
 63 </div>
Maybe that's just an issue with the manual (missing .Ed), but maybe on
hitting the second .Bd mandoc(1) could assume an .Ed for the prior
.Bd.


Example of column count issue in tmux.1:
 475 .Bl -column "XXXXXXXXXX" "X"
 476 .It Sy "Token" Ta Sy "" Ta Sy "Meaning"
becomes
 372 <table class="Bl-column">
 373   <colgroup>
 374     <col style="width: 15.00ex;"/>
 375     <col style="min-width: 1.00ex;"/>
 376   </colgroup>
 377   <tr class="It-column">
 378     <td class="It-column"><b class="Sy" title="Sy">Token</b></td>
 379     <td class="It-column"><b class="Sy" title="Sy"></b></td>
 380     <td class="It-column"><b class="Sy" title="Sy">Meaning</b></td>
 381   </tr>
v.Nu says "error: A table row was 3 columns wide and exceeded the
column count established using column markup (2)," but in practice I'm
not sure how important this is.


Finally, one bonus oddity I just noticed, an opening paren "(" going
stray in dnstop.8:
165 .It *
166 show sources + 8th level query names
167 .It (
168 show sources + 9th level query names
becomes
178   <dt class="It-tag">*</dt>
179   <dd class="It-tag">show sources + 8th level query names</dd>
180   (
181   <dt class="It-tag"></dt>
182   <dd class="It-tag">show sources + 9th level query names</dd>


Cheers, Jackson

www.mankier.com
--
 To unsubscribe send an email to discuss+unsubscribe@mandoc.bsd.lv