References: <20170607101428.GA17779@hzdr.de>
            <acfdbc42-9291-c724-ca65-c39da836f59e@morningstar2.co.uk>
            <1502234312.27138.49.camel@sipsolutions.net>
            <d2f013ff-3f09-e369-8cf6-19fa40baf4ec@morningstar2.co.uk>
            <1503265946.27303.13.camel@sipsolutions.net>
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101
            Thunderbird/52.3.0
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Language: en-GB
Content-Transfer-Encoding: 8bit
Message-ID:  <86fd0451-18f7-98ed-e8b2-5dea720ae6da@morningstar2.co.uk>
Date:         Tue, 22 Aug 2017 09:29:58 +0100
Reply-To:     Mailing list for the LaTeX3 project
              <LATEX-L@LISTSERV.UNI-HEIDELBERG.DE>
Sender:       Mailing list for the LaTeX3 project <LATEX-L@LISTSERV.UNI-HEIDELBERG.DE>
From:         Joseph Wright <joseph.wright@MORNINGSTAR2.CO.UK>
Subject: Re: \hbox_set:Nn issue after expl3 [2017/05/13] update
To:           LATEX-L@LISTSERV.UNI-HEIDELBERG.DE
In-Reply-To:  <1503265946.27303.13.camel@sipsolutions.net>
Precedence: list
Envelope-To: <rainer.schoepf@GMX.NET>
Status: R

On 20/08/2017 22:52, Benjamin Berg wrote:
> Hi,
> 
> On Wed, 2017-08-09 at 09:50 +0100, Joseph Wright wrote:
>> On 09/08/2017 00:18, Benjamin Berg wrote:
>>> just ran into the same issue myself. I haven't been able to solve my
>>> usage yet but noticed in the processes that \vbox_set_top:Nw seems to
>>> be still missing.
>>
>> Could you give a bit more detail of the case you need to solve?
> 
> I would expect it to be added either way, just so that the API is the
> same for the different boxes :)

Yes and no :)

The current box support goes back a *long* way in its current form, but
that in part reflects the fact that much of what the team have done with
expl3 hasn't really been box-oriented. It's at least possible that this
rather low-level approach may be revised. So we are adding things we
need, but not necessarily 'all the things'. The vbox vs. vtop issue is
one we probably need more thought on for lots of reasons ...

That said, at least at present it's likely best to have a full set:
if/when we need to revisit provision, knowing what people are using is
going to be important.

> Anyway, the case I am trying to solve is a macro which in reality puts
> the content into a box (this can be a vtop or hbox depending on the
> command).
> 
> So my user would type something like:
> 
>   \question[text=some text]{\verb^some text^}

Eeek: verbatim-in-an-argument. I assume you know this isn't a great plan.

> The reason to use \vtop is a bit more odd. I was unhappy that I
> wouldn't get the correct interline skip for the first row of the vbox.
> So my solution was to put it into a \vtop to measure the height and
> then unpack it into a normal \vbox to get the correct \prevdepth value
> after the box has been added to the vertical list.

This sort of thing is the point I was eluding to above: one could well
argue that the team should be providing interfaces that deal with such
things without ever asking programmers to worry about them.

Joseph