Re: [Texmacs-dev] Applied patches for version 1.0.1.15

[Nix N. Nix]

On Mon, 2003-06-23 at 08:05, Joris van der Hoeven wrote:
> 1286, 1400, 1453, 1505, 1506, 1509, 1514, 1529: Applied

Cool !

> 
> 1439: Applied, but I got some errors when using figures:
> 
> ------------------------ error -------------------------
> ! Argument of address@hidden has an extra }.
> <inserted text>
>                 \par
> l.662    \end{array}\right)$.} \]}}
>                                    }
> ------------------------ source ------------------------
> {\image{\tmfloat{h}{small}{figure}{\quad \epsfig{file=Uconv1.ps}
> \quad}{\label{Uconv1-fig}A region of the form
> $\tmmathbf{\mathcal{H}}_{\tmmathbf{U}} \cdot \tmmathbf{\zeta}$, where
> \[ \text{$\tmmathbf{U}= \left(\begin{array}{cc}
>      - \sqrt{1 / 2} & \sqrt{1 / 2}\\
>      - \sqrt{1 / 2} & - \sqrt{1 / 2}
>    \end{array}\right)$.} \]}}}
> --------------------------------------------------------

Unfortunately, LaTeX does not seem to allow passing \[ \] as an argument
to a function.  I'm surprised this didn't fail before, since, before my
patch, the LaTeX from such a figure contained

...
    \caption{\[ ... \]}
...

which also fails (as I have confirmed).

Do you think it would be a good idea to force $ $ instead of \[ \]
inside an argument to a function ?
> 
> 1447: Assigned to David; should I apply it, or should David still
>   make some modifications?

It seems OK to me - minus perhaps the 2D vector stuff, which is probably
too generic to be in the tmtable.scm file.  However, I don't know where
that code belongs, but tmtable needs 2d vectors, if it is to do
cell-level borders.