[Xy-pic] Xypic infinite loop

Ross Moore ross at ics.mq.edu.au
Wed Mar 12 00:55:09 CET 2008

Hello Richard,

On 06/03/2008, at 7:53 PM, Richard Garner wrote:

> Dear all,
> Below is a minimal example of a tex file which seems to send xypic  
> into an
> infinite loop, leading to a stack overflow.

Yes, it does indeed.

> ----
> \documentclass{article}
> \usepackage[arrow, matrix]{xy}
> \begin{document}
> \[
>      \xymatrix{
>      **{!/l 10em/} a \ar[r] & b
>      }
> \]
> \end{document}
> ----
> The problem seems to go away if:
> - you remove the \ar[r];
> - you change the /l 10em/ to any other <vector> (cf. xyrefer, p.8);
> - you add the option \CompileMatrices to the preamble.
> I think this may be a bug in xypic. Does anyone have any
> suggestions?

I've traced the problem to how the \xy@@ix@  macro
is redefined at different stages of the processing.
The problem only occurs when there are object-modifiers,
specified using  &**{....}  and perhaps not all modifiers
are affected, but sliding most certainly is.

Put the following coding into your (LaTeX) document preamble,
after the Xy-pic package has been loaded.

   \edef \next@{={\DirectionfromtheDirection@ }}%
   \expandafter \toks@ \next@ \plainxy@
   \xyFN@ \OBJECT@}
\let\xy at entry@@norm=\entry@@norm
\def\entry@@norm at patched{%
   \xy at entry@@norm }
\AtBeginDocument{\let\entry@@norm\entry@@norm at patched}

It certainly fixes your example.
Hopefully it will not introduce errors in other situations.

Kris Rose, if you are listening, there is quite a problem here
that needs fixing. Contact me please, so we can discuss whether
the above patch has any hidden land-mines.

> Thanks,
> Richard

Hope this helps,


Ross Moore                                         ross at maths.mq.edu.au
Mathematics Department                             office: E7A-419
Macquarie University                               tel: +61 +2 9850 8955
Sydney, Australia  2109                            fax: +61 +2 9850 8114

More information about the xy-pic mailing list