invisible arrows using xypic have zero length

First, thank you very much for this wonderful editor. I have a question regarding invisible arrows with xypic. When I was trying to use ar@{}, the arrow was not only invisible, but also of zero length (so that the label appeared at the beginning of the arrow instead of at its middle). Some examples are the blank arrows in the following xy-matrix:

Is there any way to overcome this problem? Thanks very much in advance!

Have you any alternative examples showing the ar@{} approach working. While we've allowed the xypic package to be used we're not easily able to support it implementation. However, its possible that the way we've implemented it is causing the broken effect.

Will.

First, it is great that you've allowed the xypic package, thanks a lot for this!

The invisible arrow issue is not nearly as important as the fact that you've allowed producing xypic diagrams.

If this sort of question raises second thoughts about supporting xypic, then I'm withdrawing the question :) (seriously, though)

To the point, I have tried a couple of examples with invisible arrows, but all had the same problem (while producing the correct result when compiled in the usual way.

Though we've not looked into this yet, I suspect the error is due to the fonts we have installed.

If you've time to send us an pdf of a perfect looking picture that you've created using some other method and the correspond code that shows the error on CodeCogs this should help with debugging.

Regarding our inclusion of xypic, unless we find it significantly degrades performance or has a security liability we won't remove it.

Thanks again for supporting xypic (unless it causes a significant problem, of course).

I've prepared a latex file including several xymatrices with ar@{}, as well as the corresponding compiled pdf file (in which the invisible arrows are looking perfect). All of the xymatrices in the file produce the error on CodeCogs.

To what email address should I send these files as attachments?

Thank you very much for considering this problem!

VC1