On Wed, Jun 13, 2007 at 09:07:36AM +0200, Stefan Schimanski wrote: > Back to topic, here is the patch. The problem was that > InsetMathKern::metrics's logic was wrong: it computed the kerning, > compared it to the old one, but forgot to update the dim parameter.
Introduced in 17899, which, according to our new logic, is my doing... Andre'
