Stefan Behnel wrote: > So an invalid value for a stylesheet parameter will stop processing but will > not signal an error in "state", which is explicitly reset from ERROR to > STOPPED here. Is that correct?
To generalise this question somewhat: there seem to be a lot of places in the source that call xsltTransformError(), which sets ctxt->state to XSLT_STATE_ERROR, and then reset the status to _STOPPED. Other places in the source then check for _STOPPED for a cascading abort. So, general question: when an xsltTransformError() was written, there usually was an error, so why not just keep the _ERROR state in this case and check for (ctxt->state != XSLT_STATE_OK) for a cascading abort? That would allow users to determine errors simply by checking the result state of the transformation context. Remaining part of the original question: how can I find out that there was an error as long as the above behaviour is not in place? Stefan _______________________________________________ xslt mailing list, project page http://xmlsoft.org/XSLT/ [email protected] http://mail.gnome.org/mailman/listinfo/xslt
