![](https://secure.gravatar.com/avatar/49e63acb01f2ca80efce7eed08310ce8.jpg?s=120&d=mm&r=g)
On 14-4-2010 8:58, Hartmut Henkel wrote:
but it is so since ancient pdftex times (and i have also stumbled over it, and since then i'm wrapping \pdfrefximage in a \hbox, as you suggest). The \pdfrefximage doesn't do any h or v mode change, it's neutral. And i guess one should keep it like this to avoid incompatibilities.
indeed, and i can imagine the why, but i can also imagine that in luatex we wrap it in an hbox as we're not bound to compatibility in luatex; it can even be conditional: if in hmode then no wrapping, else wrap it in an hbox; alternatively when issues in vmode we couldl set a flag that indicates that in the backend the image is lapped up (tlap) so that the effective height in the page stream is zero (actually that would be the cleanest solution as something similar happens with annotations) a comparison with specials is not right as specials are not looked at this way for pdftex i can imagine that a warning is issued, something: "injecting whatsit node with dimensions into a vertical list; this confuses the page builder" or something like that actually, it's a problem with all whatsits that cary dimensions so when we're further along the luatex road we can consider a special kind of hlist subtype for this kind of things Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------