Subject:
|
Re: Ghost bug in 3.11
|
Newsgroups:
|
lugnet.cad.mlcad
|
Date:
|
Wed, 3 Nov 2004 19:03:37 GMT
|
Viewed:
|
9003 times
|
| |
|
|
In lugnet.cad.mlcad, Lee Gaiteri wrote:
> In lugnet.cad.mlcad, Kevin L. Clague wrote:
> > Did you use 1.1.5 and the -mf or -mF options? I know these options make
> > ldglite ignore STEP.
>
> Didn't try those, though I doubt they'd make it ignore CLEAR.
>
> > The step dats that LPub makes do have STEP and ROTSTEP, but eliminate buffer
> > exchage and clear and such, before passing the step dat to ldglite.
>
> Sounds good to me.
>
> > > Well it turns out, you can ghost meta instructions just fine in MLCad, but it
> > > never saves them as ghosted. The only way to ghost them properly is to manually
> > > edit the .mpd file in a text editor. MLCad recognizes on load that they're
> > > supposed to be ghosted, but once you make any changes they go back to normal. So
> > > MLCad is saving ghosted meta commands as unghosted.
> >
> > I don't think you want to GHOST out those commands. This is not what GHOST is
> > supposed to be used for. In MLCAD you might want to preceed these commands with
> > WRITE (which MLCAD uses to mean comment), also // would also work.
>
> I'm not sure I follow why ghosting these would be a bad idea. Since ghosting is
> only supposed to hide something if it's not the main model, ghosting ROTSTEP and
> STEP and CLEAR would cause no different behavior in MLCad, but would prevent
> programs like ldglite from incorrectly executing them in a submodel.
The thing is, that STEP and ROTSTEP are implicitly ghosted for a given sub-model
used in a model.
When you are generating images for the sub-model itself, STEP and ROTSTEP have
meaning and affect the generated images.
When you add the sub-model to the model, the STEPs and ROTSTEPs in the added
sub-model do not imact the rendering of the model. Only STEPs and ROTSTEPs
within the model impact the rendering of the model.
This behavior is the same as for GHOSTed parts.
CLEAR is the same way. A CLEAR intermixed with STEPS/ROTSTEPs in the sub-model,
impact the rendering of the sub-model's step by step images. Yet, when we use a
submodel in the main model, and the submodel has CLEARs, those CLEARs do not
impact the rendering of step images for the main model.
The issue here is that STEP, ROTSTEP, BUFXCHG, and CLEAR in ldglite does not
work the same as MLCad. Using GHOSTing as you propose merely provides a
workaround for the deficiencies of ldglite. Once Don gets those meta-commands
working, why bother to use GHOST?
I don't think it is a good idea to overload the meaning of GHOST to work around
bugs in renderers.
Your proposal would require that the deficient renderers support GHOST. Why do
that work, when you could just get those renderers to function properly.
MLCad has the ability to render ignoring steps/rotsteps or honoring them. LPub
does the same. Why not just use LDGLite's -D option to make it do the same?
Welcome into the LDraw fold.
> Lummox JR
Kevin
|
|
Message is in Reply To:
 | | Re: Ghost bug in 3.11
|
| (...) Didn't try those, though I doubt they'd make it ignore CLEAR. (...) Sounds good to me. (...) I'm not sure I follow why ghosting these would be a bad idea. Since ghosting is only supposed to hide something if it's not the main model, ghosting (...) (20 years ago, 3-Nov-04, to lugnet.cad.mlcad)
|
6 Messages in This Thread:
 
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
