I have to make sure that a state can be saved. This means that all information leading to that state must be available. Typically the way this is done is to re-read a data file on its way to the ZIP part of the PNGJ output stream. This involves all data files and the state script.
When a model is loaded from a variable, we have to create and maintain a data structure to hold that model data so that when/if a PNGJ file is created, we have it on hand -- it was never "in" a file, and Jmol does not lay down temporary files. We recreate that data structure in the state as part of a LOAD DATA .... END command. Now, if you in memory only are creating binary data structures and loading them, then these will need to be cached and restored in their entirety. That's the issue. It's not that we couldn't *load* the data. The issue would be how to *save* it again. But I can think about it. By the way, load xxx.png doesn't really use the LOAD command -- it switches to the SCRIPT command, runs the state.spt script, which then loads files from the (cached) PNGJ data. So perhaps that is the answer to my own question. In some ways this might be like how we read data in JSmol from drag-dropped files -- we don't get the filename; we only get the byte stream. OK, so that might work... Bob
------------------------------------------------------------------------------
_______________________________________________ Jmol-users mailing list Jmol-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jmol-users