At 16:49 12.07.2001 +0200, I wrote:
>Forgot this one. With \write18 enabled, everything works fine now, but not
>without... \runMPgraphicsfalse (or at least not ..true) in cont-sys.tex is
>necessary, otherwise texexec assumes on-the-fly generation and does not
>call mp afterwards.
>The problem is: With
>texexec --automprun --pdf test
>mp is called after the tex runs, but no picture is produced (and possibly
>that's why no final tex run follows). This is the entire log file mpgraph.log:
>
>This is MetaPost, Version 0.641 (MiKTeX 2.1) (preloaded mem=metafun
>2001.7.12) 12 JUL 2001 16:42
>**mpgraph
>(mpgraph.mp )
>
>However, compiling mpgraph.mp from the command line via
>mp mpgraph
>works fine!
>
>I can't explain why mp is terminated prematurely when called by texexec.
>Even more strange, Hans might remember having told me about the --autmprun
>switch. Everything worked properly on my machine, perhaps one of the
>latest updates causes problems (but I can't tell which one, if so, for I
>didn't use mp since then)?
I have to make some additions to this. In principle, everything works well,
so no problem with texexec due to updates. But there are problems with more
complex MP code, possibly with metaobj. My test file:
\startreusableMPgraphic{nometaobj}
draw unitsquare scaled 1cm;
\stopreusableMPgraphic
\startreusableMPgraphic{justinput}
input metaobj;
draw unitsquare scaled 1cm;
\stopreusableMPgraphic
\startreusableMPgraphic{metaobj}
input metaobj;
newBox.testbox(btex Where is it? etex);
testbox.c=origin;
drawObj(testbox);
\stopreusableMPgraphic
\starttext
\reuseMPgraphic{...}
\stoptext
witth {...}={nometaobj} as well as {...}={justinput}, everything works
fine. But with {...}={metaobj}, I get the nearly empty log above. Compiling
via mp mpgraph works. So, this might be some interference with metaobj, or
perhaps a general problem with complex MP code (but I can't imagine this
looking at the available examples...)
And as I wrote before, with runtime mp via \write18, all three test
graphics work.
Eckhart
