Re: metafun manual

[Joop Susan <jgsusan@xs4all.nl>]

Hallo,

> >'texexec test1' complains:
> >
> >Randomizer initialized to 949889678.
> >figures        : figure mpgraph can not be found
> >[1.1{/usr/share/texmf/dvips/config/pdftex.map}]
> >.........
> >Sorry, I can't find the mem file `metafun.mem'; will try `mpost.mem'.
> >(mpgraph.mp )
> >Transcript written on mpgraph.log.
> >
> >mpgraph.log says:
> 
> texexec --make --alone metafun 
> 
> will make you a metafun format (check if the mem is moved to the right
> web2c path and run mktexlsr)

Ok. I only used 'fmtutil -all' after every ConTEXt/pdftex update. Better 
to use 'texexec'?

The file is created in the current directory and I moved it to 
/usr/share/texmf/web2c. After 'mktexlsr' the file is now found.

> >This is MetaPost, Version 0.641 (Web2C 7.3.1) (mem=mpost 2000.4.26)  26 
> >APR 2000 18:56
> >**&metafun mpgraph
> >(mpgraph.mp )

No change. 'mpost mpgraph.mp' later also writes to this file.

This is MetaPost, Version 0.641 (Web2C 7.3.1) (mem=mpost 2000.4.26)  26 
APR 2000 18:56
**mpgraph.mp
(mpgraph.mp
{randomseed:=222}
 (/usr/share/texmf/metapost/context/mp-tool.mp)
(/usr/share/texmf/metapost/context/mp-spec.mp)
(/usr/share/texmf/metapost/context/mp-page.mp)
(/usr/share/texmf/metapost/context/mp-core.mp) [1] [2] [3] )
3 output files written: mpgraph.1 .. mpgraph.3

> >
> >'mpost mpgraph.mp' does generate mpgraph.1.
> >'gv mpgraph.1' shows the correct figure.
> >A second run 'texexec test1' now generates a complete PDF file.
> >
> >Can you tell me why texexec cannot get this right in one run?
> 
> Normally it generates the graphic runtime, in which case one run is needed.
> Otherwise it will run mp between runs, but it does not parse all mp files
> for changes. So, then you may need a second run. 

Is there any way of automatically detecting this situation.

Or does this mean that the only sure automated way to get a complete 
guaranteed up-to-date PDF file is to run:

rm -f test1.tui test1.tuo test1.tmp test1.top
rm -f mpgraph.[0-9]* mpgraph.mp mpgraph.log tmpgraph.* cont-opt.bak
texexec --pdf test1
mpost mpgraph.mp
texexec --pdf test1

Would it be safe to specify 'texexec --pdf --once test1' for the second 
run given that the temp-files are now all up-to-date? It seems to work for 
my current testfile. Specifying '--once' for the first run makes the 
'mpost' fail.

The first run is also necessary when arranging pages. So I could add 
'--noarrange' there as well, right?

This is the brute force solution, possibly at the expence of extra 
processing time. Is this a bullet-proof final run?

Thanks,

Joop