# [metapost] Metapost online, but mproof failed

Dan Luecking luecking at uark.edu
Thu Nov 17 22:08:44 CET 2011

At 08:20 PM 11/17/2011, you wrote:
>Hi,
>
>my mp goes here, thanks for your kind hearted help, it can work with mpost now
>
>The problem is that
>tex mproof chapter_1-1.eps
>
>still caught an error,  but
>(pdf)tex mpsproof chapter_1-1.eps
>is fine.
>
>along with the mproof.log

Your mproof.log file indicates you are tex-ing the wrong
mproof. Perhaps it is the conTeXt version sent by another
poster. The real mproof (the one distributed with metapost)
also produces an error ("missing \$"), but it is not the one
indicated in your log file.

The problem that the real mproof has is with the "_" character
in the file name because it is interpreted as introducing a
subscript (which causes a "missing \$" error) when it tries to
print the file name. mpsproof manages to avoid this particular
problem by "sanitizing" file names prior to printing.) mproof
has no problem with your figure if the name is changed.

I believe we would have diagnosed this several messages ago
if you had supplied the error message.

>epstopdf chapter_1-1.eps
>also failed,  as well as pdflatex on the document which include this
>file( with *.eps extension, of course, i \usepackage{epftopdf} in the prologue
>of tex file);

pdflatex cannot handle eps and so either relies on epstopdf
or fails. epstopdf fails for me on your file. It appears to
be because the fonts that are used are not embedded in the
file and Ghostscript cannot find them. To embed them, one
needs to set a value
prologues:=3;
When I do that in your file, epstopdf succeeds. The fonts used
are CM fonts supplied with TeX, but normally not accessible to
Ghostscript (which performs the actual eps to pdf conversion).

On the other hand, dvipdfm succeeds because it has access to
all the TeX fonts.

>Is it enough to post only mp file and mproof log file, at least i
>can confirm nothing wrong in my latex file.

What was requested was chapter_1.mp and chapter_1.log.
Normally you would include (directly in the first email)
a complete listing of an .mp file that failed (reduced to
only the elements needed to exhibit the stated problem),
together with a verbatim report of the error messages.

You would usually supply your complete .mp file and its log
only if requested.

I did also suggest sending mproof.log, as it could
explain why the "tex mproof" command failed. However,
in this case, only the error message would have been
needed.

Regards,
Dan

Daniel H. Luecking
Department of Mathematical Sciences
Fayetteville, Arkansas
http://www-cs-faculty.stanford.edu/~knuth/iaq.html