Just the version-string in the ispin script is still set to "1.1.0" which was a little confusing at first.
]]>* there are additional linebreaks in statements (not critical I guess), and
* there is no ordering of statements any more, i.e. there used to be all nodes first, then all edges listed in the dot file.
The grammar for the dot language does not specify a certain ordering. However, older versions of dot seemed to produce this ordering in their output files. If any parser implicitely relies on a certain ordering of the dot-statements, then it now may fail.
I did not look very deep in the spin script, but I interpret the error "invalid bareword e" while parsing the dot file as follows:
Nodes and transition have an attribute "pos=..." with a numeric value for nodes, while the value for transitions starts with "e,...". So parsing a transition while expecting a node gives a symbol that makes no sense in numerical calculations.
Maybe this helps to make ispin work with the actual dot-versions.
A workaround could be to sort the dot file before parsing.
]]>dot - graphviz version 2.31.20130115.0545 (20130115.0545)
Would be interesting to see if you can reproduce the error with a version 2.29 or higher. For Ubuntu there is version 2.30 available.
Otherwise it's a specific problem of the macos-version. Older version-numbers for mountain lion are not available.
I will also try to find out if the format has changed and compare it with an example file from a different machine.
]]>invalid bareword "e"
in expression "50 * e=";
should be "$e" or "{e}" or "e(...)" or ...
invalid bareword "e"
in expression "50 * e=";
should be "$e" or "{e}" or "e(...)" or ...
(parsing expression "50 * e=")
invoked from within
"expr 50 * $mf"
(procedure "find_field" line 10)
invoked from within
"find_field "width=" $line"
(procedure "display_graph" line 129)
invoked from within
"display_graph init"
(command bound to event)
The automata view is very useful in teaching, I appreciate any suggestions.
Many thanks,
Heinz