search

uuverifiers / eldarica

The Eldarica model checker
Other
81 stars 23 forks source link

Template parser #2

Closed vladrich closed 8 months ago

vladrich commented 10 years ago

This is an attempt to improve error handling, in particular to avoid masking errors.

Disclaimer: it is not clear to me what the code at https://github.com/vladrich/eldarica/blob/template-parser/eld-client#L118 is supposed to do. It might or might not work.

pruemmer commented 10 years ago

Thanks for the changes!

I am getting some strange behaviour with eld-client on my machine though (and don't really understand all the shell magic you are using):

philipp@hal3:~/tmp/eldarica/eldarica$ ./eld-client regression-tests/horn-prolog/listcounter.error.nts.horn trap: ERR: bad trap NOT SOLVABLE philipp@hal3:~/tmp/eldarica/eldarica$ ./eld-client regression-tests/horn-prolog/listcounter.error.nts.horn -log trap: ERR: bad trap Initially: 35 clauses philipp@hal3:~/tmp/eldarica/eldarica$

wrong shell?

The original idea of https://github.com/vladrich/eldarica/blob/template-parser/eld-client#L118 was to only start nc once to connect to the server. If the connection fails, in line 118 an attempt will be made to start a new server.

I'm a bit concerned about the additional delay created by calling nc twice for every run of eld-client (once to check whether the server is still alive, once to actually connect to the server).

Philipp

pruemmer commented 10 years ago

Thanks for the changes!

I am getting some strange behaviour with eld-client on my machine though (and don't really understand all the shell magic you are using):

philipp@hal3:~/tmp/eldarica/eldarica$ ./eld-client regression-tests/horn-prolog/listcounter.error.nts.horn trap: ERR: bad trap NOT SOLVABLE philipp@hal3:~/tmp/eldarica/eldarica$ ./eld-client regression-tests/horn-prolog/listcounter.error.nts.horn -log trap: ERR: bad trap Initially: 35 clauses philipp@hal3:~/tmp/eldarica/eldarica$

wrong shell?

The original idea of https://github.com/vladrich/eldarica/blob/template-parser/eld-client#L118 was to only start nc once to connect to the server. If the connection fails, in line 118 an attempt will be made to start a new server.

I'm a bit concerned about the additional delay created by calling nc twice for every run of eld-client (once to check whether the server is still alive, once to actually connect to the server).

Philipp