SIM works strange... about hustoj

[GoogleCodeExporter]

copy the A+B answer from faq,write into a file called Main.c

cd /home/judge/run0
sim.sh Main.c 1000

after that look into file "sim" at /home/judge/run0/sim
see what is print in it

if the big 559939584 like things found
check if the SIM is working well

sim -p Main.c Main.c

should report two 100% likeness

if sim is reporting error message , we'll need to port SIM to 64bits manually

Original comment by [email protected] on 29 Nov 2010 at 3:08

[GoogleCodeExporter]

root@lavidaus:/home/judge/run0# sim.sh Main.cc 1000
(No Output)

root@lavidaus:/home/judge/run0# sim_cc -p Main.cc Main.cc
File Main.cc: 28 tokens
File Main.cc: 28 tokens
Total: 56 tokens

Main.cc consists for 100 % of Main.cc material
Main.cc consists for 100 % of Main.cc material
root@lavidaus:/home/judge/run0# sim_c -p Main.cc Main.cc
File Main.cc: 28 tokens
File Main.cc: 28 tokens
Total: 56 tokens

Original comment by method4libe on 29 Nov 2010 at 3:28

[GoogleCodeExporter]

try
modify
http://code.google.com/p/hustoj/source/browse/trunk/core/judge_client/judge_clie
nt.cc#956
to
fscanf(pf,"%ld%ld",&sim,&sim_s_id);

Original comment by [email protected] on 29 Nov 2010 at 4:22

[GoogleCodeExporter]

I found what's the problem.

Parameter settings in wiki page for AntiCheatEnable was incorrect. It works 
well when I changed.

I suggest some optinion Anti Cheat Enabler.

1. Sometimes a person submit same or similar code several times but it can't be 
judged as cheat. However Anti-cheat system in hustoj doesn't consider it.

2. I think wrong-answer source must check not only accepted solution. I'm doing 
Teaching Assistant for Data Structure course of Ajou University in this 
semester, some people submit WA solutions do cheating.

Original comment by method4libe on 29 Nov 2010 at 5:15

[GoogleCodeExporter]

And I found sim table doesn't update in rejudging. If row in a solution is in 
the table then 'insertion query' doesn't affected to tables.

Original comment by method4libe on 29 Nov 2010 at 5:17

[GoogleCodeExporter]

try the latest web, admin/rejudge.php

Original comment by [email protected] on 29 Nov 2010 at 5:30

[GoogleCodeExporter]

'admin/rejudge.php' works well, but that problem caused by when I judge using 
'judge_client'.

Original comment by method4libe on 29 Nov 2010 at 6:18

[GoogleCodeExporter]

check r606

Original comment by [email protected] on 29 Nov 2010 at 7:24

[GoogleCodeExporter]

OK. It works well.

1. Sometimes a person submit same or similar code several times but it can't be 
judged as cheat. However Anti-cheat system in hustoj doesn't consider it.

2. I think wrong-answer source must check not only accepted solution. I'm doing 
Teaching Assistant for Data Structure course of Ajou University in this 
semester, some people submit WA solutions do cheating.

What do you think about that, newsclan?

Original comment by method4libe on 30 Nov 2010 at 8:13

[GoogleCodeExporter]

1.The status to students will warn them about similar code no matter who's code 
is similar with,even themselves.It's true but only considering to prevent the 
system from overload.The status show to 
TEACHER(source_browser&&contest_creator) will give a interface of setting SIM 
level and query. In this situation, copy of self code will not be reported, and 
will give a links to both source code and show the account who's code is copied 
from.
2.SIM works too sensitive in those case the correct code is very short like 
A+B, I don't want to push students too hard. I found copy others code and 
modify it can be a better way than doing nothing for those students who's not 
capable of finish the code by themselves. That's why I prefer to by pass WA 
codes.

any way, it's open-source you can customize your system by tweaking the code.

http://code.google.com/p/hustoj/source/browse/trunk/core/judge_client/judge_clie
nt.cc#1061

remove the second condition expression after &&  

Original comment by [email protected] on 30 Nov 2010 at 3:10

[GoogleCodeExporter]

Original comment by [email protected] on 16 Jul 2011 at 3:06

• Changed state: Done