So, your tester is part of your submitted code…
I solved it similarly, but I defined interface in java and set implementation in my testing only.
I had problems with flushing I guess, so I left C++ and moved to Java. I’m wondering how to write two programs that one writes to others stdin…
Not needed actually, at least for this problem. You can virtually get input from a function as if you were getting from another program!
I solved FIRESC thanks to one of the past contests problems. My point is if you keep practicing you will eventually know a lot of helpful algorithms and will gain experience.
I learned a lot here in codechef and I’m glad to say that I grew up a lot here, looking back I can see how much I have improved, in this contest I could solve 5 problems and believe it or not in the first contest I participated I couldn’t solve even 1. I know there is a long way to go but what would be the fun of achieving something without any effort, right ?
My only advice is keep practicing, never give up.
@kuruma >> If I am allowed to give a suggestion, although I know that you are learning graph algorithms in your course, but I would like to tell that if you want to learn graph algorithms from other efficient source, do check the NPTEL videos by IIT Delhi, by Prof. Naveen Garg. He explains really well, and he formats the proofs so well that it seems like daily conversation. Just go through some video lectures, and you will never find graph algorithms tough. For example, http://youtu.be/CIm6RzdoPCI teaches you DFS! Very interesting and valuable. Good Luck!
I really don’t know what you have against me, but FYI I thought it was one of the best Codechef problems I’ve seen.
Thank you very much @Bugkiller, I will definitely look into these videos when I have more time to spare 
Altough atm univ is a bit more important and different things are required there…
I have nothing against you. But all my previous tries to add some tricky corner case to the problem met a strong critic from you.
@gojira: Could it be that, on the systems where your code was run, an int was larger than 32 bits ? (e.g. they had 64 bit ints) This could explain why your code reached i above 50 (as the result did not become 0 at i=34). I know this seems unlikely, but maybe it’s possible. I tried to test this behavior by submitting a simple code at several practice problems, but sizeof(int) was always 4 – but maybe there’s a different set of machines used during contests ?
My opinion is that the March 2013 contest was harder than the usual Codechef long contests. It seems that there were more HARD problems than in other such contests (I was expecting 1 or 2 HARD problems, and instead we got 3 HARD problems, with 2 tricky MEDIUM problems - ZENCALC and SUBTREE : although SUBTREE may also be considered borderline HARD, I guess). This made the contest more stressful than usual for me 
(because I wanted to solve all the problems) and also left me with less time than usual to focus on the CHALLENGE problem (which is one of the things I like best in these contests).
Anyway, on the other hand, I am also nicely impressed with the difficulty of the problems. It is not easy to come up with HARD problems which are tough to crack in a 10-day interval (many problems in other contests are HARD also because the length of the contest is shorter, but HARD problems in a 10-day contest need to be harder than hard problems in short contests).
Still, if I had to choose, I would prefer slightly easier contests in the months to come
I really don’t want to derail another thread, so this will be only my only further reply on the topic, but I have never and will never critique anyone for deliberately adding tricky corner cases. That’s what programming contests are all about. The previous issues were unrelated: a) solutions/editorials which tell you to solve the problem, you must use unsigned variables, which rules out languages like Java entirely, and b) when the problem author did not expect or handle all corner cases properly, which you admitted yourself for ANDOOR. Had they been intended, I would have had no problems
I should mention that I am not the author of ANDOOR.
I would think 1000 times before posting such problem.
Because of many issues with other problems I had no time to consider it properly, hence such crap 
So I am not the only one with such feelings. :D. Except that I took almost 6 approaches and 6-7 days to solve the FIRESC after approx and TOTR which were tip of the iceberg…
I never thought of any algorithm to solve FIRESC but feeling ashamed now that It was something I learnt in second year of Engineering…
Dude, a topper at Codechef is expecting easier problems. Not expected! Just kidding bro, all the best.
I feel ashamed at every contest basically!! 
Especially because the kind of Maths used on these contests (Discrete Maths, series, graphs theory, recurrences) is all about practicing and well, also learning from editorials! There are problems where if you don’t know the concept behind them, you simply won’t solve them!! And I really appreciate having a place where I can share my feelings towards my learning curve on these contests