FET Forum

FET Development => Report a Bug => Topic started by: Devrim Altınkurt on November 03, 2016, 07:21:03 PM

Title: a bug, maybe?
Post by: Devrim Altınkurt on November 03, 2016, 07:21:03 PM
hi again,
are you sure that line is correct:
v5.30.5 generate.cpp line #2384

assert(nConflActivitiesRooms.at(t)==conflActivitiesRooms.at(t).count());

because i get error at this line sometimes.

Can you examine it please?
Title: Re: a bug, maybe?
Post by: Devrim Altınkurt on November 03, 2016, 07:27:14 PM
if i delete this line,
everything is working correctly.
Title: Re: a bug, maybe?
Post by: Volker Dirr on November 03, 2016, 08:42:27 PM
can you please sent me or liviu an example file, so we can reproduce the bug?
Title: Re: a bug, maybe?
Post by: Liviu Lalescu on November 03, 2016, 08:47:02 PM
Volker, I think Devrim uses a modified FET.

Devrim, I examined the code and I believe official FET-5.30.5 code is correct. See above that line how I initialize these variables - one is populated with count of a QList, the other with again the count of the same QList.
Title: Re: a bug, maybe?
Post by: Devrim Altınkurt on November 03, 2016, 09:34:27 PM
i converted fet to c#.
everything is OK but this assert line.
i will continue to investigate..
Title: Re: a bug, maybe?
Post by: Devrim Altınkurt on November 03, 2016, 09:54:55 PM
ok, i found the problem.
i have written you before at http://lalescu.ro/liviu/fet/forum/index.php?topic=2963.0 (http://lalescu.ro/liviu/fet/forum/index.php?topic=2963.0)
if i use constraint "Activity tag preferred room(s)"
and if there is not enough room for all activities
i get the error wrote before.

by the way, i converted all fet source to c#/asp.net.
and i will publish at fet.website domain in a few weeks/months.
it will be free but not open source, i will only show ads for hosting costs.
:)

P.S. : i converted base algoritm to javascript but couldn't all source.
if you want, i can share with you js codes.
Title: Re: a bug, maybe?
Post by: Volker Dirr on November 03, 2016, 10:03:25 PM
Quote from: Devrim Altınkurt on November 03, 2016, 09:54:55 PM
by the way, i converted all fet source to c#/asp.net.
and i will publish at fet.website domain in a few weeks/months.
it will be free but not open source, i will only show ads for hosting costs.
:)

FET is released under AGPL. You MUST release your converted source open source! You are not allowed to keep it closed source! Please reread the AGPL which FET use! Why do you try to steal the FET source? Open source doesn't work like that. If all guys do it like you open source software will die (or at least grow very very slow). Please read again the idea of GPL and AGPL. Looks like you didn't understood that idea.
Title: Re: a bug, maybe?
Post by: Devrim Altınkurt on November 04, 2016, 06:03:23 AM
hi Volker,
please relax.
of course i know what the AGPL is.

it was my fault.
i should have been more clearly.
I meant that "converted fet source will be open source surely"
but not entirely of my website will be open source.

best regards..
Title: Re: a bug, maybe?
Post by: Liviu Lalescu on November 04, 2016, 06:40:57 AM
Yes, you can show us the JavaScript, but I warn you that I don't know this language :)

I am interested to see your page and more details.

I hope you will find a solution that that Volker also agrees.

That assert should work for any situation. See the official code a few lines above - it is clear that the tests are OK (see lines 2314 and 2315).

It would be best for you to work on more files, starting with the same random seeds - after the generation the random seeds should also be the same. You have many official files to test on.
Title: Re: a bug, maybe?
Post by: Devrim Altınkurt on November 04, 2016, 12:05:14 PM
is randomknuth function just for generate well randomly numbers?

so, i can use like "ISAAC" or any other random number generator, can't i?
Title: Re: a bug, maybe?
Post by: Liviu Lalescu on November 04, 2016, 12:20:38 PM
Yes, you can use any another RNG (random number generator) in FET and in your program, and check that the behavior should be the same.
Title: Re: a bug, maybe?
Post by: Volker Dirr on November 04, 2016, 05:59:12 PM
i fear a bit that you didn't understood, but we will see as soon as your webpage is online. i highly recommend to reread the AGPL.