[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gnugo-devel] paul_3_13.8 and paul_3_13.9
From: |
bump |
Subject: |
Re: [gnugo-devel] paul_3_13.8 and paul_3_13.9 |
Date: |
Fri, 20 Dec 2002 11:29:20 -0800 |
> i have no objections besides that paul_3_13.9 is rather large and
> it will be not very convenient for me to redo it by hand. i believe
> i know the reason why merged paul_3_13.8 crashed.
OK, then let's try to merge paul_3_13.8. Arend seemed to think
he can then merge his patch. I won't back out paul_3_13.9 yet.
> these lines are wrong. paul_3_13.9 _must_ move the call to
> write_elements() below the call to write_to_dfa() (see original
> patch) in order to work properly. this is because write_elements()
> now also sorts some of them out (effectively, it only changes
> the number of elements in pattern[patno] structure) and
> write_to_dfa() needs them all.
I tried switching these and I still get crashes. That is,
it now reads:
if (database_type == DB_DFA)
write_to_dfa(patno);
write_elements(output_FILE);
state = 4;
The crashes are quite varied.
***assertion failure:
board.c:2257 - board[pos] == EMPTY near C3***
That's the most common, but we also get:
***assertion failure:
reading.c:446 - color == board[bstr] near R10***
***assertion failure:
reading.c:671 - board[epos] == EMPTY near R12***
***assertion failure:
reading.c:679 - board[Fpos] == other near G4***
Dan