[Housebot-developers] [ housebot-Bugs-1585761 ] loadsgf fails an assert() and aborts the program

[SourceForge.net <no...@so...>]

Bugs item #1585761, was opened at 2006-10-27 08:19
Message generated for change (Comment added) made by j_house
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=774851&aid=1585761&group_id=149506

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
>Status: Closed
>Resolution: Fixed
Priority: 7
Private: No
Submitted By: Urban Hafner (urbanhafner)
Assigned to: Nobody/Anonymous (nobody)
Summary: loadsgf fails an assert() and aborts the program

Initial Comment:
u@koala:~$ cd test/regression/cgtc-2.0/
u@koala:~/test/regression/cgtc-2.0$ ls
Makefile                 cut_connect.tst          ko.tst
Makefile.am              double_threat.tst       
life_death.tst
Makefile.in              enclose_breakthrough.tst
low_liberties.tst
NEWS                     endgame.tst              sgf
README                   escape_capture.tst      
territory.tst
blunder.out.html         index.html
blunder.tst              index_all.html
u@koala:~/test/regression/cgtc-2.0$ ../../../src/housebot 
loadsgf ./sgf/blunder/blunder.1.sgf 1
./status_tracker.hpp:322: failed assertion
`object_location != current_object_status.end()'
Abort trap
u@koala:~/test/regression/cgtc-2.0$ 

This happens in head (rev 308) as well as in the
urbanhafner_incremental_benson branch. It might be
worthwile to go back and find out at which revision it
started to fail.

----------------------------------------------------------------------

>Comment By: j_house (j_house)
Date: 2007-02-18 12:27

Message:
Logged In: YES 
user_id=1353569
Originator: NO

Definitely fixed in r467, but I suspect it was fixed a little bit before
that.

----------------------------------------------------------------------

Comment By: Urban Hafner (urbanhafner)
Date: 2007-01-16 11:10

Message:
Logged In: YES 
user_id=159737
Originator: YES

Having looked at it again it seems that the problem is that loadsgf
doesn't call the setup functions that are needed. It does work when
"boardsize N" is called beforehand. So loadsgf should probably call
boardsize and clear_board first.

----------------------------------------------------------------------

Comment By: Urban Hafner (urbanhafner)
Date: 2006-11-08 06:03

Message:
Logged In: YES 
user_id=159737

I'm not so sure about it. Copying the SGF file to the root
directory and then trying fails, too.

$ cp test/regression/cgtc-2.0/sgf/blunder/blunder.1.sgf .
$ ./src/housebot
loadsgf blunder.1.sgf 1
./status_tracker.hpp:335: failed assertion `object_location
!= current_object_status.end()'
Abort trap


----------------------------------------------------------------------

Comment By: j_house (j_house)
Date: 2006-10-30 18:17

Message:
Logged In: YES 
user_id=1353569

Could this be the connection brain acting up?  If so, maybe
we should alter this bug to state that we need to
standardize on which directory housebot gets run from and
how it loads its config files and pattern databases.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=774851&aid=1585761&group_id=149506