Re[2]: [gnugo-devel] sgf output in play_solo.c

[Paul Pogonyshev]

Gunnar wrote:
>>   - trees in load and analyze, ascii and possibly some other modes
>>     doesn't work well with -L option. however, previous approach
>>     didn't work well either. fixing it can involve some deeper changes
>>     in sgf handling. i think migration to trees should be finished
>>     before dealing with this problem.

> In what ways doesn't it work well?

it appends new sgf data at the end of the tree (as if there was no -L
option). that is because gameinfo_play_sgftree_rot() doesn't indicate
where it stopped on the tree in any way. i may fix it after i have
finally done the current job.

>> [part of code from load_and_analyze_sgf_file()]

> I might be missing something, but shouldn't some kind of play_move()
> call remain here?

it is used to generate a single move and quit. there's no need to play
the move on the board. however, we can add play_move() back to keep
generality (what word should i have used here? :)

>> -      if (!sgffile_open_file(outfile)) {
>> +      /*if (!sgffile_open_file(outfile)) {
>>         fprintf(stderr, "Error: could not open '%s'\n", gg_optarg);
>>         exit(EXIT_FAILURE);
>> -      }
>> +      }*/;

> Wrong way to comment out code.

yes, sorry, can't get rid of the habit. actually, i intended to remove
comments when diff'ing. fixed in the new patch i'll submit in a
minute.

Paul