When I started working on pokersource, Loic told me to always make sure trunk passes the tests. Strangely, I discovered tonight that "someone" committed r4046 and didn't make sure the tests in test-pokeravatar.py.in still pass correctly. ;) And here I thought I'd broken something when suddenly 'make check' stopped working tonight.
Ok, enough giving Loic a hard time. ;) In fact, I believe that the fix done in r4046 fixes the bug I described in test-pokeravatar.py.in back in February: # Note that the 'start' packet test that follows is different from all # the other packets in this group. The point is that if a # game.isEndorNull() is true at line 411 of pokeravatar.py, the message # sent back -- regardless of the serial -- assumes that the player with # the right serial has requested it but cannot because a game is already # going. I am not sure this behavior by pokeravatar.handlePacketLogic() # is entirely correct, but I am leaving it, and have instead rewritten # this test to look for the response given by handlePacketLogic when # isEndorNull() is true rather than the "someoneElse" response. I had been waiting for someone to fix this, but didn't feel I knew the code base well enough to fix it at the time. The tests make a lot more sense now. ;) I have redone some of the tests on PACKET_POKER_START to deal with the expected behavior, and they pass now on trunk. However, we've lost full coverage on pokeravatar.py because of it. I can look into this more after some sleep. Looking forward to IRC discussion on this tomorrow, -- -- bkuhn _______________________________________________ Pokersource-users mailing list [email protected] https://mail.gna.org/listinfo/pokersource-users
