Miroirs/xwords
1
0
Fork 0
mirror of git://xwords.git.sourceforge.net/gitroot/xwords/xwords synced 2025-01-04 23:02:02 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
5097 commits 21 branches 243 tags 33 MiB
Commit graph

420 commits

Author SHA1 Message Date
Andy2
b506673946 use only one thread until race condition's fixed 2011-01-19 18:36:50 -08:00
Andy2
3ae62e4704 use an index rather than a ptr to track fd array (so a parallel array 
can be added)
 2011-01-19 18:36:24 -08:00
Eric House
720f78d18a get rid of ivars used only for http UI 2011-01-04 21:43:10 -08:00
Eric House
11d7de5f0d display the number of devices (pids) running 2010-12-21 20:52:01 -08:00
Eric House
b65338b643 script for use on relay host 2010-12-07 20:46:02 -08:00
Eric House
d1244da284 script encapsulating sql query that's useful to run (with 'watch') along with discon_ok2.sh 2010-12-02 22:29:36 -08:00
Eric House
ae9dbd1e5c create mtimes array rather than mtimes timestamp 2010-12-01 21:09:25 -08:00
Eric House
4bd5553596 track mtime per host rather than per game -- make it an array in the db. 2010-12-01 21:08:22 -08:00
Andy2
1720fbbd50 fix parsing of relay ids by replacing separating \n with \0. 2010-11-23 17:46:16 -08:00
Andy2
daecc807c1 don't init nPerDevice, and deal with it not being inited. 2010-11-22 20:43:37 -08:00
Eric House
b91187d6a5 test result when parsing; drop if didn't parse 2010-11-16 20:30:43 -08:00
Eric House
4c6c3c1ded when marking a device in a dead game, negate its nperdevice entry rather than always using -1. 2010-11-16 20:28:53 -08:00
Andy2
635ae053b6 cut down on printing; don't assert-fail when relay's crashed. 2010-11-15 20:40:59 -08:00
Andy2
2c5c7a0dd1 init m_nRoomsFilled -- not sure when this http view feature broke 2010-11-15 20:08:59 -08:00
Andy2
ea292c7e2f exclude dead games when looking for open public rooms 2010-11-15 19:01:11 -08:00
Andy2
c6bd2125a4 java uses '/' to delimit device id, so we'll make that standard. 2010-11-15 18:27:36 -08:00
Andy2
c36679fe13 init isDead to false. Fixes prolem where if DB entry's missing the 
value is uninitialized and so device gets warned that other device
resigned even if none did.
 2010-11-15 18:21:40 -08:00
Andy2
d4e810c9cb delimiter changed 2010-11-15 06:51:05 -08:00
Andy2
ce5738a4e1 write the count 2010-11-15 06:50:09 -08:00
Andy2
edf0c40904 track isDead when creating SafeCref for devgone 2010-11-15 06:49:41 -08:00
Andy2
2c9f2572a8 use GameFull() 2010-11-15 06:48:57 -08:00
Andy2
feca7345dc add GameFull -- database, not in-memory state, knows if game is full. 2010-11-15 06:48:36 -08:00
eehouse@eehouse.org
14579c6e30 Merge branch 'relay_proxy' of ssh://xwords.git.sourceforge.net/gitroot/xwords/xwords into relay_proxy 2010-11-14 19:40:35 -08:00
Eric House
528ae0c986 allow more than one device death per message 2010-11-14 15:19:38 -08:00
Eric House
a8ac30b180 add -d option to test dev-gone on relay 2010-11-14 15:18:49 -08:00
Andy2
43896ff187 remove unused method and variable 2010-11-12 18:25:15 -08:00
Andy2
0787a06eb5 accept new message from device announcing the death of a game 
(including seed so it's harder to spoof); respond to that by setting a
DEAD column in the db and flagging the device as gone.  Notify any
connected device of the fact.  Refuse to accept new connections to
that game.  As already-connected devices reconnect, allow them to do
so but send a new status message that their game is dead.  Not heavily
tested yet.
 2010-11-11 06:40:50 -08:00
Andy2
3d59bc7e39 add new DEAD column to track when a game's been deleted 2010-11-11 06:29:21 -08:00
Andy2
04e1b0c22b new constants, include new status event type, for relay to accept 
notification of deleted game and communicate that to others in the
game as they connect.
 2010-11-11 06:26:59 -08:00
eehouse@eehouse.org
7929167129 Merge branch 'android_branch' into relay_proxy 2010-11-10 07:06:10 -08:00
Andy2
c8d7ca40a5 use \1 instead of \n as delimiter -- in case somebody passes as string 
containing a \n.
 2010-11-09 18:02:22 -08:00
Andy2
0f111ab363 use PQexecParams instead of PQexec for queries involving strings from 
users (i.e. room names) to avoid crashes when the contain apostrophes
-- and in general to be safer.
 2010-11-09 17:27:33 -08:00
Eric House
2a2f4d4395 been a while since cvs... 2010-11-09 05:53:49 -08:00
Andy2
d2782af487 do case-insensitive comparison of room names 2010-11-06 18:52:10 -07:00
Andy2
d1fcdb69da double max room name len since the old value was too small -- only 
filled half a text field on android
 2010-11-03 18:38:43 -07:00
Andy2
c7febf0d10 nuke unused constant 2010-11-03 18:30:05 -07:00
Andy2
0148a93f3e add a constraint to the stored messages table to avoid dupes. Then 
remove the artificial limit on number of stored messages.  This seems
to fix problems where lots of chats in a row clog up the relay so that
messages never get flowing again.  Works more cleanly than
artificially ACKing.
 2010-11-03 18:14:19 -07:00
Andy2
633d4c95a7 fix buffer overflow sql error with large messages; combine logged messages. 2010-11-03 18:12:26 -07:00
Andy2
b237277544 put back enum just removed; it's screwing up existing code having all 
the followers change.
 2010-10-25 18:06:48 -07:00
Andy2
e909e4b343 remove redundant message; advance state to ALLCONNECTED on receipt of 
RECONN or CONN that lists the game as full.  This means we get the
all-green icon (or the A in the gtk case) when in a consummated game
rather than only when all devices are actually connected.
 2010-10-23 20:29:24 -07:00
Eric House
8a038dc748 Merge branch 'relay_proxy' of ssh://xwords.git.sf.net/gitroot/xwords/xwords into relay_proxy 2010-10-23 19:45:03 -07:00
Andy2
7cf4535b4c new dbmgr API returns number of players in game; use that rather than 
currently connected devices to populate [RE]CONNECT message
 2010-10-22 22:34:43 -07:00
Eric House
85d29cf560 if can't send allcond to a device because it's not connected queue the 
message to be sent next time it does.  And since the cookieID will
change and should be there already anyway, remove it.  (Should remove
the hostID too for the same reason.)
 2010-10-22 19:02:52 -07:00
Eric House
9ba2d2f177 default msgs entry timestamps to 'now' 2010-10-18 20:19:09 -07:00
Andy2
2bfb26cded sleep (in thread) before closing proxy socket. This works around 
problem where, on PRX_HAS_MSGS path only, device reading socket gets
EOF early though logging here confirms additional bytes have been
written.  Gross but effective.
 2010-10-17 06:38:28 -07:00
Eric House
34df149046 include age of room with info returned for open public rooms 2010-10-14 22:16:21 -07:00
Eric House
e5c7f476ba get rid of microseconds in timestamps 2010-10-14 22:15:41 -07:00
Eric House
0d1542ca9d tweaks to logging 2010-10-12 05:18:44 -07:00
relay
680b2f31b3 use pid to avoid permission problems for logfile in /tmp 2010-10-07 19:48:38 -07:00
Eric House
f2ad6d92b4 use full paths so pidof doesn't pick up instances of xwrelay launched elsewhere 2010-10-07 19:41:02 -07:00
Eric House
86d73392d5 track bytes sent in db; use defaults to init ctime 2010-10-06 21:45:05 -07:00
Eric House
a444786a28 reduce logging 2010-10-05 20:57:19 -07:00
Eric House
2e844d798f replace single db connection protected by a mutex with one connection per thread kept in thread-local storage. 2010-10-04 20:04:14 -07:00
Eric House
d9ca6f5e11 block TERM and INT from all but the main thread. (Pretty much untested; doesn't fix hang-on-SIGINT.) 2010-10-04 20:03:19 -07:00
Eric House
1ff4d67bc0 block TERM and INT from all but the main thread. (Pretty much untested; doesn't fix hang-on-SIGINT.) 2010-10-04 20:03:00 -07:00
Eric House
43bf3cd50f improve log statement 2010-10-04 20:01:00 -07:00
Andy2
89f967a016 pass remaining reconnect params in case connName's missing from db. 
Should only matter if the relay's db's been deleted.
 2010-10-04 18:39:26 -07:00
Eric House
d35bc7d748 refuse to store more than four messages at a time for a single 
connName/hid pair.  This works around a problem where a device,
continuing to connect and send all pending messages for another device
that is not connecting, winds up with hundreds of identical messages
stored.
 2010-10-03 18:18:15 -07:00
Eric House
1f87379cd0 change default port 2010-09-29 07:37:26 -07:00
Eric House
1b93654f7e use stored procedure sum_array instead of C macro 2010-09-29 06:21:08 -07:00
eehouse@eehouse.org
d0797af6e5 Merge branch 'relay_proxy' of ssh://xwords.git.sourceforge.net/gitroot/xwords/xwords into relay_proxy 2010-09-28 21:11:17 -07:00
Andy2
978fa010f1 commented-out sleep positioned to give download progress time to come up. 2010-09-28 18:35:33 -07:00
Eric House
bc65377cfd track device seeds in db. 2010-09-28 18:06:50 -07:00
Andy2
1de029f3f0 skip public rooms if full. (Should do this in the query....) 2010-09-28 06:42:01 -07:00
Eric House
8f77886f72 make column name shorter 2010-09-24 20:18:20 -07:00
Andy2
f7bad05390 send msg count query and print results 2010-09-24 05:51:42 -07:00
Andy2
b0c6e6da9c add db method to count messages for connName/hid pairs and handle 
query using it.
 2010-09-24 05:50:02 -07:00
Andy2
c3e1b243b6 add timestamp to msgs db and set it on adding new records. I will 
likely need to add periodic culling on the server for games that will
never fetch their messages.
 2010-09-24 05:46:30 -07:00
Andy2
c44324a32a log, for now as I've seen assertions failures suggesting a race, the 
set of checked-out sockets at insert and remove time.
 2010-09-23 06:34:18 -07:00
Andy2
40e1d29243 use new methods to store undeliverable messages in a db rather than in 
memory.  Confirmed that devices can hookup without ever being
connected at the same time and that the messages survive a relay
reboot.  No further testing yet.
 2010-09-23 06:32:57 -07:00
Andy2
74764a493b create the new db 2010-09-23 06:31:42 -07:00
Andy2
eee41377f4 new methods for storing/retrieving messages in new db. 2010-09-23 06:31:29 -07:00
Andy2
9a035c66ff add assertion 2010-09-23 06:28:13 -07:00
Andy2
85b4f3f075 move hostid assignment and tracking into db so it survives reboots and 
recycling of crefs between when devices in a game connect.  This plus
movement of messages into the DB will re-enable games played without
the two devices ever being connected at the same time.
 2010-09-23 05:33:27 -07:00
Andy2
1d0c207521 add test command for proxy feature to fetch for connNames whether 
there are messages pending for them.  Not handled yet on relay.
 2010-09-21 18:28:31 -07:00
Andy2
e06baf566f get public rooms from the db rather than stubbing it out. 2010-09-21 06:54:47 -07:00
Andy2
be23e06b60 use read_packet to replace duplicate code 2010-09-21 06:38:43 -07:00
Andy2
e7b5923883 finish first pass at sending open public room names -- but with actual 
list harvesting stubbed out.
 2010-09-21 06:27:31 -07:00
Andy2
2ec67c8ed1 finish rq to work with current format; add to makefile 2010-09-21 06:26:49 -07:00
Eric House
4ad735124c add logging and assertions. There's a race condition in cref 
recycling and may be one in tpool socket checking-out.
 2010-09-20 22:11:13 -07:00
Eric House
047f41b516 send cookieid as part of [RE]CONNECT_RESPONSE in addition to ALL_COND (where it's now redundant). 2010-09-20 22:06:35 -07:00
Andy2
a549ade1fa a bit more work on rc 2010-09-20 18:31:17 -07:00
Andy2
15cd3af62a just a comment 2010-09-20 05:35:29 -07:00
Andy2
e1c99f53cd work on detecting duplicate packets and devices that have reconnected 
on a different socket; print all present seeds for debugging.
 2010-09-20 05:35:19 -07:00
Andy2
ac8891016f work on detecting duplicate packets and devices that have reconnected 
on a different socket; print all present seeds for debugging.
 2010-09-20 05:35:09 -07:00
Andy2
430b033ae0 a bit of progress on cmdline tool 2010-09-20 04:59:09 -07:00
Andy2
df1ec1628a replace killSocket(), which jumped right into crefmgr and could get 
ahead of processing data arrived on the same socket, with EnqueueKill
that adds to same queue from which data's taken.  So if device dies
immediately after sending data there won't be a race between closing
the cref (if this is the last open socket) and handling the data.  I'm
still dying with assert fails when running 100 games at once, but much
less frequently
 2010-09-18 08:47:56 -07:00
Andy2
a4913596db remove logging 2010-09-18 08:42:15 -07:00
Andy2
c5d113a8da remove _locked from name of method since it's not called when holding a lock. 2010-09-18 07:46:10 -07:00
Andy2
53b3d7c32f enque sockets in new object that contains type so later other events 
(e.g. socket closes) can also be queued; track what sockets are
currently being handled so no two threads are ever processing events
on same socket.  This makes running 'NGAMES=50
./scripts/discon_ok2.sh' crash very infrequently rather than every
time, but the race condition in crefmgr remains.  Queuing socket
closes should help.
 2010-09-18 07:44:14 -07:00
Andy2
9c8889b429 get rid of mutex guarding reconnect. There shouldn't be any need 
since game slots are already assigned, and it's slowing down
reconnects in their race with kills.
 2010-09-18 05:07:23 -07:00
Eric House
bb1e46fa81 Merge branch 'relay_proxy' of ssh://xwords.git.sourceforge.net/gitroot/xwords/xwords into relay_proxy 2010-09-16 20:13:54 -07:00
Eric House
d0e707a48e get started with test/cmdline app to query relay's new proxy interface. 2010-09-16 20:13:24 -07:00
Andy2
c5eef29bcc some suspicious crashes and subsequent reading say that multiple 
threads can't share the same db connection.  So synchronize all
queries.  Multiple threads *can* access the db concurrently as long as
each has its own connection, so I could add connections to the tpool
threads.  But this will probably be performant enough for the first
10K simultaneous users. :-)
 2010-09-16 20:08:32 -07:00
Andy2
1490b2def9 simplify state table, getting rid of a bunch of states: we never leave 
ALLCOND until it's time to kill the cref.  When one device connects
and leaves the cref is recycled, so when a new device joins and is
assigned the same game and a new cref is initialized the number of
players already registered must be set from the db entry; do that.
 2010-09-16 18:59:56 -07:00
Andy2
50c387d75e remove players from db as well as runtime when ack times out 2010-09-16 06:49:35 -07:00
Andy2
25c1824c45 fix broken assertion and add missing state transitions to fix crashes. 2010-09-16 05:39:25 -07:00
Andy2
2b702b48d6 need to clear CID of record in DB when killing corresponding in-memory 
cref or reconnect will fail.
 2010-09-16 04:47:17 -07:00
Andy2
d1a9c3962a rename cookie->room in db; reposition columns. 2010-09-16 03:30:24 -07:00
Andy2
b25579b893 greatly simplify state machine for connections. Now (with DEVACK=20 
on conf file) newrelay.sh will complete running 50 4-device gtk games.
Haven't tested reconnect case which I almost certianly broke.
 2010-09-16 01:53:43 -07:00