Commit graph

361 commits

Author SHA1 Message Date
Andy2
821e4d9c39 add ACK to protocol, returned by device after gets connName on initial
connection.  Space is held for device on assumption ack will come
back, then given up if it fails.
2010-09-14 13:54:52 -07:00
Andy2
f8c1c016cc remove special-case states and handling of first connect; track number
of players sought in the cref itself, adding in constructor, and not
in HostRec instances.  This is cleanup with no behavior change.
2010-09-13 13:49:13 -07:00
Andy2
100ed8532f timestamp records on creation 2010-09-12 05:04:52 -07:00
Andy2
ded6e22180 do finding of open games using database rather than cref list, which
requires updating the number of players when new devices join.  Drop
requirement that cid be unique.
2010-09-12 04:49:03 -07:00
Andy2
663551fcb5 improve db, adding uniqueness contraints. Pass in connName so can
meet that constraint.  Figure out next cid using a query at startup.
2010-09-12 03:57:23 -07:00
Eric House
c1da64b96a instantiate and use the new dbmgr class 2010-09-11 19:45:04 -07:00
Eric House
87de53cf96 add new class to interface with postgres database in an experiment to
simplify things by using a db to survive a crash/reboot.  So far (with
next checkin) manages to update a db.  Next need to do queries instead
of searches through in-memory data.
2010-09-11 19:44:37 -07:00
Andy2
54d49d450a add back state and table entries to get rid of an unused cref. 2010-09-11 01:40:40 -07:00
Andy2
463e9c3ccf pass langCode in reconnect case too 2010-09-11 01:23:39 -07:00
Andy2
dc957e1a92 add match test on language code and number of players 2010-09-10 13:45:40 -07:00
Andy2
e60495ff93 read, pass, store and display langCode. Not yet used to route
connections to crefs.
2010-09-10 01:59:37 -07:00
Andy2
20312144dd tons of changes, still rough, snapshoted here: a gtk device vs device
game works to completion with both signing up as guests (no -s) with
one local and one remote player (identical commandlines.)  Not yet
tested: if any signs up as a host, reconnecting rather than
connecting, etc.  This is just a snapshot.
2010-09-10 01:30:40 -07:00
Andy2
f7c5ff619d turn off heartbeats 2010-09-10 01:24:03 -07:00
Eric House
89e7a506a4 reduce amount of logging 2010-09-08 20:28:14 -07:00
Eric House
90d505abea call accept() in main thread rather than in thread proc to fix
long-standing bug where main thread kept finding socket to be readable
and forking new threads to call accept() on it and then block forever.
2010-09-08 20:24:36 -07:00
Andy2
33d60452a3 fork a thread to handle proxy connects. Later may want to use thread
pool as for game threads.
2010-08-23 18:52:52 -07:00
Andy2
a34ccb424f device now passes list of "relayID"s, consisting of connName plus
device id, to relay, one for each stored game that's communicating via
the relay.  Relay parses out each relayID.  Next relay can use these
to look up whether messages are available and reply with that, and
device can put up a notification.
2010-08-22 12:16:57 -07:00
Andy2
56b6a425e2 proxy now writes one byte to relay and reads one byte reply. 2010-08-19 18:34:28 -07:00
Andy2
019da83d60 meant to check this in 2010-08-12 06:42:59 -07:00
Andy2
416b379781 add second class of socket to be listened on, meant for per-device
rather than per-game connection, and when adding connected sockets to
active set track the type.
2010-08-12 06:42:33 -07:00
eehouse@eehouse.org
3b7abc2a31 Merge branch 'android_branch' of ssh://xwords.git.sourceforge.net/gitroot/xwords/xwords into android_branch 2010-07-29 21:13:00 -07:00
Eric House
e857b01783 ignore logs and core files 2010-07-29 21:07:20 -07:00
Eric House
d7a3795355 add static decl 2010-07-29 21:04:33 -07:00
Eric House
d7d2a7523b kill cref when in bad state. 2010-07-29 19:46:52 -07:00
eehouse@eehouse.org
13ecb2f929 Merge branch 'android_branch' of ssh://xwords.git.sourceforge.net/gitroot/xwords/xwords into android_branch 2010-07-29 05:41:48 -07:00
Eric House
b170a04dec drop duplicate [re]connect packets before they enter the state
machine.  Fixes problems where crefs get stuck in the
XWS_CHK_ALLHERE_2 state.  Still should probably kill a cref where
there's no state table entry.
2010-07-28 22:21:01 -07:00
Eric House
f70b9e1d10 use constructor to save some code; no behavior change 2010-07-28 20:17:54 -07:00
eehouse@eehouse.org
958ec15a19 Merge branch 'android_branch' of ssh://xwords.git.sourceforge.net/gitroot/xwords/xwords into android_branch 2010-07-27 21:32:17 -07:00
Eric House
ad1be65236 Merge branch 'android_branch' of ssh://xwords.git.sourceforge.net/gitroot/xwords/xwords into android_branch 2010-07-27 20:47:42 -07:00
Eric House
b335151172 log when no event pushed while adding players -- this may be how state machine is getting screwed up 2010-07-27 20:47:04 -07:00
Andy2
51ffdce185 comment out pthread_create failure log; seeing it too often and it's
just for the web view.
2010-07-26 18:37:44 -07:00
Andy2
f05bfcb3d8 log when dropping message that would mess up cref player counts 2010-07-16 06:23:20 -07:00
Eric House
a52fc5a4fd don't call removeSocket when none was added 2010-07-15 18:21:32 -07:00
Eric House
8e5abb189b add mutex for reconns too; log socket and ip addr connecting on it 2010-07-15 18:14:44 -07:00
Eric House
5ea1f3b470 list newest crefs first 2010-07-15 18:13:01 -07:00
Eric House
78f965b979 replace asserts that have been firing on integrating reconn events
with dropping the reconn messages; move log of state table lookup failure
into cref so cref id can be included
2010-07-14 21:32:19 -07:00
eehouse@eehouse.org
f9d8df8cea don't track xwrelay.conf so local changes don't show as local mods 2010-07-14 06:29:50 -07:00
Andy2
41f8f574b1 use git describe for version info/stamp 2010-07-07 20:26:25 -07:00
Andy2
5365ea4f4e fix crash when empty string given on ctrl port 2010-07-07 18:59:48 -07:00
Eric House
6456165a92 replace svnversion with gitversion.sh 2010-06-05 07:27:14 -07:00
Eric House
8aaaa06bcb identify rev number as from git 2010-06-05 07:25:24 -07:00
ehouse
43181f5cd2 accept multiple cref ids and names on console 2010-06-03 03:00:35 +00:00
eehouse
8bb2fb9b09 track and display in web view the number of games actually joined
rather than the number attempted.
2010-05-28 02:31:27 +00:00
eehouse
c92d919934 log error from read 2010-05-26 02:25:58 +00:00
eehouse
7a279702c0 add sterrror to what's logged when pthread_create fails -- as is
happening often.
2010-05-20 04:28:07 +00:00
eehouse
3e83f50c84 use strcasecmp rather than strcmp on cookies (room names) to make
rooms case insensitive.
2010-05-01 02:49:19 +00:00
eehouse
4ed87fecb4 add once-per-day date line to logs since they can span weeks now. 2010-05-01 02:48:47 +00:00
eehouse
f4b64fe97d add option, off by default, to stall each message forward by some time
to better simulate network conditions when testing against local relay.
2010-03-28 16:09:07 +00:00
eehouse
e5f3384050 another .gitignore 2010-02-24 04:30:21 +00:00
ehouse
7ac98bfcfb Don't require a timeout waiting for a game to connect initially, and
make no timeout the default.
2009-12-27 18:17:24 +00:00
ehouse
a1be2fb8fc add ports relay is listening on to http display 2009-12-16 02:43:52 +00:00
ehouse
fc17df2a7e When guest connects, check if providing too many players before
reponding to connection request so that we just deny the request (with
new error code) rather than revoking it later.
2009-12-14 04:10:23 +00:00
ehouse
015b0d42d8 include title in head with summary readable in tab 2009-12-08 14:51:03 +00:00
ehouse
26469d734c Modify state machine and return new error codes to force host to
connect first, guests second, with clear error messages if the order's
wrong.  This seems to make it simpler for users to get a connection
right.  Code holding multiple games worth of hosts and guests in a
cref is gone.
2009-12-04 08:03:27 +00:00
ehouse
337c36dfd2 remove unused parameter 2009-12-04 07:59:07 +00:00
ehouse
7ad59e3e2b cleanup 2009-12-04 07:54:14 +00:00
ehouse
55478369e4 listen on only one port by default 2009-11-22 16:29:23 +00:00
ehouse
c1827562dc be consistent is use of variable vs real name 2009-11-10 14:37:42 +00:00
ehouse
2ecf415618 Remove on case where HandleEvents was being called recursively and
creating unexpected state transitions; add assert to catch any future
recursion.
2009-11-09 05:30:42 +00:00
ehouse
393cccb86c Add MSGONLY state to indicate cref being kept alive only to preserve
messages for a latter connection.
2009-11-09 05:29:06 +00:00
ehouse
aae12428da Add limited support for store-and-forward (where limitations include
that all devices must be connected initially and that it's all memory
based so a crash wipes stored messages.)  Accept messages for
forwarding when in the MISSING state, not just the ALLHERE state.
Store messages that can't be sent now, and send any that have
accumulated when a host reconnects.  When a cref loses its last
connection, keep it around unless it has no messages stored (as will
be the case when a game ends.)
2009-11-08 21:35:39 +00:00
ehouse
2d4ae3e44a When forwarding message, lookup cref by socket if cookieID == 0, as
can be the case when through to a non-all-connected game.
2009-11-08 21:32:36 +00:00
ehouse
8e0dad43c7 get rid of overlogged message 2009-11-08 21:25:43 +00:00
ehouse
ed52582ea4 get rid of states and code for checking numbers of players -- since
any number is allowed now.
2009-11-03 05:03:20 +00:00
ehouse
3bdfda6548 Make it possible for multiple games to connect using the same room
name.  All new connections are stored together, and after each
connection an attempt is made to build one complete game with a host
and however many guests.  All remaining devices are moved into a new
pending record in the same state, and the completed game is treated as
always.  Seems to work, though nearly 20% of linux instances are
failing to connect the relay run from the new test script samename.sh.
Need to figure out why.

Also added logging of seed and connname to comms.c since games
launched together can no longer be certain to connect on the relay.
This allows the test script to identify joined games from their logs
and detect success or failure.

This checkin changes the relay protocol, so relay and clients will
both need to be upgraded.
2009-11-02 01:01:47 +00:00
ehouse
11b586cd4b lookup hearbeat recipient by socket rather then cookieID since
cookieID isn't always correct until later
2009-11-02 00:45:54 +00:00
ehouse
7a71c4636b change varible names; remove instance variable than can be on the stack. 2009-11-01 01:38:03 +00:00
ehouse
0e43675abf Address problems matching [re]connections to games: first, go back to
setting connName when all in a game are present.  Second, have every
host include in connections a random number.  That number is made part
of the connName and in general used to test whether a host belongs in
a particular game.  Add this "seed" to web interface.  Means new
versions for relay protocol and game stream format.  Latter is handled
correctly so older games can be opened.
2009-09-26 14:37:49 +00:00
ehouse
030ae52a4b Store start time as time_t rather than string. 2009-09-26 14:05:51 +00:00
ehouse
b219e8b279 Send notification when device disconnnects without closing socket (in
addition to the existing one sent when a socket is closed).
2009-09-20 18:47:47 +00:00
ehouse
ffacbcb36a Move "For" column left. 2009-09-19 04:59:19 +00:00
ehouse
45bdccd4e8 Format time more compactly 2009-09-18 03:50:01 +00:00
ehouse
a9e8f5b493 shorten column headings; don't assign cid until game is full. 2009-09-17 03:42:12 +00:00
ehouse
d5d20293ce print log timestamps with padding 0s for better sorting 2009-09-15 12:54:40 +00:00
ehouse
135048a3cf Extend name change to relay's web interface 2009-09-14 03:31:54 +00:00
ehouse
18e9301e87 rename Cookie to Invite, a change that may change. Rename everywhere
in clients but not in relay where cookie is too pervasive.  (There are
no code changes in this checkin.)
2009-09-14 03:11:11 +00:00
ehouse
6603949b0e Fix bug where device quitting one game to start another with same
cookie winds up connecting to the old game -- by flagging a cref as
full and no longer accepting connections without connNames even when
in the MISSING state.  Required a protocol changes so that devices get
their connNames as early as possible.
2009-09-14 01:55:03 +00:00
ehouse
7eeede45c3 exit if unable to open socket 2009-09-13 15:27:19 +00:00
ehouse
7ca629a2f5 don't log socket closing as an error 2009-09-04 11:02:09 +00:00
ehouse
95a70862be Lock cref before calling method that assumes lock. 2009-08-30 15:18:40 +00:00
ehouse
2a569df97f For privacy, don't show host IP unless connection is local. 2009-08-30 15:16:06 +00:00
ehouse
b15665e041 serve http view from cached data and make cache refresh rate
configurable.  Should make it harder to attack the relay on http port.
2009-08-21 12:32:57 +00:00
ehouse
dbf9daf71b Fix problems reconnecting hosts where some but not all received the
ALLHERE message and connName: change relay protocol so cookie is
included in RECONNECT message, and hostIDs are not assigned until
ALLHERE, and change host-to-game matching to use connName first but
fall back to cookie.  This fixes nearly all cases failing to reconnect
after relay goes down.
2009-08-21 12:00:09 +00:00
ehouse
493d3c226f log when crash 2009-08-20 12:12:43 +00:00
ehouse
53133839a4 don't log list of sockets by default 2009-08-20 03:50:40 +00:00
ehouse
e96becb05f variable name change only (for clarity) 2009-08-13 12:59:55 +00:00
ehouse
47d6d53419 make log file location configurable; remember start time as used for
permenent IDs on spawn rather than overall boot to avoid duplicating
cids.
2009-07-31 13:02:05 +00:00
ehouse
7077a11af6 pass .css file 2009-07-31 12:58:47 +00:00
ehouse
acbf8180ef add missing lock(); track and display uptime of daemon child as well as parent 2009-07-31 12:56:04 +00:00
ehouse
80ccdd6d04 spelling... 2009-07-31 12:47:58 +00:00
ehouse
f537169db5 Assign 0 not NULL to pthread variables. 2009-07-31 12:43:04 +00:00
ehouse
bcc887c0b1 Need to remove before inserting. Todo: is there a replace? 2009-07-31 12:39:47 +00:00
ehouse
503db9cc6f up copyright 2009-07-30 12:54:17 +00:00
ehouse
f35f2fe696 add css file path 2009-07-30 12:52:05 +00:00
ehouse
ba97f096dc make css file for http interface configurable; fix bug where
connections arriving too close together would wind up with different
crefs (using new mutex).
2009-07-30 12:49:21 +00:00
ehouse
631bf4b179 write to named file for use on ISP from cron 2009-07-30 12:46:40 +00:00
ehouse
2641ab6430 bug fix: mark after locking so will unlock in destructor 2009-07-30 12:39:45 +00:00
ehouse
4a5d74a0af rename state to take up less space in browser; add number of respawns
to status display.
2009-07-29 04:25:21 +00:00
ehouse
8be31357ad track per-cref lifetime and display it via http 2009-07-28 05:41:15 +00:00
ehouse
8284799abe Fix crashes resulting from deleting cref instances in one thread while
another wanted to operate on them.  The root problem is that you can't
dispose of a mutex while somebody's blocking on it.  So now the
locking mutexes live inside the cref class.  When the lock owner
realizes the cref needs to die, it sets a flag and it's moved to a
recycled list.  A thread blocking on the mutex will then get it, but
checks the flag and releases it immediately if it's being recycled.
(Also improve the http interface a bit.)  With these changes I've run
31K (and counting) games against the relay without a crash or deadlock
(using sim_real.sh.)  The main problem that remains is that sometimes
two games using the same cookie wind up with two crefs (and so never
connect.)
2009-07-28 05:15:26 +00:00
ehouse
0d03ed71bb cleanup 2009-07-28 05:09:24 +00:00
ehouse
6b1cf247bf Don't alloc storage for fds and log buffer each time through poll
loop; instead reuse and realloc when needs to grow.
2009-07-28 05:08:15 +00:00
ehouse
8c7c457a54 shorten state names so they take up less webpage space. 2009-07-28 05:06:25 +00:00
ehouse
3b62e04d07 don't use -lmcheck. It causes a lot of crashes in places where it
certainly looks as if nothing is wrong.
2009-07-28 05:00:18 +00:00
ehouse
dddd5e1dd8 clean up logging. I got some cores that suggest overflow. This fixes
that if indeed it was happening.  Problems with sockets disappearing
while in poll may remain.
2009-07-16 04:43:33 +00:00
ehouse
0a0d4804dd store keys and values as const char* instead of string since the
string code was involved in a lot of crashes.
2009-07-16 04:23:02 +00:00
ehouse
8a0d963c95 include abbreviated uptime in stat row 2009-07-16 02:56:25 +00:00
ehouse
c5e73fd725 Add to http display number of games currently being hosted. 2009-07-13 03:01:22 +00:00
ehouse
582eb88f9a int->bool 2009-07-13 02:58:16 +00:00
ehouse
854b3d9349 use synchronized method to add map entries 2009-07-06 02:02:59 +00:00
ehouse
2ef16d44ca Include number of games played in html output; don't report closed
sockets found while reporting closed sockets (to avoid deadlock);
remove sockets from crefmgr's map when closing them so new connections
using same (re-used) socket aren't treated as belonging to open games.
2009-07-06 01:50:51 +00:00
ehouse
a680a6e5fe Always close sockets we're done with. Otherwise they leak -- don't get reused. 2009-07-06 01:32:30 +00:00
ehouse
37dc9b7b88 running under valgrind I need to handle SIGPIPE. 2009-07-06 01:31:16 +00:00
ehouse
f4db39f490 synchronize reads of std::map to fix crash due to memory corruption.
Apparently containers aren't thread-safe even for reads.
2009-07-03 14:54:21 +00:00
ehouse
9b29e45ea4 plug valgrind-found leak; prefer localtime_r. 2009-06-29 12:58:30 +00:00
ehouse
4a308f4f8e Don't use unitialized variables (thanks valgrind :-) 2009-04-05 18:47:45 +00:00
ehouse
fbe95478ba rename now() as uptime(); drop NTHREADS to 1 for now; include time of
server boot in connName and keep counter in memory rather than a file.
2009-03-10 12:52:17 +00:00
ehouse
1b49a26736 do loglevel and ports via the general-purpose config mech. No special
cases remain.
2009-03-05 13:49:01 +00:00
ehouse
3469dc7bf3 Make refresh interval configurable. 2009-03-02 05:29:06 +00:00
ehouse
6b3265ac47 Change config class from having a method for every configurable value
to returing values for string keys so adding a new config doesn't mean
modifying that class.  Then add new config for addr passed to bind so
relay can be run to accept http connections from remote machines.
2009-03-02 01:50:14 +00:00
ehouse
9a8abb770d in web console list hosts present rather than just the count 2009-02-28 19:35:32 +00:00
ehouse
3bf116c936 Add an http interface (read-only web console). 2009-02-28 16:15:59 +00:00
ehouse
03c806e961 add tarball target to help 2009-02-21 17:15:09 +00:00
ehouse
644dc3b5ad detect duplicate initial connect packets by checking what socket they're on. 2009-02-07 18:17:17 +00:00
ehouse
53b27c96c9 Helps to compile platforms too before checking in... 2009-02-01 16:20:18 +00:00
ehouse
1ac4ef3bc6 Attempt to detect connections that are duplicates by checking if that
socket is already in a connection.  This doesn't seem to always work,
and I may be leaking cref instances when it does, so needs debugging
still.
2009-02-01 16:00:20 +00:00
ehouse
6ba2032187 clean up event logging 2009-02-01 15:55:45 +00:00
ehouse
552f6350f9 tweaks to stop/start script to run on an ISP's server better 2008-12-31 04:26:37 +00:00
ehouse
c53fe0857c Don't parse the same conf file twice in a row. 2008-12-31 04:21:03 +00:00
ehouse
1c3be799b7 start counter for uptime at launch 2008-12-31 04:19:23 +00:00
ehouse
9fc325a334 include new file in tarball 2008-12-31 04:16:55 +00:00
ehouse
32121c5634 include svnversion output in tarball 2008-12-30 05:17:00 +00:00
ehouse
ea6fff8f28 replace int with bool where appropriate; fix to compile with newer g++ 2008-12-30 05:13:30 +00:00
ehouse
4c191e0755 Don't specify g++ version; add some consts to fix warnings with newer g++. 2008-05-21 03:49:11 +00:00
ehouse
4beb818de3 use sigaction rather than signal 2008-03-15 14:04:02 +00:00
ehouse
acc7b5be52 Don't respawn if child died in less than a second. 2007-12-05 06:30:45 +00:00
ehouse
da93ed5b31 Don't use uninitialized values as port numbers; use match for set as well as get. 2007-12-03 00:45:09 +00:00
ehouse
25c4368231 Start adding ability to listen on multiple sockets and to dynamically
change the set listened on.  There's still some debugging to do but
nothing that worked before is broken.  Also begin to accept unique
prefixes (e.g. g for get) for commands and attributes on the control
port.  Note that relay-related code in comms seems broken now, but is
without this checkin.
2007-12-01 15:00:30 +00:00
ehouse
022317b47a Exit on unexpected params; better testing when launching; use bash to create pid file. 2007-11-22 04:59:11 +00:00
ehouse
cb1222c0f9 Make cooperation in a heartbeat scheme a compile-time option. This is
meant to allow comms to take over heartbeats.
2007-11-18 23:38:56 +00:00
ehouse
8dbfe4ed79 fix new compiler warnings; daemonize, and add handler to kill child if
parent dies so that script can kill based on pid file.
2007-11-13 03:53:10 +00:00
ehouse
74a30ffe82 init.d style script for starting and stopping relay 2007-11-13 03:50:41 +00:00
ehouse
0291ec1e75 Turn on self-spawning: main loop should never crash but exists only to
respawn child when it does.  Add crash command to test this.  Add
ability to set logging level from ctrl port.
2007-11-10 05:41:49 +00:00
ehouse
9f23c4b4d9 Replace ints with bools; add rev command; fix bug by reiniting args with each
command so they don't hold previous values.
2007-02-07 13:22:40 +00:00
ehouse
a8d3dbda66 Print svn rev number as part of usage() 2007-02-07 11:56:30 +00:00
ehouse
767b9f2908 initialize mutex before using; fixes a hang since could never aquire
lock.  (This bug has always been there, but didn't matter before;
maybe pthreads changed.)  Comment out self-spawning stuff until can
fix df-transfer problems.  Other minor changes around logging.
2007-02-06 05:52:22 +00:00
ehouse
5fbc89ae76 add tarball target for exporting enough code to build relay. 2007-02-04 01:22:31 +00:00
ehouse
4c2259400a include result of strerror in messages logging errno 2007-02-03 21:35:07 +00:00
ehouse
d123d6c352 Make self-restarting: fork child and wait (to deal with asserts). 2006-10-11 02:06:20 +00:00
ehouse
3da83805cc handle cmdline opts in a less-fancy way that works on freebsd too. 2006-03-22 04:33:34 +00:00