IRC log for #bzflag on 20170701

00:21.42*** join/#bzflag infobot (~infobot@rikers.org)
00:21.42*** topic/#bzflag is https://BZFlag.org || http://ohloh.net/p/bzflag || https://bzflag.org/help || Channel Logs: http://infobot.rikers.org/%23bzflag/ || 2.4.10 released! https://forums.bzflag.org/viewtopic.php?f=8&t=19401
00:21.42*** mode/#bzflag [+v infobot] by ChanServ
01:09.20*** join/#bzflag alfa1 (~alfa1@host64.190-30-52.telecom.net.ar)
01:09.38*** part/#bzflag alfa1 (~alfa1@host64.190-30-52.telecom.net.ar)
03:57.03allejoJeffM: https://github.com/BZFlag-Dev/bzflag/blob/2.4/src/bzfs/bzfsAPI.cxx#L1089-L1096 why isn't delete() being called before returning NULL for the `*playerRecord` that was already new'd
03:57.35JeffMshould call delete
03:57.50JeffMprobalby shoudl break it into two parts
03:58.40JeffM<PROTECTED>
03:58.41JeffM<PROTECTED>
03:58.41JeffM<PROTECTED>
03:58.41JeffM<PROTECTED>
03:58.41JeffM<PROTECTED>
03:58.41JeffM<PROTECTED>
03:58.41JeffM<PROTECTED>
03:58.42JeffM<PROTECTED>
03:58.50JeffMthat way you don't need a delete
03:59.24JeffMit's just to handle error cases, so may as well not allocate if the player is bad
03:59.46allejowhen would `new bz_PlayerRecordV2` return a null pointer?
04:00.29JeffMwhen the system has no memory
04:00.36allejoah
04:00.37JeffMaka never in a modern OS
04:00.50JeffMso if it returns null you don't need to delete it :)
04:00.55JeffMbecause the allocation failed
04:01.10JeffMmaybe on a tiny computer with out virutal memory
04:01.13JeffMlike pi or something
04:01.24JeffMso it's a good thing to handle if you want to care about those peoples
04:01.56JeffMbut at that point you are basicly screwed since you can't allocate anymore data at all... hope nobody joins
04:02.18allejohehe
04:02.43allejowonder how little memory would necessary for that
04:02.56JeffMmaybe like an 8-10 meg system
04:03.03JeffMor an OS with a fixed memory allocation
04:03.08JeffMlike old MacOS
04:06.42JeffMbut that's a good catch. it will leak a player record if someone asks for a player that doesn't exist
04:06.46JeffMso you should make the change.
04:07.34allejochange coming up in a bit
04:09.55*** join/#bzflag BZnotify (~BZnotify@192.30.252.34)
04:09.55BZnotify[13bzflag] 15allejo pushed 1 new commit to 062.4: 02https://git.io/vQRo1
04:09.55BZnotify13bzflag/062.4 146dfc1ae 15allejo: Fix possible leak if non-existent player is searched for
04:09.55*** part/#bzflag BZnotify (~BZnotify@192.30.252.34)
04:14.49JeffMprobably never happened, because everyone gets player IDs from events and lists and stuff
04:14.55JeffMbut still a good catch
04:16.36allejo:)
04:17.30JeffMprobalby not the only place that can happen too
04:18.05JeffMcould run visual studio's code analysis on it all
04:18.36allejooh yea, probably happens else where as well
04:20.17JeffMlets see what it comes up with :)
04:20.22JeffMtime to use all the cores!
04:21.05JeffMfinaly.. none of them are parked
04:21.52allejohow many cores?
04:22.08JeffM8, 16 thread
04:22.31JeffMsadly it's only using 11% of the total CPU, but at least it's doing a lot of things at the same time :)
04:22.49JeffMit's why I named the machine octa
04:24.16JeffMchrist on a stick chrome.. why do you need 1.2 gigs for 3 friggin tabs?
04:25.10JeffMat that rate I could only open like 38 tabs......
04:25.33allejoapt name
04:26.15allejoI abuse tabs a lot... which is why this comes in handy: https://chrome.google.com/webstore/detail/the-great-suspender/klbibkeccnjlkjkiokjodocebajanakg?hl=en
04:26.32JeffMI think one of the pages was just being a javascript hog
04:26.45JeffMthis machine has plenty of ram, 32 gigs
04:27.06JeffMfigured it should last at least another 5 years
04:28.34JeffMwant the list of what it found in bzfs?
04:28.57allejoyou'll probably get kicked from IRC for spam if you paste here :p
04:29.17JeffMhttps://www.irccloud.com/pastebin/KnN9syhp/
04:29.25JeffMIRCCloud !!!
04:29.33JeffMit won't let me paste big stuff, it autopastebins :)(
04:30.07allejohuh. the list isn't as big as i thought it would be
04:30.28allejoand that's a pretty neat feature actually from irccloud
04:30.49JeffMI only did the bzfs proejct, so not common, date, game, net, or obstacle
04:31.53JeffMI don't know how complete it is, but it looks like valid results
04:32.16JeffMexcept the link stuff at the end, that's probalby just windows specific stuff that has to do with the way the deps are built.
04:34.28JeffMit shows nice little note about what it found when you double click on it in VS.
04:34.38JeffMand highlights the lines involved
04:35.07JeffMlike the 4th item, it shows where p is checked for null, but also shows how later outside the check it can still be derefed.
04:36.02allejowhere's p being deferenced in that?
04:36.19JeffMline 414
04:36.31JeffMin the loggin call, that's outside the null test
04:38.31JeffMsnprintf(kickmessage, MessageLen, "%s kicked by %s, reason: %s",
04:38.31JeffMp->player.getCallSign(), playerData->player.getCallSign(),
04:38.31JeffMkickEvent.reason.c_str());
04:39.17JeffMthere should be a p == nullptr check after p is gotten on line 391, but the only place it's checkfed for is if the playerdata is operator
04:39.56JeffMnow the ID comes from an event, so it's probalby always a valid player ID and not ever null, but the tool doesn't know that
04:40.41JeffMprobalby should just return false if p is null after the lookup
04:43.06allejobut aren't you already checking for p not being null right after the isOperator check?
04:43.18JeffMwhat if the player isn't an operator?
04:43.29JeffMthen it never checks p as null and continues on
04:44.01JeffMthat IsOperaator CAN return false
04:44.04allejoOH
04:44.28JeffMyeah
04:44.35allejook, our line numbers are off and i can't read apparently lol
04:44.40JeffMheh
04:44.50allejohttps://github.com/BZFlag-Dev/bzflag/blob/2.4/src/bzfs/BanCommands.cxx#L414
04:44.50JeffMI did reformat the entire file :)
04:44.59allejoand https://github.com/BZFlag-Dev/bzflag/blob/2.4/src/bzfs/BanCommands.cxx#L391
04:45.02allejo...
04:45.03allejothat's why
04:45.45JeffMyeah, look at line 428
04:46.08JeffMthat uses p, but is outside that isOperatator check
04:49.27allejoright right
04:49.42allejoi didn't see that reference there, was looking at line 413
09:12.54*** join/#bzflag I_Died_Once (~I_Died_On@unaffiliated/idiedonce/x-1828535)
10:05.47*** join/#bzflag Notify (~notify@104.225.5.10)
15:02.41*** join/#bzflag [dmp] (~dennis@unaffiliated/dmp/x-546784)
15:06.38*** join/#bzflag [dmp] (~dennis@unaffiliated/dmp/x-546784)
15:43.16JeffMFixed size array all, I am disappoint
15:43.35JeffMallejo even
15:43.55JeffMBut the tutorials look good
15:44.15JeffMI'd just change function to method in the chapter one
18:58.08*** join/#bzflag khazhyk (~khazhyk@unaffiliated/khazhyk)
21:29.44*** join/#bzflag nadir (uid134094@gateway/web/irccloud.com/x-geaexdknvqqnyqtd)
21:30.40*** join/#bzflag Leppp (~Leppp@cpe-24-25-206-147.san.res.rr.com)
21:43.53*** join/#bzflag alfa1 (bee7eba0@gateway/web/freenode/ip.190.231.235.160)
22:22.01*** mode/#bzflag [+o blast007] by ChanServ
22:24.11*** topic/#bzflag by blast007 -> Temporary database maintenance/repair || https://BZFlag.org || http://ohloh.net/p/bzflag || https://bzflag.org/help || Channel Logs: http://infobot.rikers.org/%23bzflag/ || 2.4.10 released! https://forums.bzflag.org/viewtopic.php?f=8&t=19401
22:34.32blast007there's been three hardware issues with this Linode VM since March...
22:34.46blast007and this one seems to have messed up the DB a bit
22:36.08blast007forums are turned off at the moment, doing a backup, and then gonna compare the DB from this backups from earlier this morning
23:51.52allejoJeffM: fixed array size is more than enough for a plugin. what'd you use? a map or vector?
23:52.15JeffMMaps for life!!!!!!!!!!
23:53.48the_maphey, I'm finally useful
23:53.53allejohahah
23:59.31JeffMI can understand the array, calling methods a function could screw someone up if they end up googling on other sites
23:59.54JeffMSpecially if they don't know much c++

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.