2006-05-03 Timothee Besset * fixed a few warning and errors when compiling with gcc-4.0 ( didn't go all the way through, will stick to 2.95 ) * 1.41-3 Linux setup is hosed. it does not install the right binaries ( looks like the 1.40 stuff actually ) going to redo an installer from scratch more or less, and put everything right in there recovered 1.41 mp_bin.pk3 cgame/ui from Win32 installation, need a qagame? ( I can compile a new one myself ) 2006-05-02 Timothee Besset + no working Mac source! latest Mac source is likely at Aspyr. requested it + revert changes to tr_shader.c and go back to the previous "broken" SkipBracedSection calls need to release an updated binary to 1.41 and that fix would probably mean changing shaders as well 2003-09-16 Timothee Besset + java master and auth code. bind to specify interface for operation on multi-homed etmaster (192.246.40.38) + wolfKEY is NOT checked in to CVS 2003-08-31 Timothee Besset + show_bug.cgi?id=626 built new RTCW 1.41 setup working for more recent distributions (glibc 2.3 based ones) 2003-01-22 Timothee Besset + fix typo in Sys_Chmod (linux) 2003-01-16 Timothee Besset + update cons scripts for building on debian sid (default gcc version has changed) + show_bug.cgi?id=570 fixed SkipBracedSection usage to actually do it's job fixed a syntax error in common.shader, updated version is REQUIRED to run current code and yeah, that brings a considerable speedup to map loading 2002-4-12 Timothee Besset + releasing 1.41 update: - modified the setup scripts to generate a 1.4 -> 1.41 incremental - only putting the qagame as updated content - adding RTCW-README-1.4.txt (server admin information) to the setup - changed the INSTALL installation instructions to state about update and not full setup - commented out the full setup builds, update Conscript-setup if you need them 2002-13-11 Timothee Besset + show_bug.cgi?id=569 merged Rhea's g_antilag.c changes - cancel out syntax changes (makes diff harder to read) - IS_ACTIVE macro replaces our former test: x->r.linked == qtrue && \ x->client->ps.stats[STAT_HEALTH] > 0 && \ x->client->sess.sessionTeam != TEAM_SPECTATOR && \ (x->client->ps.pm_flags & PMF_LIMBO) == 0 \ - we only walk through active clients 2002-12-11 Timothee Besset + show_bug.cgi?id=569 fix for medic bug was validated with today's testing\ haven't merged against Rhea's code yet. will do 2002-11-11 Timothee Besset + show_bug.cgi?id=568 Rhea: better fix for loss of weapon functionality 2002-11-5 Timothee Besset + added in_subframe to toggle X subframe event handling + reworked the timing code to be more reliable + cleaned up dgamouse/in_mouse code, removed unnecessary dgamouse var + made the mouse grabbing an in_nograb cvar, no longer a compile time option in_nograb 1 forces in_dgamouse 0 and r_fullscreen 0 (any of those two will b0rk) + last safe check for screwed up X window server was too agressive tweaked to ignore when the subframe correction is too high 2002-28-10 Timothee Besset + ATVI Wolfenstein Misc #270 checking in a fix that doesn't crash g_syncronousClients + no longer prevent demo record on g_synchronousClients 0, just warn + ATVI Wolfenstein Misc #472 gun clipping, made the leaning bbox a bit bigger (along z<0) to cover gun clipping 2002-25-10 Timothee Besset + ATVI Wolfenstein Misc #474 spectator following a client, client gets kicked, was not sending spectator back to a correct free fly mode + ATVI Wolfenstein Misc #462 added a 'noToggle' flag to itemDef, uses a CVAR_NOTOGGLE this allows for ITEM_TYPE_YESNO that rely on an action uiScript to set their value (see sv_punkbuster in createserver.menu) added trap_PbSetSvStatus, can't set Punkbuster on server side if not available new mp_pak5 with updated createserver.menu (for all i18n) + updated ui scripts for mp_pak5 against ATVI's latest version (merged back new elements) 2002-24-10 Timothee Besset + ATVI Wolfenstein Misc #469 MG42 switching view angles to the player and back (prediction issue) + ATVI Wolfenstein Misc #470 fixed MG42 repair hint to work again when crounching or standing on top of MG42 (breakage was a side effect of mg42 mount hint fix, see 2002-17-10) + ATVI Wolfenstein Misc #413 maintain correct mp_currentTeam while doing free fly spectating (was causing problems joining teams) 2002-21-10 Timothee Besset + ATVI Wolfenstein Misc #461 'voting disabled on server' was not setting cg_draw2d to 0, parasite text was showing tweaked the translation for 'or callvote a start_match' by 'or call a vote to start match' wrote French traductions, others are pending (ATVI) 2002-19-10 Timothee Besset + ATVI Wolfenstein Misc #439 making sure we have a default empty error message for the connection diagnostics so that we don't pop ghost ones back + ATVI Wolfenstein Misc #460 cleanly shutting down listen server when using "exit to main" from UI + ATVI Wolfenstein Misc #467 mg42 client and client being pushed both become CONTENT_CORPSE till they get in a correct position again problem was, the call to trap_EntitiesInBox returns the mg42 client to the list when it queries for itself so numtouch stuck to 1, mg42 never returning to solid again added a safe check that hit != self + added g_dbgRevive to verbose the WolfReviveBBox stuff, was getting too noisy 2002-17-10 Timothee Besset + merged in mp_pak5.pk3 update from ATVI + ATVI Wolfenstein Misc #433 Fixed the screen shaking situation on the client Arnout: + Fixed double MG42 fire sound playing. + ATVI Wolfenstein Misc #436 Fixed MG42 muzzle flash not always drawing. + ATVI Wolfenstein Misc #437 Fixed revived client getting stuck in MG42 user client + ATVI Wolfenstein Misc #438 Fixed ability to use MG42's from above + ATVI Wolfenstein Misc #449 Fixed client getting stuck in walls after MG42 gets destroyed + Fixed MG42 hint icon not being consistent with the actual usage 2002-16-10 Timothee Besset + show_bug.cgi?id=562 Rhea - sniper zoom bug when playing demos applied fix to demo playing / sniper zoom 2002-13-10 Timothee Besset + ATVI Wolfenstein Misc #414 anytime enterTime gets reset, player is forbidden to switch team for 30s (with the exception of warmups and 10s after a map_restart) if player went back to free fly from follow spectating, enterTime was reset. Fixed. map_restart (i.e. warmup ends) was causing a reset of enterTime for all players. Fixed. now we only reset enterTime and prevent joining new team only when player explicitely sets new team 2002-12-10 Timothee Besset + ATVI Wolfenstein Misc #374 on first run the ui_vote* variables where not initialized, causing a g_voteFlags 0 (even if UI showed 'Voting Enabled', as long as you didn't raise the precise voting menu) added a uiScript voteFlags open in the open menu to cope with this affects all translations, new mp_pak5.pk3 + show_bug.cgi?id=560 LT getting too much ammo (thks Rhea) 2002-7-10 Timothee Besset + ATVI Wolfenstein Misc #304 setCvar "com_errorDiagnoseIP" "" is misread by the parser, second argument is ignored adding a clearCvar ui script, and updating UI accordingly 2002-2-10 Timothee Besset + ATVI Wolfenstein Misc #345 server URL message text lights up like a button - fixed tweaked the layout of URL confirm and serverinfo dialogs for all i18n made the various buttons look right (and not overflow their box and stuff) + ATVI Wolfenstein Misc #281 removed map_restart line from server creation menu (we never had that one in the UIs) 2002-1-10 Timothee Besset + ATVI Wolfenstein Misc #339 added switch back to free spectating, binded to the activate key (usually enter) you can't free spectate if you are in a team already (otherwise you could send indications) had to go around some ugly things in the limbo management on client side + ATVI Wolfenstein Misc #350 made server prediction off by default 2002-28-9 Timothee Besset + merged a new mp_pak5.pk3 with ATVI updates 2002-24-9 Timothee Besset + SV_CheckPaused using Cvar_Set "sv_paused" instead of setting the integer directly could have bad consequences over PB cvar hack checks actually this is not relevant to RTCW because cl_paused is never set to 1, always remains 0 but it might matter later on, better be on the safe side of things + cons post build on win32 copies PB stuff 2002-19-9 Timothee Besset + show_bug.cgi?id=543 more rcon fixes propagated from Q3 2002-18-9 Timothee Besset + ATVI Tracker Wolfenstein Misc #322 global 'Allow Voting' UI toggle was broken g_allowVote toggles all g_voteFlags on or off when used + NOTE: we have newer ui_mp/french/filter.menu than the one that was in ATVI's Wolf14 zip newer mp_pak5.pk3 altogether with #322 2002-17-9 Timothee Besset + ATVI Tracker Wolfenstein Misc #320 line wrap + skipnotify bug + ATVI Tracker Wolfenstein Misc #336 show_bug.cgi?id=550 quick fix to new rcon code 2002-5-9 Timothee Besset + ui_mp/ translations update from ATVI translated french/filter.menu FIXME: other filter.menu tranlsations still have to be done + update build system with an optional 'pk3' on build command line, defaults to off + show_bug.cgi?id=443 Linux client: propagated sub-frame event timing from Q3 2002-2-9 Timothee Besset + added g_antilag in filter.menu (Both/Yes/No) + handy, keep around: __asm__ __volatile__ ("int $0x03"); + show_bug.cgi?id=540 'ui failed pure server check' not doing UI properly was caused by the search path reorder on server join using FS_Restart in FS_PureServerSetLoadedPaks to fix + copied over ingame.txt to menus.txt, even if menus.txt is not used, it's safer to have the same unique ver as ingame.txt (which is our actual reference) + fixed Com_DPrintf being fucked up in ui code + show_bug.cgi?id=541 XP backport (script_mover_blocked) + ATVI Tracker Wolfenstein Misc #263 non anti-lag servers have their ping highlighted in yellow 2002-31-8 Timothee Besset + translations: got updated translations from John Fritts on 30/08 integrated, fixed french/createserver.menu french/serverinfo.menu which wouldn't load (missing newline) French/German 'Friendly Fire' in filter.menu was still 'Both Yes No', replaced by correct phrasing server browser was displaying server lines with Yes/No for punkbuster. Made it translation sensitive 2002-30-8 Timothee Besset + ATVI Tracker Wolfenstein Misc #284 rcon commands where sent after being tokenized and rebuilt that was breaking any quoting, for instance 'rcon g_motd "hooka pooka"' added Cmd_Cmd() to retrieve the un-tokenized command and transmit as is on both ends 2002-28-8 Timothee Besset + ATVI Tracker Wolfenstein Misc #263 renamed g_hitscanPredict into g_antilag, made it a CVAR_SERVERINFO added UI in server creation menu (defaults to 1) + internationalisation: added placeholder ui_mp/french/ ui_mp/german etc. .menu files need the new ones to put into mp_pak5.pk3 hardcoded translations: translations/translation-1_4.cfg looks like the "WARNING: You are missing" is a mistake on my side. That has been removed. NOTE: trailing \n can be omitted in the translation file NOTE: no 'need paks' translation for spanish / 'need paks' is only printed to console, was not strongly necessary NOTE: PROTOCOL_MISMATCH_ERROR is a server print, we can't translate this (we do PROTOCOL_MISMATCH_ERROR_LONG) added CL_TranslateStringBuf and support for \n in translation files (for PROTOCOL_MISMATCH_ERROR_LONG) NOTE: PB_MESSAGE is server side too, can't deal with it 2002-27-8 Timothee Besset + ATVI Tracker Wolfenstein Misc #218 using [skipnotify], only printing to console + ATVI Tracker Wolfenstein Misc #244 fixed Refresh button being broken in server browser server info dialog + ATVI Tracker Wolfenstein Misc #273 voting config defaulted to no, the g_voteFlags was initialized in game code only if you go to ui first, you were getting g_voteFlags 0 (doesn't exist yet) had to add the init in sv_init.c (which makes it kinda hardcoded, something I don't like) for testing the fix, need to start from a fresh install (or delete wolfconfig_mp.cfg) + ATVI Tracker Wolfenstein Misc #270 disabled lean while tossing dynamite + ATVI Tracker Wolfenstein Misc #271 have serverinfo menu be modal over the ingame options window had to remove the menus.txt + ingame.txt loading order and put everything seperate in menus.txt and ingame.txt NOTE: the UI_LoadMenus is a mess, ingame or not, that's still ingame.txt that gets loads (previously menus.txt + ingame.txt) in short, menus.txt is no longer used 2002-26-8 Timothee Besset + show_bug.cgi?id=472 propagate Linux client keyboard fix from Q3 + show_bug.cgi?id=62 propagate the invisible players/entities fix from Q3 (not sure this had ever been happening in Wolf though) 2002-23-8 Timothee Besset + ATVI Tracker Wolfenstein Misc #200 #201 fixed allies timeleft warning not working (was loading the wrong sound) issue is not new in 1.4, we had that in 1.33 already + ATVI Tracker Wolfenstein Misc #221 colored g_motd string is now supported + improved the pre-map_restart ignoring of cient messages, handles multiple map_restarts in a row + SV_InitGameVM: clearing the entity array before calling into the VM (reported by djbob, XP backport) (NOTE: does this fix show_bug.cgi?id=62 invis player?) + reworked the server 'client text ignored' message to only trigger when there's actually a message that doesn't get to the game VM + show_bug.cgi?id=376 had to shrink the rcon buffer to 256-16, seems it's the only value that won't cause to loose part of the output + ATVI Tracker Wolfenstein Misc #191 FF off + airstrike -> spectator was still hurting teammates + updated cons script to run under win32 and do misc handy tasks 2002-22-8 Timothee Besset + ATVI Tracker Wolfenstein Misc #196 fixed voting UI to remember settings with low ACCEPT and BACK buttons fixed parasite grpVoting still showing when you go back to the menu + ATVI Tracker Wolfenstein Misc #210 fixed callvote UI getting messed up in Misc tab + ATVI Tracker Wolfenstein Misc #193 & #195 overflow causing stack trashing. We are lucky it didn't cause a crash limit the numer of lines to 23 to keep in screen + ATVI Tracker Wolfenstein Misc #219 made g_motd a CVAR_ARCHIVE + ATVI Tracker Wolfenstein Misc #212 correct blue background on buttons in error diagnostic and server info 2002-20-8 Timothee Besset + show_bug.cgi?id=536 tweaking the SV_ExecuteClientMessage code some more to also let through the messages with bad serverId but last client command was a nextdl kind of hackish, but that's really the only fix that I can see for now 2002-15-8 Timothee Besset + show_bug.cgi?id=534 fixing rcon being broken on NT/XP with > 23 days uptime (or so) NOTE: this is ATVI Tracker Wolfenstein Misc #203 2002-9-8 Timothee Besset + show_bug.cgi?id=500 re-enabled ingame (meaning, non PB) IP banning code with the needed fixes: + addip doesn't cause an ERR_DROP and overflow on the g_banIPs cvar anymore + syntax for subnet matching has changed: use 192.168.1.* the former 192.168.1.0 syntax would have bad consequences, see show_bug.cgi?id=529 + added a 'Reconnect' button next to the 'OK' button on the diagnostic screen (realized that not everyone does have the reflex to toggle console and hit 'reconnect') + updated build system, pushed version to 1.4 2002-8-8 Timothee Besset + show_bug.cgi?id=405 dedicated server signal handling fix, propagated from Q3 + show_bug.cgi?id=480 dropped clients under heavy load, applied the patch, haven't been able to verify that it's an effective fix at this point + show_bug.cgi?id=425 splines code cleanup + fixed 'vote, die, vote again' exploit 2002-7-8 Timothee Besset + latching com_maxfps, this should considerably annoy everyone who's using it for various cheats + fixing lean / see through walls. this was happening because of bg_pmove.c and cg_view.c not computing the view offsets the same way. 2002-6-8 Timothee Besset + prevent centerview exploit on sniper rifle recoil blocking centerview for 1s after a zoomed sniper rifle shoot 2002-3-8 Timothee Besset + configure voting with the serverinfo cvar g_voteFlags UI on server side startmenu UI on client side: retrieve the vote flags, only show the allowed votes 2002-1-8 Timothee Besset + fixed the AfterBuild code, now checking for unresolved in the DLLs unfortunately, this is broken with pcons! + show_bug.cgi?id=498 fixed NET_AdrToString to print the port as unsigned int (for ports > 1^^15, was showing negative) + show_bug.cgi?id=524 changed default GL driver from libGL.so to libGL.so.1 see LSB 1.2 spec: http://www.linuxbase.org/spec/refspecs/LSB_1.2.0/gLSB/libgl.html 2002-31-7 Timothee Besset + show_bug.cgi?id=517 ERR_DROP if PB client off / server on conflict when starting local server 2002-30-7 Timothee Besset + flame thrower cheat, run and aim at your feet, you don't take damage fixed by a server side trace, check hit on the ground within a small radius of the player (might need to configure the tests some more if this triggers too easily) 2002-24-7 Timothee Besset + on FF-off servers, if you drop an airstrike and join spectator, teammates would get killed now the airstrike won't hurt anyone, neither team or opponents + cleaned up cg_emtpySwitch. replaced by cg_autoReload, pollutes userinfo + move Sys_PBSendUdpPacket to qcommon/common.c to avoid dupe code win_net.c unix_net.c 2002-23-7 Timothee Besset + merged autoreload branch back into RTCW-1_34 cg_autoReload cvar (defaults to 1) finalized auto reload code, added UI in the Controls > Shoot menu (next to the reload weapon binding) + backported pmoveExt_t from XP, storing the auto reload settings for bg_* code there (deals with transmission to the server too) + win32 .dsp: added RTCWBASE env variable, copying as post-build step + show_bug.cgi?id=512 Wolf server buffer overflow vulnerability, several vsprintf -> Q_vsnprintf to secure + MAX_SHADERS bumped from 2048 to 8192 (XP backport) + backport of g_antilag.c from XP (Gordon code) included in build with appropriate modifs, udpated the data init and trace calls (Bullet_Fire_Extended) TODO: in XP, we'll have to make antilag the default of regular game also fixes the headshot behaviour, headshot was checked only if there's collision with main player bbox + all new g_antilag.c stuff is controled by g_hitscanPredict cvar (default 1) 2002-12-7 Timothee Besset + lieutenant ammo pack gives 8 colt/luger rounds NOTE: the max rounds are 32 for colt/luger, 90 for 2-handed weapon + using /kill while at MG42 forced player to stay in MG42 anim mode after respawn, fixed 2002-10-7 Timothee Besset + Soldier/lieutenant could loose all weapon functionality when dropping weapon while throwing a grenade fixed, can't drop weapon while holding grenade + updated the Linux build system, files with the same name compiled to both cgame and game or ui and game files with same names are confusing gdb when setting breakpoints. dupe cgame files are prefixed with cg_ and dupe ui files are prefixed with ui_ (cons rocks) 2002-9-7 Timothee Besset + added ui in advanced server menu for g_useralliedmaxlives g_useraxismaxlives g_fastres and fastres milliseconds + added fastres milliseconds config to the game code (I thought this was in already, but it was not) 2002-8-7 Timothee Besset + show_bug.cgi?id=509 strategy for referenced pk3 in mod dirs (fs_game) is now the same as in the basegame post-connection diagnostic screen shows the list of missing pack files if any 2002-6-7 Timothee Besset + not drawing motd while doing auto-download (there's download progress instead) + using a grey filled rectangle below autodownload messages for readability misc UI tweaks 2002-5-7 Timothee Besset + show_bug.cgi?id=509 see branch 'test_dl_ratio' for an experiment with pushing up the rate ratio specifically for downloads (foxed) improved speed of the auto-download regular server tops at 20kb/s max autodownload tweaked update server tops at 30kb/s (by pushing up the sv_fps to 60) added sv_dl_maxRate for independant autodownload max rate (autodownload used to be limited by sv_maxRate) + show_bug.cgi?id=491 SV_SendMessageToClient was defeating the sv_lanForceRate setting by forcing snapshot emitting for LAN clients 2002-4-7 Timothee Besset + auto-update UI: now showing a modal dialog box on the main menu, forcing the user to click either 'OK' or 'Later' + misc Linux build system tweaking for auto-update server build 2002-3-7 Timothee Besset + show_bug.cgi?id=490 draw motd on connection screen + drawing a grey rectangle below error message on connect screen, changed font scale (better readability) fixed the global motd which was not displaying on connect screen (y offset was offscreen) 2002-2-7 Timothee Besset + modified edit field to prevent text getting written outside of it's bounding box the editfield scrolls like the console does fixes the "Say: " and Name edit field in limbo menu + added Paint_AutoWrapped_Text the hardcoded ui areas (such as UI_LIMBOMENU) didn't allow text wrapping on proportional fonts so far used the Paint_Text code to make a text wrapping version also allows to evaluate the size requirements (number of lines) without actually drawing anything + making limbo chat text wrap and be multiline (based on above) limbo chat prints text and scrolls irc-style, much more intuitive 2002-28-6 Timothee Besset + modified the protocol error code: the connect screen can't draw too much information, 256 chars is the max then we print a more complete message in the error diagnostic screen (i.e. after the client hits esc and aborts connection) TODO: i18n of those error messages? + tweaked the diagnostic window, have to click through a button to show server info in a new window the diagnostic window itself only has the error message and the URL buttons + reworked the URL button disclaimer completely, prompting each time an URL is going to be opened (reverted multi_setup.menu to it's old version, removed the persistent pref toggle) + unified the com_errorMessage code to use the same UI scripts wether or not a server IP is available for diagnostic + in-game server info button, in the Options menu the button is present / disabled depending wether the server is local NOTE: it would be useful to have customizable origin for modal windows too 2002-27-6 Timothee Besset + show_bug.cgi?id=507 adding a "com_errorDiagnoseIP" if this is set to a server IP, the UI error dialog switches to a different and more complete dialog which displays server information with clickable URL buttons 2002-26-6 Timothee Besset + making protocol mismatch sending a print OOB message with a markup tag to propagate into a com_errorMessage + made CL_DisconnectPacket use com_errorMessage too to provide a dialog box 2002-25-6 Timothee Besset + adding ui_doURL CVAR_ARCHIVE, editable in mutliplayer > multiplayer setup + adding MAIN/ui_mp/urls.menu (and binding into the loaded set) + adding SERVERINFO "URL" to g_main.c, defaults empty + adding 'modal' menuDef keyword, causes menus to be called as truly modal (the caller menu to return to is stored in a stack) added WINDOW_MODAL window flag existing menuStack array, re-using (only for modal windows though instead of at all times) 2002-24-6 Timothee Besset + adding CVAR_SERVERINFO|CVAR_ROM "mod_url" to g_main.c defaults to http://www.activision.com/games/wolfenstein/ this is meant to be configured programmatically by mod authors and not altered by server config + addding 'Go to server URL' 'Go to mod URL' to the server browser's serverinfo_popmenu (with toggles visible / non visible depending on presence of the variable) 2002-21-6 Timothee Besset + foreign characters in name confuse clients that don't use the matching locale adding securities and cleanups to remove all foreign chars from CVAR_USERINFO cvars (check in Cvar_Set2, and in Cvar_Get) the cvar in the general case can hold foreign chars (ex. ui_objective), only CVAR_USERINFO are restricted we could let in foreign chars with the userinfo if there's a cleanup client side (wether or not the incoming name is in the same local as we are) that would be a bunch of extra work, and is kinda overkill.. 2002-20-6 Timothee Besset + show_bug.cgi?id=506 multiple/on-demand mp_bin management (mod request, and general utility otherwise) made mp_bin referenced by server (that's for solving test case #1) + show_bug.cgi?id=506 client reorganizes the pk3 search order to match the server's this solves test case #3, and makes likelyhood of 'invalid .PK3 files referenced' much lower. + win32 fixups (win32 Sys_GetDLLName) 2002-19-6 Timothee Besset + cleanups, dead code removal + tagging last 1.33 source before starting some 1.34 work post-release-1_33 + added securities ldd -r on the .so needs backporting to 1.33 mod source release + applying Xian g_maxlives and max lives enforcements code (with some tweaking) + propagate tr_image.c ResampleTexture overflow fix from Q3 2002-13-6 Timothee Besset + SAD spawn bug has reappeared (Wolfenstein Misc #114) attempting a fix: putting a correct srand call in SAD, storing a randomized table of the feeds 2002-11-6 Timothee Besset + built a seperate GOTY map pack building release candidate for 1.33 full, 1.32->1.33 update, and Map Pack + make lightded setup go look for the correct version string (in SP) + partly renamed some 'Light Dedicated' to 'Standalone Dedicated' + using new loki_setup code on the way 2002-9-6 Timothee Besset + tweaked the 'Server uses protocol %i' to 'Server uses protocol %i (your client is using %i)' + show_bug.cgi?id=502 changed CopyDLLForMod to handle copy error correctly (only affects Linux client, suspect this could have caused sv_pure bypassing) 2002-5-6 Timothee Besset + fix a screwup in light ded code (something very weird broke the checksum generation) 2002-4-6 Timothee Besset + nobinaries="yes" in sdk setup, run sdk tests only on spoutnik 2002-3-6 Timothee Besset + build_sdk.sh script, added setup content, ready to release mod sdk for Linux 2002-31-5 Timothee Besset + cleared up bad fix on cl_guid (PB) + made g_enforcemaxlives a CVAR_ARCHIVE, and defaulted it to 1 (cheat exploit protections should always be on by default) + moved ClearMaxLivesIP call post G_RegisterCvars (otherwise, filterban is not initialized) + fixed the g_filterBan disabled (forgot to comment out a line) + the IP information was not enforced in userinfo on server side, added a check in SV_UserinfoChanged + made sure we only do maxlives enforcement when a game with maxlives is actually running 2002-30-5 Timothee Besset + PB DLLs for Mac (v 989) from Tray + show_bug.cgi?id=500 disabled g_filterBan again (not functional yet, need good source for 1.33) + imported changes from Xian, polished and cleaned up a bit for Linux build + show_bug.cgi?id=491 added sv_lanForceRate, defaults to 1 if 0, the clients on a LAN are not forced by the server to max rate, server listens to their settings request + new WolfReviveBbox code from Splash Damage (fixes a medic revival stucking) Xian changes: * Added g_axismaxlives and g_alliedmaxlives. These cvars superscede g_maxlives when set. Used for setting maxlives independantly for each team * Added g_fastres 1 (Defaults to 0). Player instantly gets up, is instantly mobile, but is only invulnerable for a second vs 3 seconds the old way * Added g_knifeonly 1 (defaults to 0). This is a novelty for the mp_destruction crowd on my servers. If this gets ganked, I won't lose any sleep over it. Players spawn in with knives only. Medics get 20 needles instead of 10 and can drop health as normally. Engineers can drop and arm dynomite. All the MG42's are removed from the world. * Added g_enforcemaxlives 1 (defaults to 0). This code is still a little buggy for reasons unknown to me. Timothee will give it a once over. We may choose to make this undocumented. The way this works, is that when a client enters the game, their IP gets added to a list. Any IP on the list is banned from connecting until the next round starts. This is to prevent people from disconnecting and reconnecting to get by the maxlives limitation. * listmaxlivesip (debug console command). Prints the current list of clients who's IP's are being enforced with maxlives. 2002-29-5 Timothee Besset + PB 989 Linux from Tony (where are win32 DLLs?) + tracker #414, missing g_filterBan, broken IP filtering added g_filterBan back, and the corresponding IP filtering calls (upon client connection) 2002-28-5 Timothee Besset + adding pcons script, configuring for parallell build + adding mp_pak4 to light ded checksum, updating the setup 2002-23-5 Timothee Besset + last minute update from Activision & SplashDamage for 1.32 setup 2002-22-5 Timothee Besset + 1.32 release, got win32 RC, building Linux RC - some changes in main/ mp_pak4.pk3 was in SplashDamage's 1.32 zip, is not installed new/modified files: mp_tram.pk3 sp_pak3.pk3 - no changes in Docs/, only installs PB manual + updated rotate.cfg to rotate on mp_tram + updated CHANGES file (public CHANGES file in setup) 2002-13-5 Timothee Besset + new code 1.32-MP from SplashDamage, rebuild + merge back into trunk, there is going to be a new build + 1.32g is in QA at Activision, creating a post-1_32g branch + show_bug.cgi?id=476 CVAR_CHEAT|CVAR_LATCH .. the latching was happening before the cheating check switched cheat check before latch + added anjuta project file and run.sh wrappers 2002-12-5 Timothee Besset + added TuxGames info in faq 2002-7-5 Timothee Besset + import of splashdamage's updated source + cleared CG_SELECTEDPLAYER_ARMOR stuff, the menudefs.h doesn't have it 2002-26-3 Timothee Besset + SOS update, fixes shader overflow bug show_bug.cgi?id=471 2002-25-3 Timothee Besset + 1.31 update, grabbed from SOS and rebuilt + build_setup.sh update: antares needs -L option for cp, spoutnik doesn't + new PB files, pb/htm/ directory 2002-16-3 Timothee Besset + new set of checksums for the light dedicated server (merged in the new mp_ice map pak) -- tagging tree as Release-1_3-tag + last minute fix to build_setup.sh script + built Wolf 1.3 setup 2002-15-3 Timothee Besset + blanking default CD path + dedicated + mod -> cl_cdkey buffer overflow, causing a corrupted ucmds static struct cleaned up the cl_cdkey and associated functions stuff, moved declaration to qcommon.h and cl_cdkey alloc from server/cl_main.c to common/common.c + media-only mods failed to load because the DLL was not found we handle this a bit differently than how SP code does, if we can't find a DLL, we will try to copy it from a search in the base game (with the usual homepath/basepath search order) + rebuilt light ded's pak0.pk3, generate new checksums 2002-13-3 Timothee Besset + merge from SOS, breaks the custom auto-update server compile time define (going through several autoupdate servers now) 2002-12-3 Timothee Besset + wolf-beta/2002-March/000079.html switched from system to execl in Sys_DoStartProcess hacked around to keep a system() call when a command line is needed (HACK!) + text dialog installation: dialog box wraps the text 2002-11-3 Timothee Besset + modified setup, PB is in a seperate option, has his own license prompt 2002-02-3 Timothee Besset + autodownload was breaking for files > something between 3Mb and 6Mb because of autoupdateFilename pointing to a char * returned by va() (va wrapping past some critical size corrupted the name) + fixed Cbuf_ExecuteText("quit") to Cbuf_ExecuteText("quit\n") 2002-28-2 Timothee Besset + show_bug.cgi?id=464 changed the 'servers not listed' message, it was misleading + added ./cons -- nograb compile option to avoid mouse grabbing / easier debug + show_bug.cgi?id=463 checking the .so version in fs_homepath against main binary version to avoid outdated ones + show_bug.cgi?id=462 send potential remaining fragmented packets before sending a gamestate 2002-27-2 Timothee Besset + 1.3.b1 with new setup code (xsu-enabled) + auto update server pointing to zerowing 2002-26-2 Timothee Besset + show_bug.cgi?id=455 removed old libMesaVoodooGL.so loading code Voodoo cards should use XF4/DRI, that load code was outdated and confusing people with broken OpenGL + bumped GL extension string buffer to avoid truncate + added NV fog extension to MP, propagated existing SP code added corresponding UI code + validated on win32 + updated Linux faq 2002-02-25 - fixed missing param in 'dropped gamestate, resending' Com_DPrintf message 2002-02-21 Timothee Besset - update_mp_bin.sh to update new stripped .so to the mp_bin.pk3 ./update_mp_bin.sh 2002-02-20 Timothee Besset - tweaks to win32 auto-update Sys_StartProcess 2002-02-19 Timothee Besset - adapted the auto-update code to linux saving in the right spot, doing a chmod +x before execution - tweaking Sys_StartProcess on linux, kernel panic if no sleep() temporisation between the spawn and the exit - added a Sys_Chmod on linux 2002-02-18 Timothee Besset - merging in auto-update server code from SOS - added the appropriate build system changes to compile update server 2002-02-17 Timothee Besset - merged DLL path in vm_s from SOS 2002-02-15 Timothee Besset - merged all dllchecksum stuff to trunk - updated from more recent SOS - fixed win32, tested on win32, build light dedicated bins on win32 --- branch dllchecksum --- TODO: hanging at map load screen on disconnect is still present (linux client only issue AFAIK) - renamed FS_ExtractFromPakFile to FS_CL_ExtractFromPakFile, only compiled with client code - light dedicated server: we won't collapse mp_bin.pk3 into the single light dedicated pak0 because it's bound to change regularly and that would mean recomputing the set of checksums each time tested light dedicated to work the same against the test cases used for full server - isolated on branch for testing and polish of the dll checksum code / testing of light dedicated TODO: if server w/ different mp_dll.pk3 than the client, doesn't pass the sv_pure check client bails with: Sys_Error: Couldn't extract ui.mp.i386.so that would be the most common error users would get, would be better if we could go back to UI / Disconnect cleanly TODO: need also to make sure that it's not possible to auto-download the mp_dll pk3? could be used to hack clients from a fake server (this situation doesn't happen right now because of the above error) - cleanup to easier cross platformness of the DLL checksuming code - plugged destData mem leak in FS_ExtractFromPakFile - moved the declaration of FS_BuildOSPath and FS_ExtractFromPakFile to qcommon.h for cleaner stuff - adding centralized per-OS SYS_DLLNAME_* defines in qcommon.h for the DLL strings (makes the code in SV_VerifyPaks_f and FS_ExtractFromPakFile unified between OSes) - string shifting func decs moved to qcommon.h 2002-02-14 Timothee Besset - modified unix_main.c Sys_LoadDll to scan pwd/homepath/basepath (homepath/basepath only for release build) made the basics of the DLL checksuming work on linux (SV_VerifyPaks_f, extraction of the DLL files) - sv_pure instead of cl_connectedToPureServer in Sys_LoadDll (so that dedicated builds too) (TODO: same on win_main.c) - modified FS_ExtractFromPakFile to take the file to check as input wrote the *nix-style path scanning in unix_main.c, passing to FS_ExtractFromPakFile (TODO: fix the win32 build) 2002-02-13 Timothee Besset - merging in sv_pure 1 / pb 0 DLL checksuming code TODO: linux / linux client server linux / win32 client server combinations light dedicated TODO: make sure disabled if PB is on TODO: start testing for mods too? Current state: builds on linux, won't run anything with sv_pure 1 (the linux parts of the checks are not in) 2002-02-11 Timothee Besset - updated the cons build script 2002-02-08 Timothee Besset - tweaking the 'out of disk space' error message for linux ver 2002-02-06 Timothee Besset - client/cl_main.c, localisation code disabled on Mac - merged in latest version from SOS - looking at various constants that are different for the Mac version client/snd_local.h MAX_CHANNELS was changed from 96 to 64 for Mac MAX_STREAMINGSOUNDS from 24 to 12 for Mac client/snd_mem.c sfxScratchBuffer moved from malloc to Hunk_Alloc (all ver) game/bg_public.h various MAX_* constants: #if defined(__MACOS__) //DAJ HOG #define MAX_ANIMSCRIPT_MODELS 32 //DAJ tried 24 // allocated dynamically, so limit is scalable #define MAX_ANIMSCRIPT_ITEMS_PER_MODEL 1024 //512 #define MAX_MODEL_ANIMATIONS 256 // animations per model #define MAX_ANIMSCRIPT_ANIMCOMMANDS 9 #define MAX_ANIMSCRIPT_ITEMS 64 #else #define MAX_ANIMSCRIPT_MODELS 32 #define MAX_ANIMSCRIPT_ITEMS_PER_MODEL 2048 #define MAX_MODEL_ANIMATIONS 512 // animations per model #define MAX_ANIMSCRIPT_ANIMCOMMANDS 8 #define MAX_ANIMSCRIPT_ITEMS 128 #endif qcommon/common.c MacOS isn't forced to a minimum 16Mb com_zoneMegs server/sv_ccmds.c sv_maxclients defaults to 16 on Mac (32 otherwise) server/sv_init.c sv_maxclients Cvar_Get default to 16 on Mac (20 otherwise) <- is this a dupe? ui/ui_shared.c Controls_SetConfig doesn't do a in_restart on Mac, does otherwise 2002-02-05 Timothee Besset - merged from WestLake-1_1b into the trunk - checked on win32 (full and light dedicated, + PB code) -- branch WestLake-1_1b -- - merging in Mac MP code update byte swap for aas_routingcache_t smaller memory footprint on some static structs (grep for HOG) renderer/tr_model.c: the r_cache 1 (R_Hunk_Begin, R_Hunk_Alloc) code is disabled in the Mac version relies on direct ri.Z_Malloc calls keeping the original version working with a 24Mb hunk in_restart command removed from Controls_SetConfig (ui_shared.c) adding the whole mac/ directory tree - guarding the #pragma export on/off with __MACOS__ - guarding DLL_ONLY for __MACOS__ - cl_main.c, 'Disconnected from server' message was commented out? (re-enabled) - fixed dupe CL_SaveTranslations_f definition - tr_image.c R_FindImageFile, uppercase search with developer messages 2002-2-4 Timothee Besset - bumped the versions to 1.2.b4 - merged the lightweight dedicated server code from Release-1_1b-branch into trunk - merged and tested linux PB code - wrote linux setup code for light dedicated server (./cons -- release light setup) - upgrade makeself to fix a CRC bug 2002-2-2 Timothee Besset - updated from SOS, more linux fixes around the PunkBuster code 2002-1-30 Timothee Besset dedicated server branch - defaulting to dedicated 2 (i.e. internet) in the dedicated server code - started linux setup code for light dedicated 2002-16-1 Timothee Besset - show_bug.cgi?id=441 adding brandelf calls to the setup building process so that our binaries run on BSD - preparing for 1.1b release, set version to 1.1b (first RC) 2002-1-16 Timothee Besset - show_bug.cgi?id=452 if doexit, spawn the process at actual process exit (after GL shutdown and other resources have been freed) - show_bug.cgi?id=439 Alt+ENTER fullscreen/windowed toggle 2002-1-15 Timothee Besset - show_bug.cgi?id=440 we were leaking 24Mb of RAM on each R_Init, re-allocating, not freeing existing chunk not doing the re-alloc now, sticking to the initial mem chunk - finalize lightweight dedicated code after recent merges allow .bsp load from pak0.pk3 in lightweight and demo.. generated 5 checksumfeed / pure checksums pairs to use randomly shifted the strings stored in the .exe to be safe validated individually each of the 5 checksum/string pairs 2002-1-11 Timothee Besset - merged in from SOS, PB stuff and new gameplay type - updated win32 project files, Dedicated Debug/Dedicated Release added PB files to the .dsp PB currently disabled on win32 light dedicated (and disabled on linux) 2002-1-10 Timothee Besset - started internal testing on light dedicated, building test binaries - win32 fix: need to #define DEDICATED and DO_LIGHT_DEDICATED 2002-1-8 Timothee Besset - implemented sv_pure checksums faking for light dedicated server currently restricted, runs with an hardcoded fs_checksumFeed - fixed openurl.sh not being executable in setup 2002-1-7 Timothee Besset - building .pk3 for dedicated only server, 22Mb - unix/extract-dedicated.pl util script - unix/dedicated-only.txt - building a dedicated on win32, added Debug Dedicated and Release Dedicated (also checked in the A3D .lib) 2002-1-2 Timothee Besset - full Wolf setup on linux now includes SP stuff will spawn cons in the SP tree for build and copy the binaries added a WolfMP.xpm icon and the relevant script stuff updated the INSTALL file displayed by the setup 2002-1-1 Timothee Besset - updated FAQ with BSD info (bug #441) 2001-12-31 Timothee Besset - bumped version to 1.2-MP - show_bug.cgi?id=419 reorganizing the Sys_OpenURL to use a working Sys_StartProcess propagating that code to Wolf SP 2001-12-25 Timothee Besset - building release setups for 1.1-MP (and Multiplayer DEMO) - tagging the tree Release-1_1-MP (Wolfenstein WolfMedia-demo WolfMedia-main modules) 2001-12-24 Timothee Besset - show_bug.cgi?id=371 hacked out to force PROT_READ and safe memset code in all cases - show_bug.cgi?id=427 opening URL through an openurl.sh script file - Sys_OpenURL and Sys_StartProcess exit with Cbuf_ExecuteText( EXEC_APPEND, "quit" ); 2001-12-23 Timothee Besset - show_bug.cgi?id=422 using .mp. (qagame.mp.i386.so etc.) for all MP stuff - protocol 50 for the demo - bringing the setups up to speed again, full setup and demo setup living together in the tree 2001-12-13 Timothee Besset - Sys_OpenURL stub 2001-12-12 Timothee Besset - show_bug.cgi?id=390 SIG_IGN on SIGTTIN SIGTTOU 2001-12-10 Timothee Besset - tweaking Sys_LoadDll load order: pwd first, then basepath (NOTE: currently there's no homepath loading, should add that) - tweaked vm_* defaults code 2001-12-06 Timothee Besset - show_bug.cgi?id=402 full scene bug fix 2001-12-02 Timothee Besset - unix/const-arg.c file, about const vec3_t[3] problems 2001-12-01 Timothee Besset - Wolf MP source passes the -Wall -Werror 2001-11-30 Timothee Besset - Updated the Wolfenstein FAQ http://zerowing.idsoftware.com/linux/wolf 2001-11-29 Timothee Besset - Wolf 1.0.b3 Linux fixed FreeBSD/OpenBSD/NetBSD setup 2001-11-27 Timothee Besset - Wolf 1.0.b2 Linux setup scripts and media for the complete multiplayer version of the game single player spawn in the menu is inoperant 2001-11-19 Timothee Besset - show_bug.cgi?id=51 propagated a Q3 fix, Wolf 0.9.20-5 2001-11-17 Timothee Besset - building 0.9.20-4 for release, tagging the tree as MPTest-0_9_20-4 - merging bug #371 fixes into trunk, 0.9.20-4 2001-11-14 Timothee Besset - show_bug.cgi?id=388 server not restarting correctly at time wrap modified the way we restart when time wrap occurs will no longer hang with no server after some time 2001-11-12 Timothee Besset - third iteration of sound code fixes, should be all good and ready to rollback to Q3 source too 2001-11-09 Timothee Besset - using temp branch Bug-371 for sound crash stuff testing (bug #371) PROT_READ|PROT_WRITE fails on Alsa, defaulting to PROT_WRITE only after initial failure (mmap failure returns -1 and not NULL, our failure check was bugged) 2001-11-08 Timothee Besset - setup script fix propagated from Q3 (correct return code) 2001-11-04 Timothee Besset - removed game/g_arenas.c (SP stuff) 2001-11-01 Timothee Besset - updating for MP test2, version 0.9.20 linux version of extractfuncs (has different cmd line) -- demo switch in the cons scripts updated the setup script, auto-extracting the Wolf version no longer hacking our way to assemble the .nasm files 2001-10-15 Timothee Besset - updated from the SOS source, rebuilt (added a dummy Sys_StartProcess) branch Files-1_30, Q3 1.30's filesystem with ~/.wolf support - merging the Q3 filesystem code for home directory support: files.c uses a MAX_ZPATH instead of MAX_QPATH for some names wolf source has a fs_verbose cvar? (only read usage, removed it) fs_basegame support is new fs_patchpath and fs_patch_oldpath in Wolf? are used to generate lists of the used files and generate a .bat of the files to remove doesn't seem to be a Wolf addition, is not in Q3 .. getting rid of it fs_filecompare removed FS_Delete is not in Q3 1.30, but used in Wolf some file streaming code not there either? maintained mp_*.pk3 sp_*.pk3 code in FS_AddGameDirectory FS_ComparePaks has unpure diagnostics (more verbosive) FS_PatchDeletedFiles_r removed leaving early exit FS_SetRestrictions (bypass copy protections) updated qcommon.h for MAX_FILE_HANDLES Sys_DefaultInstallPath / sys_DefaultHomePath we have implementation in unix/unix_shared.c missing the win32 implementation, grabbing the one from Q3 (win_shared.c) declaration is in qcommon.h TODO: qcommon/files.c has some Com_ReadCDKey code and stuff - fixed recursive errors related to logfile 1 2001-10-08 Timothee Besset - added PROT_READ to mmap on dma sound buffer (cf Q3 bug) show_bug.cgi?id=371 Com_Memset on dma buffer in Wolf doesn't happen systematically on startup but happens on map load - applied Sys_LoadDll changes from Q3 2001-09-28 Timothee Besset Planning the merge with wolfMPTest branch updated from SOS, merged current trunk content cvs merge with WolfMPTest branch checked build system, all fine there is new media, building from the MP Test and new stuff in SOS specifically ui_mp/ directory need to build as regular binary, dropping PRE_RELEASE_DEMO define dropping the -debug suffix on debug .so name, useless NOTE: should drop the loading into base path too, and use ./ ? Checked the merge result on win32: - cleaned CritialSection into CriticalSection - fixed extractfuncs to ignore "qmin" and "qmax" (OSX) WolfMPTest branch Changelog ========== BEGIN 2001-09-28 Timothee Besset Updated the FAQ (http://zerowing.idosftware.com/linux) Answered some tech support email and started bug tracking This version is going to be merged back in the trunk Tagging it to WolfMPTest-Merge 2001-09-21 Timothee Besset Current SOS is not compatible with MPTest branched in the last comaptible state building linux clients for release removing wolfconfig.cfg from the setup patched vm.c to vm_* 0 added cons -- setup to call build_setup.sh heavily patched setup script to turn it into a full setup. updated the INSTALL file merging in Graeme's source merged in - cgame/cg_main.c Graeme particles re-init fixes I have newer - cgame/cg_players.c CG_TrailItem Darin / Version 2 on 09/18/2001 Don't draw icon above your own head built cg_players.c.patch I have newer - cgame/cg_scoreboard.c WM_DrawClientScore Darin / Version 2 on 09/18/2001 added drawing of treasure holder in scoreboard built cg_scoreboard.c.patch merged in - cgame/cg_servercmds.c Graeme CG_ClearTrails call commented out in CG_MapRestart replaced - cgame/cg_trails.c Graeme various changes merged in - client/cl_cin.c Graeme commented out some OSX only code I have newer - client/cl_main.c up to Sean / Version 11 on 09/18/2001 morv cvar inits in CL_Init mostly (other small fixes) built cl_main.c.patch leaving as is - client/snd_dma.c I have fixed Sys_InitializeCriticalSection (was missing a 'c') leaving as is - client/snd_mix.c I have moved some vars to static for asm implementation snd_mixa.s leaving as is - client/snd_public.h only comments differ I have newer - extractfuncs/extractfuncs.h SMF - workaround for Graeme's predifined MACOSX functions built extractfuncs.c.patch I have newer - game/bg_misc.c DHM - Nerve :: updated delay so prediction is correct (panzerfaust) built bg_misc.c.patch leaving as is - game/g_client.c comments differ I have newer - game/g_cmds.c Darin / Version 7 on 09/17/2001 Can't mount an mg42 while ducking built g_cmds.c.patch reverting - game/g_misc.c my version has Darin / Version 3 on 09/17/2001 Testing new firing rate and damage since it was testing we should probably not release a new server that does the MG42 differently I have newer - game/g_weapon.c Darin / Version 7 on 09/17/2001 Don't allow engineer to repair mg42's if he is mounting one! built g_weapon.c.patch merging in - game/q_shared.c MACOS_X implements qmax / qmin FIXME: I also added #ifndef max #define max.. for linux in q_shared.h a while ago should we use qmax everywhere instead? (prolly so) merging in - game/q_shared.h idSqrt ppc implementation FIXME: this is currently guarded for OSX, it should be a CPU guard? linux ppc for instance should benefit? replaced - macosx/ copying over the whole macosx/ tree in my source NOTE: newly added SOS files might be missed in the process, hopefully there's none I have newer - qcommon/common.c changes 'generated by quake' into 'generated by RTCW' built common.c.patch leaving as is - qcommon/qcommon.h critical section missing a 'c' leaving as is - qcommon/unzip.c my version has gcc warnings bug fixes, not important for OSX building partial merge - qcommon/vm.c I have cleaned up vm_* init, but OSX should still init on vm_* 0 so I'm not building a patch for this Graeme switched MACOS_X to original id code in VM_DllSyscall (merged in) conflict - renderer/tr_animation.c merging in R_CalcBoneLerp, exiting in some cases (OSX fails on random memory reads) I had a different solution for baseIndex = ((int)pIndexes - tess.indexes); in RB_SurfaceAnim .. used baseIndex = (int)(pIndexes - (int *)tess.indexes); that does a int difference Graeme's fix denotes a better understanding of what's going on using his code I have newer - server/server.h Sean / Version 3 on 09/17/2001 TA Merge - adds a qboolean connected; to challenge_t built server.h.patch I have newer - server/sv_client.c Sean / Version 4 on 09/17/2001 TA Merge, see above built sv_client.c.patch NOTE: also has VM_Call( gvm, GAME_CLIENT_DISCONNECT, newcl - svs.clients ); commented out I have newer - server/sv_main.c Sean / Version 5 on 09/17/2001 TA Merge, see above also has some comment of my own about rcon redirect issues built sv_main.c.patch leaving as is - splines/splines.cpp gcc warning fix and MSVC-isms NOTE: still not sure the splines code is used in the client leaving as is - splines/splines.h comments differ I have newer - ui/ui_local.h Sean / Version 2 on 09/18/2001 NERVE - SMF - new font stuff built ui_local.h.patch I have newer - ui/ui_main.c Sean / Version 14 on 09/18/2001 NERVE - SMF - new font stuf built ui_main.c.patch I have newer - ui/ui_shared.c see above built ui_shared.c.patch I have newer - ui/ui_shared.h see above + MAIN/UI include case sensitivity fix built ui_shared.h.patch leaving as is - unix/ChangeLog / Cons* / build_setup.sh that's my stuff merging in - unix/unix_net.c OSX fixes merging in - unix/unix_shared.c Sys_DefaultBasePath deals with installPath init critical section returns (void *)-1 NOTE: some changes by me, leaving as is for now leaving as is - unix/setup subdirectories leaving as is - win32 got a crash on zerowing server, fixed in g_cmds.c rebuilt g_cmds.c.patch for zaphod same as previously happened: #0 0x805ee66 in CM_Trace (results=0xbfff6d60, start=0x14, end=0xbfff6ef8, mins=0x45a1e3c0, maxs=0x45a1e3c0, model=0, origin=0x80b97a0, brushmask=1, capsule=0, sphere=0x0) at debug-x86-Linux-2.1/dedicated/qcommon/cm_trace.c:1170 start=0x14 is fucked (gdb) frame #7 0x459bc9a1 in ClientDamage (clent=0x47841e3c, entnum=3, enemynum=74, id=182) at debug-x86-Linux-2.1/game/game/g_cmds.c:1970 1970 in debug-x86-Linux-2.1/game/game/g_cmds.c (gdb) print enemy->client->ps.origin Cannot access memory at address 0x14. WolfMPTest branch Changelog ========== END 2001-09-20 Timothee Besset First native linux Wolf client ran on: Thu Sep 20 09:53:38 CEST 2001 some crashes because of my tr_animation.c stuff, need to use Graeme's .. patching from his info cleanup of cons script, building debug and release merging again against latest SOS recompiling and checking back in? NOOO .. the protocol has changed .. tagging the WolfMP test compatible as WolfMPTest no conflict M src/cgame/cg_event.c no conflict M src/cgame/cg_main.c no conflict M src/cgame/cg_newDraw.c no conflict M src/cgame/cg_players.c no conflict M src/cgame/cg_weapons.c no conflict M src/client/cl_main.c M src/game/bg_animation.c M src/game/bg_misc.c M src/game/bg_pmove.c M src/game/bg_public.h M src/game/g_client.c M src/game/g_cmds.c M src/game/g_missile.c M src/game/g_session.c M src/game/g_weapon.c M src/game/q_shared.h no conflict M src/qcommon/msg.c M src/qcommon/qcommon.h no conflict M src/ui/ui_players.c 2001-09-19 Timothee Besset Merging new SOS stuff in my CVS, modified files are: M src/cgame/cg_main.c M src/cgame/cg_players.c M src/cgame/cg_scoreboard.c M src/cgame/cg_trails.c M src/client/cl_main.c merged M src/client/snd_public.h M src/extractfuncs/extractfuncs.c merged M src/game/bg_misc.c merged M src/game/g_client.c merged M src/game/g_cmds.c merged M src/game/g_misc.c merged M src/game/g_weapon.c merged M src/qcommon/common.c M src/qcommon/unzip.c M src/server/server.h merged M src/server/sv_client.c merged M src/server/sv_main.c M src/ui/ui_local.h M src/ui/ui_main.c M src/ui/ui_shared.c M src/ui/ui_shared.h Also need to add on Vendor side the files I added in SOS those files might have been modified by Id since then too! merged unix/unix_net.c merged unix/unix_shared.c #1 merge the ones we know #2 do a diff merge completed fine, gonna rebuild dbg of server and restart it on zerowing unzip.c modified to use statics (OSX conflict), merged with the Q3 version / comment out unused statics to fix gcc warnings using new version tag: Wolf 0.7.16-1 ( -1 for patchlevel 1 on linux, not checked in SOS) writing cons script for full client now would need to share some stuff? botlib dependency etc.? (different targets will need different compile flags and config, but the idea is that the files need to be listed only once in our scripts) additional issue: the splines code is that going to require g++ everywhere? or g++ for splines compile gcc for regular stuff g++ for linking copied back a bunch of stuff from Q3 source using snd_dma.s again for S_WriteLinearBlastStereo16 2001-09-17 Timothee Besset Update from SOS, post MP Test from Nerve, OSX check-ins, merging everything and checking in too 2001-09-16 Timothee Besset added -fshort-enums to the flags, catches enum warnings adding a -1 valued enum in enum declarations .//botlib/be_aas_entity.c:453: // TTimo WTF?? qboolean blocking // warning: comparison is always false due to limited range of data type // if (VectorCompare( absmin, absmax ) && blocking < 0) { if (VectorCompare( absmin, absmax )) { game/ai_cast_characters.c:1117 // TTimo gcc: warning: comparison is always false due to limited range of data type // initial line: if (trap_GetTag( ent->s.number, painTagNames[0], &or ) < 0) if (!trap_GetTag( ent->s.number, painTagNames[0], &or )) { game/ai_cast_characters.c:1128 // TTimo gcc: warning: comparison is always true due to limited range of data type // initial line: if (trap_GetTag( ent->s.number, painTagNames[tagIndex], &or ) >= 0) if (trap_GetTag( ent->s.number, painTagNames[tagIndex], &or )) { 2001-09-15 Timothee Besset cons-based building system is now equivalent to existing one based on "regular makefiles". not gonna use it for MP test builds though, don't want to take the risk. final MP Test source and media is ready updating the code and building final binaries merging with recent SOS changes new media! adding wolfconfig.cfg to the distribution (vm_game 0 etc.) there is a potential problem with the mp_beachtest.script it doesn't seem to be broadcasted to clients 0.7.15 win32 release doesn't use Id's SOS source tree the linux dedicated got built upon. Patched qcommon/files.c to let .script files go through FS_FOpenFileRead outside of the directory (fs_restrict 1 test) 2001-09-14 Timothee Besset updated from SOS, this is supposed to be the source used for the MP test release merging in with my CVS tree NOTE: CVS conflicts solving, only looking at the diff if the file is in the list of modified files from SOS some of Zaphod's fixes overlap my stuff (OSX builds with gcc) nothing in botai nothing in botlib client: cl_cgame.c cl_parse.c not affecting my changes game: ok M src/game/ai_cast_script_actions.c ok M src/game/bg_animation.c ok M src/game/bg_misc.c ok M src/game/bg_pmove.c ok M src/game/bg_public.h ok M src/game/g_active.c ok M src/game/g_client.c ok M src/game/g_cmds.c ok (no diff) M src/game/g_local.h ok M src/game/g_mover.c ok M src/game/g_script.c ok M src/game/g_tramcar.c ok M src/game/g_weapon.c ok M src/game/q_shared.h nothing in qcommon in server: ok (no diff) sv_init.c 2001-09-13 Timothee Besset patched the build system and pak checksum to run in demo mode 2001-09-11 Timothee Besset setting -Werror back on cleaning up gcc warnings some non-benign things: ai_cast_think.c:739: warning: left-hand operand of comma expression has no effect // TTimo gcc: left-hand operand of comma expression has no effect // initial line was: for (i = 0; i < aicast_maxclients, clCount < level.numPlayingClients; i++, ent++) for (i = 0; ( i < aicast_maxclients ) && ( clCount < level.numPlayingClients ) ; i++, ent++) same on ai_cast_think.c:869 g_items.c:1000: warning: left-hand operand of comma expression has no effect now builds completely without warnings next: build with cons? started putting up cons scripts, saw the major issues, no functional build system yet. have to switch to compile new binaries for Id testing this afternoon. 2001-09-08 Timothee Besset building dedicated server: - game/bg_public.h uses key_t conflicts with /usr/include/sys/types.h:122: previous declaration of `key_t' renamed to bg_key_t - some FS_ReadFile casting to (void **) - stricmp usage is banned for portability reasons, use Q_stricmp - pak checkum with unsigned ints (gcc warning) - md4.c guarded #pragma between #ifdef WIN32 - net_chan.c comment out Netchan_ScramblePacket (defined but not used) - added () around TFL_DEFAULT #define - fixed some %d into %ld (warning: int format, long int arg (arg 3)) - moved the key completion stuff from cl_key.c to common.c - be_aas_routetable.c is new - suspicious be_aas_routetable 954, while (trav = next) - SV_GetTag in server/sv_game.c - new function Sys_DefaultBasePath in win_main.c used in Cvar_Get "fs_basepath" default in files.c (filesystem init) Q3 uses Sys_DefaultInstallPath instead, stubs to Sys_Cwd as Sys_DefaultBasePath does for win32 adding equivalent Sys_DefaultBasePath to unix_shared.c building game dll: - game compilation defines GAMEDLL - duplicate trunk_states in g_tramcar, named plane_states correctly - unresolved BoxOnPlaneSide after looking at Q3 qagame source, the symbol is not used but it's extracted from g_funcs.h and g_func_decs.h patched the two files to solve the problem - other unresolved: undefined symbol: strlwr (debugi386-glibc/baseq3/qagamei386.so) undefined symbol: stricmp (debugi386-glibc/baseq3/qagamei386.so) use of strlw and stricmp should be PROHIBITED for portability reasons, use the provided Q_strlwr and Q_stricmp instead undefined symbol: max (debugi386-glibc/baseq3/qagamei386.so) undefined symbol: min (debugi386-glibc/baseq3/qagamei386.so) added #define in q_shared.h the very first linux RTCW dedicated server was born on Mon Sep 10 9:45:00 CEST 2001 ran a few minutes, checked that I could join the game from win32 client since the g_func.h and g_func_dec.h had to be patched, gonna use linux_ prefixed one and maintain by hand