Entropy Report post Posted May 23, 2007 This is an exe only update for the RC2 which can be found here: http://www.eternal-lands.com/forum/index.php?showtopic=34304 So please download that, then download this http://el.beplacid.com/downloads/el.zip and unzip it in the directory where you had RC2, overwritting the old exe. If you don't know what that means, please do not do it, or else you risk getting baned from the test server and from the forums. Linux users, just compile the latest CVS, there is no need for extra data files besides for the RC2. Please report any bugs here, and also report the message you get regarding the support for the future version. Share this post Link to post Share on other sites
ttlanhil Report post Posted May 23, 2007 Importing from the RC2 thread, since it isn't fixed yet... well, I had a look, but didn't see where the previous details were, so... I think I've figured out the cause of the fighting crash... trying to do an update of an actor with a timedelta using maths that equates to 0.0f/1000.0f ... which may be 0.0f, or may even be negative (yay for floating point) anyone who has experienced this is encouraged to update CVS and see if you still crash (if there's much more in the way of bug fixing, we may have to have another RC soon for wide-spread testing) ed: or not. darn googling showed me something else of interest, this is not a new problem (4th Dec) if someone familiar with cal3d can look at it, it'd be great... failing that, maybe someone cluey about opengl could look at the cal3d code, to see if we're causing trouble with our opengl calls before the cal3d call... other than that, dunno, because the data otherwise all seems valid ed: the heck? the fix I applied didn't get used correctly... arg... Share this post Link to post Share on other sites
CelticLady Report post Posted May 23, 2007 [18:15:42] Your graphic card/driver don't support the minimumrequirements for the next el release. Please upgrade your driver. If this don't help, you need a better graphic card. [18:16:38] Video card: Intel 915G [18:16:38] Vendor ID: Intel [18:16:38] OpenGL Version: 1.4.0 - Build 4.14.10.3971 I'll see if a newer driver is available. I've never upgraded a driver before so I'll cross my fingers I don't break something I still have blackouts in Idaloran. Didn't test anything else yet. Share this post Link to post Share on other sites
Entropy Report post Posted May 23, 2007 ed: the heck? the fix I applied didn't get used correctly... arg... What do you mean? Share this post Link to post Share on other sites
Tanyia Report post Posted May 23, 2007 Continued from RC2... Karen: The LOD still requires me to relog before taking effect. I'm using RC3. Just a note...sorry, I can't test CVS. I can only get it to compile on one computer, which currently can't handle EL for some crazygonuts reason, and I am not willing to allocate time to look into why until Thursday night. But if there's more info I can provide you, just let me know. Share this post Link to post Share on other sites
ttlanhil Report post Posted May 23, 2007 ed: the heck? the fix I applied didn't get used correctly... arg...What do you mean?well, the times when it crashed, gdb told me that the last update time for the actor animation, and the cur_time, are the same. I figured that's likely to be a problem, so I added an if() so that it wouldn't render if the last time was the same as cur_time.except that I got the crash again, and the two were the same once more so either something is getting changed in another thread (possible, but scary), the cal3d call is altering ELC data (don't think it's possible), gdb is giving bad data (possible, but I think pretty unlikely), or something even funkier is going on (like the cal3d call continuing after it returns, threaded style, and crashing several steps later in the EL code) I'm still looking into it, since I still crash, but meh... Share this post Link to post Share on other sites
Entropy Report post Posted May 23, 2007 Any way to reliably reproduce the bug? Share this post Link to post Share on other sites
ttlanhil Report post Posted May 23, 2007 (edited) Any way to reliably reproduce the bug?not that I know of. I get it maybe once or twice an hour fighting ogres in melinis also, I just added some code to force the given value to 0.0f, and it didn't crash... yet despite it being 'impossible' for it to be 0.0f otherwise, it did when it crashed... so it's not a direct cause. it may be a coincidence, related, or a symptom when tracing through data in gdb, I got the following: (gdb) print act $1 = (actor *) 0x861ae40 (gdb) print *act $2 = {actor_id = 3091, actor_type = 27, tmp = {have_tmp = 1, x_pos = 45, y_pos = 246, z_pos = 0, x_tile_pos = 90, y_tile_pos = 492, x_rot = 0, y_rot = 0, z_rot = 315}, calmodel = 0x84d9220, cur_anim = {anim_index = 3, kind = 1, duration = 0.317431152, duration_scale = 0.634862304}, cur_anim_sound_cookie = 0, cur_idle_anims = {{ anim_index = 0, kind = 0, duration = 0, duration_scale = 0} <repeats 16 times>}, IsOnIdle = 0, anim_time = 0, last_anim_update = 3449173, x_pos = 45, y_pos = 246, z_pos = 0, scale = 1, x_tile_pos = 90, y_tile_pos = 492, x_rot = 0, y_rot = 0, z_rot = 315, max_z = 1.78974867, boots = 0, hair = 0, skin = 0, pants = 0, shirt = 0, cur_weapon = 0, cur_shield = 0, is_enhanced_model = 0, body_parts = 0x0, remapped_colors = 0 '\0', texture_id = 287, skin_name = '\0' <repeats 255 times>, actor_name = "Male Ogre", '\0' <repeats 246 times>, que = {attack_up_4, pain1, nothing <repeats 19 times>}, last_command = 5 '\005', busy = 1 '\001', sitting = 0 '\0', fighting = 1 '\001', move_x_speed = 0, move_y_speed = 0, move_z_speed = 0, movement_frames_left = 0, rotate_x_speed = 0, rotate_y_speed = 0, rotate_z_speed = -0, rotate_frames_left = 0, after_move_frames_left = 0, moving = 0 '\0', rotating = 0 '\0', stop_animation = 1 '\001', stand_idle = 0 '\0', sit_idle = 0 '\0', dead = 0 '\0', damage = 10, damage_ms = 960, last_health_loss = 3448441, cur_health = 39, max_health = 105, ghost = 0 '\0', has_alpha = 0 '\0', kind_of_actor = 15, buffs = 0, current_displayed_text = '\0' <repeats 59 times>, current_displayed_text_time_left = 0, x_speed = 0, y_speed = 0, z_speed = 0, async_fighting = 1, async_x_tile_pos = 90, async_y_tile_pos = 492, async_z_rot = 315} (gdb) print *act->calmodel $3 = {m_pCoreModel = 0x603d1e0, m_pSkeleton = 0x967d440, m_pMixer = 0x830b2d0, m_pMorphTargetMixer = 0x8acc720, m_pPhysique = 0x818c470, m_pSpringSystem = 0x9ff6080, m_pRenderer = 0x9642690, m_userData = 0x0, m_vectorMesh = {<std::_Vector_base<CalMesh*,std::allocator<CalMesh*> >> = { _M_impl = {<std::allocator<CalMesh*>> = {<__gnu_cxx::new_allocator<CalMesh*>> = {<No data fields>}, <No data fields>}, _M_start = 0x858ad20, _M_finish = 0x858ad28, _M_end_of_storage = 0x858ad28}}, <No data fields>}, m_boundingBox = {plane = {{a = 8, b = 0.072915636, c = 0.22441116, d = 8}, {a = -0.0729156435, b = 0.13869366, c = 8, d = -0.190895602}, {a = 0.0729156137, b = -8, c = -0.224411175, d = -0.0729156137}, {a = -8, b = -0.224411175, c = -0.190895557, d = -8}, {a = -0.138693735, b = -0.235959873, c = -8, d = 1.03141344e-08}, {a = -0.190895557, b = -8, c = 0.13869372, d = -0.0729155689}}}} (gdb) print (*act->calmodel)->m_boundingBox $9 = {plane = {{a = 8, b = 0.072915636, c = 0.22441116, d = 8}, {a = -0.0729156435, b = 0.13869366, c = 8, d = -0.190895602}, {a = 0.0729156137, b = -8, c = -0.224411175, d = -0.0729156137}, {a = -8, b = -0.224411175, c = -0.190895557, d = -8}, {a = -0.138693735, b = -0.235959873, c = -8, d = 1.03141344e-08}, {a = -0.190895557, b = -8, c = 0.13869372, d = -0.0729155689}}} the only thing of interest that I see is that m_vectorMesh appears to be empty ed: I'm still working at it... maybe if I get a few backtraces to compare I might see a similarity (eg certain actor commands in the queue or something funky like that) Edited May 23, 2007 by ttlanhil Share this post Link to post Share on other sites
Entropy Report post Posted May 23, 2007 Did you check your errors log? Maybe some mesh didn't load properly or something.. Share this post Link to post Share on other sites
Kazoo Report post Posted May 23, 2007 [23:53:48] Please update your graphic card driver! [23:53:48] Your graphic card supports the absolute minumin requirements for the next el release, but don't expect that you can use all features. [23:53:48] Point Particles disabled. [23:53:48] Framebuffer disabled (need newer driver) [23:54:46] Video card: Intel 915GM [23:54:46] Vendor ID: Intel [23:54:46] OpenGL Version: 1.4.0 - Build 7.14.10.4764 I have the newest drivers installed, and I also get the blackouts in Idaloran and Iscalrith storage. Share this post Link to post Share on other sites
KarenRei Report post Posted May 23, 2007 Tanyia: I'll take another look at that. Surprised that my changes didn't do it for you. Anybody with blackouts use CVS? Share this post Link to post Share on other sites
ttlanhil Report post Posted May 23, 2007 (edited) Did you check your errors log? Maybe some mesh didn't load properly or something..`grep -in ogre ~/.elc/error_log.txt` gives no hits.after stripping timestamps, there's still 1708 unique lines in it, a lot of errors... in 5 places, there's reports of an opengl error(though I'm not sure just how old that is, and it's in 4 different files. including mapwin)... other than that, nothing I can see of interest to this problem on the up(?) side, I just went over an hour on ogres without a crash maybe my new debugging code is scaring the bug away (lots and lots of printf()ing, so it will change the timings) Edited May 23, 2007 by ttlanhil Share this post Link to post Share on other sites
Entropy Report post Posted May 23, 2007 Are you sure it's only the ogres? And is it only when fighting? Do you have to be fighting them to crash, or anyone fighting them crashes it? Share this post Link to post Share on other sites
KarenRei Report post Posted May 23, 2007 Tanyia: Which option are you changing -- I get instant response by everything except for light columns threshold (this is due to the fact that light columns switch their state only after the framerate significantly differs from the threshold so that you don't get a bunch of flickering when the framerate is on the border). If this is a problem, I can make it force an update when you change the threshold. Share this post Link to post Share on other sites
ttlanhil Report post Posted May 23, 2007 Are you sure it's only the ogres? And is it only when fighting? Do you have to be fighting them to crash, or anyone fighting them crashes it?that's the only thing I've experienced the crash on myself... but ogres are what I train on, so I wouldn't expect that I'd see an occasional problem with any other fighting.so far, it's always been when I was fighting them. I'm unsure if there have always been other people fighting nearby when I get crashes... Share this post Link to post Share on other sites
Entropy Report post Posted May 23, 2007 Well, if you want, I can put a low level char of mine in god mode to fight them forever, and you can idle in there waiting for a crash, and experiment different things. So let me know if that helps. Share this post Link to post Share on other sites
RallosZek Report post Posted May 23, 2007 [18:16:38] Video card: Intel 915G [18:16:38] Vendor ID: Intel [18:16:38] OpenGL Version: 1.4.0 - Build 4.14.10.3971 [23:54:46] Video card: Intel 915GM [23:54:46] Vendor ID: Intel [23:54:46] OpenGL Version: 1.4.0 - Build 7.14.10.4764 I see a pattern Especially since my card is older, but has higher OpenGL version: [13:57:49] Video card: GeForce4 4200 Go/AGP/SSE2 [13:57:49] Vendor ID: NVIDIA Corporation [13:57:49] OpenGL Version: 1.5.8 NVIDIA 96.31 I also don't have framebuffers, by the way and not experiencing the blackouts. My intel chipset is 82801DB/L/M, which has ICH4 audio and the 915 has ICH6 HDA, so defenitely older. Share this post Link to post Share on other sites
Segor Report post Posted May 23, 2007 I'm reading engineering books and noticed that when i finish reading the books they hightligh in white in the knowledge section (normally) but when i restart my client that reverses, like i haven't read them. I have read so far blackpowder building caltrops building rope building nails building I haven't seen this reported yet, or maybe i didn't search hard enough... http://img504.imageshack.us/img504/1667/el076lg4.jpg Share this post Link to post Share on other sites
2coolfool Report post Posted May 23, 2007 (edited) Using these specs : [15:16:41] Video card: RADEON 9200 Series DDR x86/SSE2[15:16:41] Vendor ID: ATI Technologies Inc. [15:16:41] OpenGL Version: 1.3.1072 WinXP Release i got [15:10:06] Please update your graphic card driver![15:10:06] Your graphic card supports the absolute minumin requirements for the next el release, but don't expect that you can use all features. didn't note any "blackouts", went to idaloran and iscalrith storages. edit: typos.. Edited May 23, 2007 by 2coolfool Share this post Link to post Share on other sites
dejan Report post Posted May 23, 2007 I have ecountered bug in RC2 and RC3 - in 800x600 mode, when you turn digital clock on, than HUD stats become invisible. Share this post Link to post Share on other sites
Beaverhunter Report post Posted May 23, 2007 (edited) I'm reading engineering books and noticed that when i finish reading the books they hightligh in white in the knowledge section (normally) but when i restart my client that reverses, like i haven't read them. I have read so far blackpowder building caltrops building rope building nails building I haven't seen this reported yet, or maybe i didn't search hard enough... http://img504.imageshack.us/img504/1667/el076lg4.jpg Are you really sure that you read them ON the TEST server? If you read them on main server it's no good on test server until ent decides to update the data of test server. Note also on MAIN server you won't see that you have read anything either of the new books since the knowledge.lst file is not updated as the RC version of it is. Dejan: That's not a bug. With lower resolutions you have to choose, either digital clock or stat bars in hud. The window is just too small to fit all items in the hud at too low resolutions. Adding the new stat bar Engineering, forced this as items were overlapping each others. There has however been mentioned that this maybe could be fixed by replacing the digital clock onto the analog clock or something like that, but I guess that's for the devs or patch makers to try and decide. Edited May 23, 2007 by Beaverhunter Share this post Link to post Share on other sites
CelticLady Report post Posted May 23, 2007 (edited) [18:16:38] Video card: Intel 915G [18:16:38] Vendor ID: Intel [18:16:38] OpenGL Version: 1.4.0 - Build 4.14.10.3971 [23:54:46] Video card: Intel 915GM [23:54:46] Vendor ID: Intel [23:54:46] OpenGL Version: 1.4.0 - Build 7.14.10.4764 I see a pattern Especially since my card is older, but has higher OpenGL version: [13:57:49] Video card: GeForce4 4200 Go/AGP/SSE2 [13:57:49] Vendor ID: NVIDIA Corporation [13:57:49] OpenGL Version: 1.5.8 NVIDIA 96.31 I also don't have framebuffers, by the way and not experiencing the blackouts. My intel chipset is 82801DB/L/M, which has ICH4 audio and the 915 has ICH6 HDA, so defenitely older. So do I, however, it isn't specific to the 915 chipsets exclusively. Icy mentioned in the last thread that hers is an Intel 845GM, OpenGL Version: 1.3.0, Build 4.14.10.3889.(btw, the build numbers aren't exactly correct that the client is showing. Mine should show as 6.14.10.4764, and not 7.14.10.4764. The first number is for the OS it's for, 6 being XP and 7 being Vista. Probably doesn't matter anyways though.) But yeah, maybe the Intel onboard chipsets are causing the problem in some way. Anyways, I just updated my driver to the same build as Kazoo has and I get the message now that I meet the minimum requirement. It isn't night, so I can't check again for the blackout, but I suspect it is still there since Kazoo and I have almost identical video. EDIT: I just checked my knowledge list and I do show as having read the 3 engineering books that I previously had read. So I'm not seeing an issue like temospena did about the books. Edited May 23, 2007 by CelticLady Share this post Link to post Share on other sites
Entropy Report post Posted May 23, 2007 Ok, please let me know about the blackness, as soon as you find out. if it's just a drivers issue, that's good, no more work for us. Share this post Link to post Share on other sites
RallosZek Report post Posted May 23, 2007 Linux users, just compile the latest CVS, there is no need for extra data files besides for the RC2. Please report any bugs here, and also report the message you get regarding the support for the future version. That message about the features, needs -DGL_EXTENSION_CHECK, which isn't in Makefile.linux or make.defaults as of 3 hours ago. Adding it to OPTIONS in top of Makefile.linux enables it, or in your CFLAGS env variable. [19:38:01] Your graphic card supports more than thedefault requirements for the next el release. Share this post Link to post Share on other sites
agis29 Report post Posted May 23, 2007 I downloaded the rc3 and after that i replaced the old .exe file. After a few fluffies the client crashed. Thats what i found written in errorlog.txt: Log started at 2007-05-24 20:09:51 localtime (GTB Standard Time) [20:09:52] Error: Can't open file "quest.log" [20:09:52] Error: Can't open file "custom_mirrors.lst" [20:09:52] Error: Can't open file "commands.lst" [20:10:00] Got invalid sound number Last message repeated 2 times [20:10:11] Client closed After that i disabled the sound. Around 8 minutes later the client crashed again but there was nothing new written inside errorlog.txt!! The error reporting service showed ModName cal3d.dll Share this post Link to post Share on other sites
