<Simon|B> oh ok2009-08-23 17:36:57
 I have the newest one and it worked2009-08-23 17:37:09
* Shelwien thinks that google should make an online browser2009-08-23 17:37:41
 they have one or what do you mean?2009-08-23 17:38:17
<Shelwien> i mean like their pdf viewer etc2009-08-23 17:38:29
 or google docs2009-08-23 17:38:33
 there're too many incompatible browsers2009-08-23 17:38:48
<Simon|B> `you mean a browser in a browser? :-D2009-08-23 17:39:03
<Shelwien> yeah2009-08-23 17:39:09
 in fact i think that adobe could do it2009-08-23 17:39:19
<chornobyl> so you can browse while you browse2009-08-23 17:39:22
<Simon|B> but where is the difference?2009-08-23 17:39:33
<Shelwien> if they'd just build a html renderer into flash plugin2009-08-23 17:39:55
<Simon|B> `only hardware problems could get passed2009-08-23 17:39:56
<Shelwien> they could capture the whole market eventually2009-08-23 17:40:11
 as flash is the same everywhere2009-08-23 17:40:20
 and having the same browser engine everywhere would be convenient for designers too2009-08-23 17:40:57
<chornobyl> not realy2009-08-23 17:40:58
<Shelwien> well, its the same or doesn't work2009-08-23 17:41:13
<chornobyl> what abut flash under *nix2009-08-23 17:41:24
<Shelwien> dunno really, but it should be there somewhere2009-08-23 17:41:39
<Simon|B> I dislike flash and hope it soon will be used much less2009-08-23 17:42:01
<chornobyl> its crippled a lot2009-08-23 17:42:01
<Shelwien> reverse-engineered too or what?2009-08-23 17:42:31
<Simon|B> because of new html tags2009-08-23 17:42:34
<Shelwien> hmm looks like there's an official one...2009-08-23 17:43:23
 and i don't like flash either2009-08-23 17:44:58
<chornobyl> it http://www.linux.org.ru/view-message.jsp?msgid=36542672009-08-23 17:45:09
 small example2009-08-23 17:45:16
<Shelwien> but imagine what'd happen if somebody tried to extend the main browser engine with all the same features2009-08-23 17:45:34
 like this we have youtube working at least2009-08-23 17:45:51
 and some services with graphs etc2009-08-23 17:45:57
<osman> hi guys2009-08-23 17:47:09
<Shelwien> hi osman ;)2009-08-23 17:47:15
<osman> did anyone realize new ccm source?2009-08-23 17:47:27
<chornobyl> maybe it would be possible to display same page in different browsers2009-08-23 17:47:28
<Shelwien> as i said, currently the easiest way to do that would be probably to add html rendering into flash2009-08-23 17:48:20
<Simon|B> http://thevideobay.org/ heard of this?2009-08-23 17:48:31
<Shelwien> osman: i worked on comptracer update instead, no new ccm from me2009-08-23 17:48:42
<Simon|B> it maybe will be the first video portal using the video tag2009-08-23 17:48:49
<osman> @shelwien: i have cleaned it completely2009-08-23 17:49:12
<Shelwien> i've seen v5, yeah2009-08-23 17:49:24
<osman> yep.2009-08-23 17:49:32
<chornobyl> youtube has it experimentally2009-08-23 17:49:34
<osman> @Skymmer: could you test v5?2009-08-23 17:49:41
<Simon|B> oh really?2009-08-23 17:49:44
 how can I test it?2009-08-23 17:49:53
<osman> www.osmanturan.com/ccm_sh_v5.zip2009-08-23 17:50:10
 in Release folder there is mine compile2009-08-23 17:50:22
 in main folder there is ccm_sh which is ccm_sh0 (or something) from Shelwien2009-08-23 17:50:49
 i think there is no reason for christian to kill us :)2009-08-23 17:51:34
 i wonder what compressor would be next for decompilation? :)2009-08-23 17:52:57
 ppmonstr -> ccm -> ?2009-08-23 17:53:05
*** Simon|B has left the channel2009-08-23 17:53:47
 as to "browser in browser" capability, seems we should give a chance to silverlight :)2009-08-23 17:53:59
*** Simon|B has joined the channel2009-08-23 17:54:03
<Shelwien> 1. we didn't really finish with ccm. there're also filters and the idea was to beat the original in both speed and compression2009-08-23 17:54:30
<osman> i don't think adding our E8 and delta filter would be troublesome. we cracked main modeling part which is most interesting part IMO.2009-08-23 17:55:34
<Shelwien> 2. i wasn't that interested in ccm myself, but the question about its decompilation appeared a few times when i mentioned deppm2009-08-23 17:55:45
 well, E8 yeah, but delta filter might be troublesome2009-08-23 17:56:15
 also ccm_sh currently works via getc/putc2009-08-23 17:56:33
<osman> i really don't care it's filter2009-08-23 17:56:34
 really2009-08-23 17:56:36
<Shelwien> well, it really affects the compression2009-08-23 17:56:57
 and its kinda interesting what kind of deltas are there, and what kind of detection2009-08-23 17:57:18
 anyway, i was thinking about just writing a delta filter of my own2009-08-23 17:57:48
<osman> yep. but as i said before main scene is not secret anymore. and this is really good for me2009-08-23 17:57:51
<Shelwien> but atm i don't see a way to make a good sequential delta filter2009-08-23 17:58:14
 the main problem is formats with 16bit+ values2009-08-23 17:58:53
 ...2009-08-23 17:59:37
<osman> sparse model can cope with this problem. we can especially enable sparse model on such data (it's disabled in case a match)2009-08-23 17:59:38
<Shelwien> it can't compare to delta2009-08-23 17:59:58
 on wavs etc2009-08-23 18:00:03
<osman> no no. i mean after delta2009-08-23 18:00:07
<Simon|B> is anyone of you using opera?2009-08-23 18:00:21
<osman> or we can use a bit more context masks for audio only2009-08-23 18:00:24
 errr... i mean a bit more better*2009-08-23 18:00:38
 like only taking high bits account2009-08-23 18:00:59
<Shelwien> Simon: I only have IE6 and mozilla 1.7 ;)2009-08-23 18:01:08
<osman> 1111 0000 | 0000 0000 etc2009-08-23 18:01:13
 i've IE8 and chrome2009-08-23 18:01:23
 and i really like chrome2009-08-23 18:01:28
 using it everyday2009-08-23 18:01:39
<Simon|B> oh if you are still happy with them why not... ;)2009-08-23 18:01:39
<osman> :)2009-08-23 18:01:40
<Shelwien> well, why don't you just try that? ;)2009-08-23 18:01:51
 of course really a special model is required for analog data etc2009-08-23 18:02:21
<osman> btw, there are some unused spaces in data structures2009-08-23 18:02:51
 seems they are left for CCMX2009-08-23 18:03:00
<Shelwien> yeah, there's even an unused SSE3 table ;)2009-08-23 18:03:17
<osman> yep :)2009-08-23 18:03:23
 final SSE stage has 64 bit wide counters %)2009-08-23 18:03:37
<Shelwien> btw, as to your matchmodel class2009-08-23 18:03:48
<osman> actually 24 + 6 = 30 bits were used2009-08-23 18:03:57
*** Simon|B has left the channel2009-08-23 18:04:04
<Shelwien> maybe its more readable like that, but also slower2009-08-23 18:04:17
 especially things like hashmask etc, which are technically constants2009-08-23 18:04:38
*** Simon|B has joined the channel2009-08-23 18:04:50
<osman> yeah. i'm sure it's not any faster. but, i thought, it's better to put it in a structure until someone really need to understand i2009-08-23 18:04:57
 t2009-08-23 18:04:57
<chornobyl> i got opers2009-08-23 18:05:07
 *opera2009-08-23 18:05:11
<Simon|B> Only wanted to know if video tag works for someone else but it seems as it isn't in beta of v10 (yet)2009-08-23 18:17:18
<chornobyl> you mean http://www.youtube.com/html52009-08-23 18:18:13
<Simon|B> yeah2009-08-23 18:19:48
 there were some versions earlier supporting it very good but dunno why it isn't inplemented finally :(2009-08-23 18:20:30
<chornobyl> it not work for me in ff 3.5 either2009-08-23 18:21:08
<Shelwien> btw osman2009-08-23 18:26:40
 which 13 lines?2009-08-23 18:26:45
<osman> final SSE updating. i've finished it already2009-08-23 18:27:28
 http://pastebin.com/m6f0aedac2009-08-23 18:27:33
* Shelwien wonders what was the maximum static allocation in PE2009-08-23 18:31:23
<Shelwien> but anyway i'm going to make a template class out of the whole ccm model2009-08-23 18:32:48
 then there'd be no need for malloc and the like2009-08-23 18:33:06
*** Skymmer has left the channel2009-08-23 18:46:55
*** Skymmer has joined the channel2009-08-23 18:56:02
<Skymmer> osman, your last decmm is included in test results. Now i'll provide link to txt cause file became too large to paste it here2009-08-23 18:57:10
 http://skymmer.narod.ru/data/deccm_speed.txt2009-08-23 18:57:14
 Crap! Sorry2009-08-23 18:57:41
 Your results are not in there2009-08-23 18:57:50
<osman> ok. no problem2009-08-23 18:58:32
<Skymmer> Check now2009-08-23 18:58:47
<Shelwien> 4042009-08-23 18:59:01
<Skymmer> wwhoof.. and now ?2009-08-23 18:59:33
<Shelwien> ok2009-08-23 18:59:41
<Skymmer> fine.2009-08-23 18:59:47
<Shelwien> there were also osmans v4 and v52009-08-23 19:00:14
<Skymmer> The results there are for 52009-08-23 19:00:54
 I forgot to change pasted text. wait a minute2009-08-23 19:01:10
<osman> hehe2009-08-23 19:01:21
 i wonder what happens if you would try to clean ccm source :P2009-08-23 19:01:38
<Skymmer> don't wanna to think about :))2009-08-23 19:02:16
<Shelwien> osman: if you like doing that, you can try cleaning deppm next ;)2009-08-23 19:02:26
<osman> errr...i can't do that2009-08-23 19:02:55
 because at ccm i can guess what's going on2009-08-23 19:03:05
<Skymmer> now should be corrected.2009-08-23 19:03:13
<Shelwien> deppm is simpler in fact2009-08-23 19:03:17
<osman> but, i don't have any experience in ppm2009-08-23 19:03:20
 so, i can get stuck frequently2009-08-23 19:03:33
<Skymmer> I think there is no reason to test 3 and 4 ?2009-08-23 19:03:38
<Shelwien> not really i guess2009-08-23 19:03:54
<Skymmer> OK.2009-08-23 19:04:03
<Shelwien> i can try properly compiling osmans v5 though2009-08-23 19:04:08
 just for later reference2009-08-23 19:04:13
<Skymmer> as you wish2009-08-23 19:04:20
<osman> @shelwien: it would be interesting to see properly optimized v52009-08-23 19:04:47
 @skymmer: thanks for testing2009-08-23 19:06:12
<Shelwien> http://shelwien.googlepages.com/ccm_sh_v5_ic.exe2009-08-23 19:07:07
<osman> @skymmer: could you test it :)2009-08-23 19:07:38
*** Skymmer has left the channel2009-08-23 19:11:31
*** Skymmer has joined the channel2009-08-23 19:19:42
<Skymmer> Results updated.2009-08-23 19:23:09
*** Skymmer has left the channel2009-08-23 19:27:58
<Surfer> http://www.youtube.com/html5 works only in chrome :)2009-08-23 19:42:45
*** Shelwien has left the channel2009-08-23 19:42:46
 safari not supported2009-08-23 19:42:57
<chornobyl> àà,2009-08-23 19:43:04
 ff?2009-08-23 19:43:08
<Simon|B> but note original ccm build isn't IC ;)2009-08-23 19:46:22
*** toffer has left the channel2009-08-23 19:52:04
*** Shelwien has joined the channel2009-08-23 20:04:42
*** toffer has joined the channel2009-08-23 20:05:40
<Shelwien> original build not IC - is that bad or good? ;)2009-08-23 20:06:07
 supposedly its gcc 4.32009-08-23 20:06:19
<Simon|B> it means only it is still faster ;)2009-08-23 20:06:24
<Shelwien> ah, that's because of osman ;)2009-08-23 20:06:36
 though not really ;)2009-08-23 20:06:54
<osman> hey! :)2009-08-23 20:06:59
<Shelwien> in fact, the original uses rc with memory i/o 2009-08-23 20:07:23
 and ccm_sh still works with getc/putc2009-08-23 20:07:38
 ...guess I'd fix that today ;)2009-08-23 20:08:00
<Simon|B> ok we will see^^2009-08-23 20:10:17
*** chornobyl has left the channel2009-08-23 20:10:25
<Shelwien> in fact a properly rewritten ccm_sh _should_ be faster than original2009-08-23 20:10:46
 just because it doesn't have the filters2009-08-23 20:11:07
 so does less processing2009-08-23 20:11:13
 also, I guess I'd compile it with gcc 4.3/4.4 too next time2009-08-23 20:11:56
* Shelwien returns to adding export forwards to comptrace2009-08-23 20:19:26
<Surfer> 662009-08-23 20:32:24
*** Surfer has left the channel2009-08-23 20:32:29
<Shelwien> huh? looks like i made it or something...2009-08-23 20:57:20
 ah... damn... some of them have to be hooked too...2009-08-23 21:00:47
 ok. anybody alive with vista?2009-08-23 21:26:34
 osman?2009-08-23 21:27:09
<Simon|B> I could have vista yeah^^2009-08-23 21:27:22
<Shelwien> i want to test the new comptrace2009-08-23 21:27:41
<Simon|B> what has to be done?2009-08-23 21:28:00
<Shelwien> a moment2009-08-23 21:28:18
 please get that - http://shelwien.googlepages.com/comptrace_v1.rar2009-08-23 21:29:35
 and run the test.bat2009-08-23 21:29:38
 if it goes well, you should get some trace*.log files around2009-08-23 21:30:00
 v0 just crashed on vista though2009-08-23 21:30:21
* Shelwien probably would have to try running that win7 VM if it won't work2009-08-23 21:31:16
<Simon|B> ok2009-08-23 21:31:43
 have to boot my laptop2009-08-23 21:32:11
* Shelwien thinks that its pretty cool even if it works only in XP - it compiles a DLL specifically for the test platform ;)2009-08-23 21:34:00
 addimports crashes but I still have some logs2009-08-23 21:34:48
<Shelwien> huh?!2009-08-23 21:35:07
 weird ...2009-08-23 21:35:16
<Simon|B> works on XP2009-08-23 21:35:49
<Shelwien> sure2009-08-23 21:35:56
 are logs on vista the same?2009-08-23 21:36:06
<Simon|B> mom2009-08-23 21:36:10
 hehe no2009-08-23 21:36:48
 only two brackets2009-08-23 21:36:56
<Shelwien> not interesting then2009-08-23 21:37:09
 but addimports?!2009-08-23 21:37:22
 can you test it manually?2009-08-23 21:37:47
 by running addimports in the src directory?2009-08-23 21:38:11
<Simon|B> sure2009-08-23 21:38:57
<Shelwien> ah, one more thing2009-08-23 21:39:16
<Simon|B> do you want a msvc dump?2009-08-23 21:39:19
<Shelwien> no2009-08-23 21:39:23
<Simon|B> prette helpful in my eyes^^2009-08-23 21:39:43
 *pretty2009-08-23 21:39:48
<Shelwien> it would also want a kernel32.dll renamed into mernel32.dll2009-08-23 21:39:50
<Simon|B> yes crashes also on XP if I start it alone2009-08-23 21:40:48
<Shelwien> as i saids thats different probably2009-08-23 21:41:06
 it opens mernel32.dll and trace.dll2009-08-23 21:41:22
 and would crash without either one2009-08-23 21:41:35
 well, i guess its still some progress2009-08-23 21:43:36
 it didn't create the trace.log before ;)2009-08-23 21:43:46
<Simon|B> ah wait2009-08-23 21:44:08
 there is no trace.log only the others2009-08-23 21:44:30
<Shelwien> i guess it didn't work at all then2009-08-23 21:44:51
<Simon|B> yeah maybe I only wondered because there were still progress after the crash2009-08-23 21:45:22
<Shelwien> by chance, maybe you know if there's any speed different between vmware player and full vmware?2009-08-23 21:48:13
<Simon|B> no I don't really know but don't think so2009-08-23 21:49:24
 I mostly used server without problems2009-08-23 21:49:38
 The progress was only the ccm part...2009-08-23 21:50:00
<Shelwien> ?2009-08-23 21:50:30
<Simon|B> it crashes on the beginning and then there is the ccm print at the end2009-08-23 21:51:24
<Shelwien> well, i guess if there's no kernel32.dll in the ccm.exe.local, then ccm would normally work...2009-08-23 21:51:42
<Simon|B> r_bytes=27677 00350000 00000001 0000086C 12009-08-23 21:52:06
 w_bytes=124222009-08-23 21:52:06
 memused=544256k2009-08-23 21:52:06
 thats in XP and not on vista2009-08-23 21:52:20
<Shelwien> on XP its ok2009-08-23 21:52:21
 but in vista kernel32.dll is different2009-08-23 21:52:38
<Simon|B> but shouldbn't be helpful I guess :D2009-08-23 21:52:46
<Shelwien> and apparently addimports crashed on it somehow2009-08-23 21:53:01
 well, if confirmed that it still doesn't work2009-08-23 21:53:37
 so i'm now setting up the win7 VM to test it myself ;)2009-08-23 21:53:57
<Simon|B> hehe yeah win7 is also on my computer but don't think it still works^^2009-08-23 21:55:29
<Shelwien> interesting... apparently full vmware requires a reboot to work2009-08-23 21:55:40
 player worked right away...2009-08-23 21:55:51
<osman> hi everyone2009-08-23 22:16:20
 seems i was late for the vista test2009-08-23 22:16:34
<Shelwien> ...2009-08-23 22:16:35
 not really, you can try too ;)2009-08-23 22:16:45
<osman> btw, i really wonder all the time why you refuse to use "hooking"?2009-08-23 22:17:01
<Shelwien> i don't refuse2009-08-23 22:17:14
<osman> it's much more easier i think2009-08-23 22:17:22
<Shelwien> its just that this specific project was made specifically to evalulate the .local method2009-08-23 22:17:47
<osman> i should not i'm not talking about "patching" executable2009-08-23 22:17:49
 *i should note that i'm not...2009-08-23 22:18:00
<Shelwien> i don't like detours2009-08-23 22:18:15
<osman> it just works :)2009-08-23 22:18:24
 and you don't need detour library actually2009-08-23 22:18:37
<Shelwien> thats exactly what i don't like ;)2009-08-23 22:18:39
<osman> it's just a wrapper2009-08-23 22:18:43
<Shelwien> its better to use non-standard methods for hacking2009-08-23 22:22:35
 because protections often check for standard methods2009-08-23 22:23:07
 and detours is easy enough to detect imho2009-08-23 22:23:53
<Simon|B> good night2009-08-23 23:04:37
*** Simon|B has left the channel2009-08-23 23:04:40
*** sami has left the channel2009-08-23 23:08:31
*** toffer has left the channel2009-08-24 00:12:38
<Shelwien> osman?2009-08-24 00:19:39
*** osman has left the channel2009-08-24 00:37:06
*** Skymmer has joined the channel2009-08-24 00:44:08
*** Skymmer has left the channel2009-08-24 00:46:24
 ok, i dunno anymore2009-08-24 01:24:04
 Simon's crash apparently happened due to array overflow %)2009-08-24 01:24:54
 because I didn't expect that there'd be more than 1024 exports in kernel32.dll ;)2009-08-24 01:25:13
 but after i fixed that, it still doesn't work in win72009-08-24 01:25:58
 says something about GetTickCount not found in kernel32.dll ;)2009-08-24 01:27:14
 but I'd appreciate if somebody could test it in vista once more2009-08-24 01:27:37
 http://shelwien.googlepages.com/ccm_sh1d.rar2009-08-24 05:05:30
*** pinc has joined the channel2009-08-24 05:49:36
 http://shelwien.googlepages.com/results.txt2009-08-24 05:54:48
 !next2009-08-24 06:03:05