Advertisement
MjnMixael

Weeding Out Old Bugs

Dec 5th, 2012
176
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 12.20 KB | None | 0 0
  1. <The_E> 1114
  2. <Valathil> i've never seen shittier thruster graphics
  3. *** MjnRoaming has quit IRC: Ping timeout: 202 seconds
  4. *** MjnPhone has joined #scp
  5. *** MjnPhone has quit IRC: Client Quit
  6. <The_E> Time to weed out some old bugs
  7. <Valathil> you sure this bug is still active
  8. <The_E> Nope
  9. <Valathil> seems man thrusters are rendered before the ship
  10. <The_E> But then, I don't have data shitty enough to repro
  11. <Valathil> hehe man thrusters
  12. *** MjnMobile has quit IRC: Ping timeout: 190 seconds
  13. <Valathil> i mean they are batched
  14. <Valathil> so the would have gone after before
  15. <Valathil> but engine glows are now batched as well since the thruster distortion
  16. <Valathil> so ill take a guess and say its fixed
  17. <The_E> I would say so as well
  18. <The_E> There are quite a few zombie issues lurking amongst the "assigned" set
  19. <The_E> Looking at 1811, and I'm just going "wat? This is holding up an mvp release?"
  20. <Valathil> da fuw
  21. *** FUBAR-BDHR_ has quit IRC: Ping timeout: 202 seconds
  22. <Valathil> this isnt even a bug its coded that way that if there are lot of hits on screen that the old ones are cut off
  23. <The_E> Indeed
  24. <Valathil> id say close with extreme prejudice
  25. <The_E> I will concede that that's behaviour that should be documented and shit, but I don't exactly see the need to "fix" it
  26. <The_E> If you want a 50-second shield hit effect, you're doing it wrong anyway
  27. <Valathil> looks fucking awesome in slow motion
  28. <Valathil> you can see every energy ripple
  29. <Valathil> plasma discharges
  30. <The_E> real-time shield mesh distortion action
  31. <Valathil> stop you making me wet
  32. <The_E> Tesselate that mesh to hell and distort the crap out of it
  33. <Valathil> i came
  34. <The_E> ogod
  35. <The_E> 1616
  36. <The_E> it's a goddamn change-ship-type bug again
  37. <Valathil> aahm shouldnt swapping the ship class invalidate all texture replacements anyway?
  38. <The_E> Yes, _except_ in the special case of the original and target ship sharing the model apparently
  39. <Valathil> cause that happens so fucking often?
  40. <Valathil> there is a sexp to do tex replacement isnt there?
  41. <The_E> If I can't deal with it in half an hour, I'm closing it as No Change Required Go Die In A Fire
  42. <Valathil> FOAD
  43. *** Swifty has joined #scp
  44. <The_E> Gods
  45. <The_E> Whoever came up with the braindead idea of storing the replacement texture info behind an int*
  46. <The_E> Memory leaks and potential overruns are _such_ nice things to have
  47. <Swifty> hey I just live here.
  48. <Swifty> but I agree it's kinda derpy
  49. <The_E> Of all the things Goober criticizes, he thinks that code is rock-solid?
  50. <Valathil> http://i3.kym-cdn.com/photos/images/original/000/361/483/bdd.jpg
  51. *** Zacam-Work has joined #scp
  52. <MjnMixael> ok, i'm home
  53. <MjnMixael> lemme eat, then I shall don my mantis boots
  54. <Valathil> i read boobs
  55. <MjnMixael> till then 1153 just needs confirmation and probably closure
  56. <The_E> http://pastebin.com/K6JY4ZmY
  57. <Valathil> cant you just reattach the pointer
  58. <Valathil> why must be copied?
  59. <The_E> True, I could just do that
  60. <Valathil> thats the beauty of pointers
  61. <The_E> Okay, moving on
  62. <The_E> In keeping with the theme, look at 1775
  63. <The_E> Sure, why not use code from a codebase very few of us have seen and that is already 3000+ revisions behind trunk
  64. <Valathil> Closed: "Too old to fix"
  65. <MjnMixael> k, ima browse for bugs
  66. <MjnMixael> 2188
  67. <Swifty> closing mantis 14 was one of the satisfying moments of my career
  68. <Swifty> most*
  69. <Valathil> soft particles?
  70. <Zacam-Work> Don't worry about 1775
  71. *** m|m has quit IRC: Quit: m|m
  72. <The_E> What
  73. <The_E> What the actual fuck
  74. <The_E> 2733
  75. <The_E> Assigned to: Kazan
  76. <The_E> When did he come back to code for FSO
  77. <Zacam-Work> He didn't.
  78. <Zacam-Work> I don't know who that was intended for, but it's not him.
  79. <Valathil> http://i0.kym-cdn.com/photos/images/original/000/195/379/1320452774001.png
  80. <Zacam-Work> Wait, it WAS explicitly intended for him? ...
  81. <Zacam-Work> Right, chartering a flight to DC.
  82. <MjnMixael> were there decisions on 1114, 1616, or 1153?
  83. <MjnMixael> derp 1616 was just done
  84. <The_E> AAAAAAARGH
  85. <The_E> Gods
  86. <The_E> And now I discover that Kazan was apparently unable to follow established convention when documenting his sexps
  87. <The_E> IS THAT SO FUCKING HARD
  88. <Zacam-Work> HAHAHAHAHAHAHAHHAHAHAHA
  89. <Valathil> i think e is going super sayian any second now
  90. <Zacam-Work> Kazan, follow -ANYTHING- that somebody else did? Fuck no. Cause he always knew bestest and that was that.
  91. <MjnMixael> Zacam, can you ad 1841 on your list to check later?
  92. <Zacam-Work> If anything, dealing with Keldor is EASY compared to Kazan.
  93. <Zacam-Work> Well, I reported it, it's assigned to me, and I'm monitoring it. But yeah, I need to retest that bastard.
  94. <MjnMixael> aha, here's a possible one for you Valathil. 2693
  95. <MjnMixael> lemme confirm it first, though
  96. <Valathil> that be nice
  97. <The_E> Just look at the help text for add-nav-waypoint
  98. <The_E> Fuck it, I need some Borderlands now
  99. *** HerraTohtori has joined #scp
  100. <MjnMixael> ok, Valathil, confirmed and I've setup a new modpack that doesn't throw all kinds of debug warnings that are unrelated
  101. <Valathil> yay
  102. <Valathil> thanks ill look into it
  103. <MjnMixael> https://dl.dropbox.com/u/6681376/Collision%20Test.7z
  104. <MjnMixael> should I try to put something together for 1114, or is that just going to be closed?
  105. <Valathil> na that is 80% already fixed
  106. <MjnMixael> The_E: close 1897?
  107. <Valathil> 90% when you think about how we dont do alpha blending on billboards anymore i think its all additive
  108. <Valathil> and that means the order of rendering doesnt matter
  109. <MjnMixael> right
  110. <MjnMixael> close eeet
  111. <Valathil> DOOOOONE
  112. <Valathil> so what im i looking for in collisiontest
  113. <MjnMixael> he set $no collisions on the Fenris and so it should behave as such, however it does collide with the Ganymede. You can fly through it, which I think is expected, but not through it's subobjects
  114. <Zacam-Work> On a purely unrelated note, the universe has annoyed me with assuming that "crap" is a decent substitute just because it has quantity.
  115. <Zacam-Work> So, my headphones finally gave up the ghost. 9 hours of on-line searching later, I finally find a decent replacement pair. I get them, and then the audio jack on the media player fails. I have yet to find a decent replacement that meets my requirements.
  116. <Valathil> wasnt there a cmdline for using the old collision system?
  117. <Valathil> i wanted to test if this comes from go_even_faster
  118. <Zacam-Work> ...yes?
  119. *** Eli2_ has joined #scp
  120. <MjnMixael> yeah, it's under 'Troubleshoot'
  121. <Valathil> wrong exe selected
  122. <Valathil> i suck cocks
  123. <Zacam-Work> '-old_collision'
  124. *** Eli2 has quit IRC: Ping timeout: 190 seconds
  125. <Valathil> funny thing is you collide ship to ship with the antenna but shots go thru
  126. <Swifty> valathil, are you kidding me? We definitely do alpha blending on billboards
  127. <Swifty> it's not additive, it's premultiplied
  128. <Swifty> that's why it's order independent
  129. <Zacam-Work> (or should be)
  130. <Valathil> yeah its complicated with the words
  131. <Zacam-Work> but if the content doesn't have the -right- shade of black or whatever...
  132. <Valathil> additive blending is also alpha blending
  133. <Valathil> src*srca + dest*(1-srca) what id call alpha blending is order dependent but src*srca +dest what id call additive blending isnt
  134. <Valathil> but i know additive also has alpha
  135. <Valathil> aaaaaaaaaaaaaaahm
  136. <Valathil> i cant find
  137. $no_collisions anywhere
  138. <Valathil> ah TAB
  139. <Swifty> both are alpha blending. blend(source, dest) = (source.rgb * source.a) + (dest.rgb * (1 - source.a)) is conventional alpha blending
  140. <Valathil> yes
  141. <Valathil> i know
  142. *** FUBAR-BDHR_ has joined #scp
  143. <Swifty> blend(source, dest) = source.rgb + (dest.rgb * (1 - source.a)) is premultiplied or additive
  144. <Valathil> i know
  145. <Valathil> no wait
  146. <Valathil> the second is wrong
  147. <Valathil> that be weird
  148. <Valathil> if you have completely black thing and 100% alpha it would result in black square
  149. <Valathil> which is not additive
  150. <Swifty> idk, read what Tom Forsyth has to say on it http://home.comcast.net/~tom_forsyth/blog.wiki.html#[[Premultiplied%20alpha]]
  151. * MjnMixael is becoming a ghost
  152. <Valathil> premultiplied is the better version of normal alpha blending
  153. <Valathil> not additive
  154. <Valathil> additive is still source.rgp + source.a + dest.rgb
  155. <Valathil> source.rgp +* source.a + dest.rgb
  156. <Valathil> fuck
  157. <Valathil> you know what i mean
  158. <Valathil> now let me get back to fixing the collision bug instead of veryfing that we both know what were talking about
  159. <Swifty> which collision bug?
  160. <Swifty> the one with weapons vs rotating subobjs?
  161. *** GhostOfMjn has joined #scp
  162. *** -NickServ- [email protected] has just authenticated as you (MjnMixael)
  163. <Valathil> 2693
  164. <Swifty> oh, that's something else then
  165. <GhostOfMjn> Swifty, have you had a chance to look into the weapon collision issues?
  166. <Swifty> no
  167. <Swifty> i don't really know where to start
  168. <Swifty> i suspect it's something to do with the sort and sweep stuff
  169. <GhostOfMjn> fubar discovered it seems to be mostly dealing with rotating subsystems
  170. <Swifty> have you guys verified that the collisions work again once you set those subsystems to not rotate?
  171. <Valathil> it might interact with this bug now that i hear rotating
  172. <Valathil> no thats just ship to ship
  173. <Valathil> never mind me
  174. *** sigtau|mac has quit IRC: Quit: *closes macbook lid*
  175. <Valathil> interesting
  176. <Valathil> there are two version of collision checks in ship_ship_check_collision both wrongly detect a collision
  177. *** chief1983 has joined #scp
  178. *** ChanServ sets mode +o chief1983
  179. <chief1983> in bug 1616, is there not a chance that behavior was relied on by someone?
  180. <Valathil> haha i think i know why the $no_collisions isnt working
  181. <chief1983> I mean, maybe someone did a same-model ship class change and expected the texture replacement not to carry over
  182. <Valathil> we dont actually check the light ships subsystems do we only the target of the collision
  183. <Valathil> thats why the player can fly through the fenris but the fenris bumps on the ganymede
  184. <Valathil> this comment is also a gold nugget "// should be no ships that can rotate this fast"
  185. <Valathil> lets hope
  186. <The_E> chief1983: That is such a highly specific case that I am not going to bother checking against it
  187. <The_E> BEsides, the typical use case referred to here is much, much more elegantly handled by team colors
  188. *** einstine909 has quit IRC: Remote host closed the connection
  189. <Valathil> did i ever mention the collision code i an ugly hacked up piece of shit?
  190. <The_E> GhostOfMjn: Have you tried reproducing 2605?
  191. <GhostOfMjn> no
  192. <GhostOfMjn> i shall though
  193. <The_E> Awesome
  194. <Valathil> is nobody talking to me?
  195. <The_E> My personal working theory is that the person who reported it did not actually check whether the weapon he put onto the turret is actually one the AI would fire at a capital ship
  196. <GhostOfMjn> yeah, that's what it sounds like to me
  197. *** Valathil has quit IRC: Quit: Verlassend
  198. <GhostOfMjn> and there's the smart spawn flag now too
  199. <GhostOfMjn> but standby whilst I test
  200. <GhostOfMjn> in theory Spawn + Capital+ should do the trick to just about any weapon
  201. <GhostOfMjn> yeah, I've an orion firing Infyrno's at a Typhon
  202. <The_E> By simply adding capital+?
  203. <GhostOfMjn> yeah
  204. <The_E> Right
  205. <The_E> So closing that bug as working as intended would be justified
  206. <GhostOfMjn> er, well I removed remote detonate, but yet
  207. <GhostOfMjn> yes*
  208. <chief1983> did Valathil just quit because he was being ignored?
  209. <GhostOfMjn> dunno
  210. <chief1983> or did he interpret your comment before he quit as directed at him, incorrectly
  211. <The_E> Closing rapidly on only 150 open issues
  212. <chief1983> awesome
  213. <chief1983> You guys are killing it, I just want to say.
  214. <The_E> A lot of it is due to mantis housekeeping, I should add
  215. <GhostOfMjn> can you look at 1153?
  216. <chief1983> Still
  217. <chief1983> clean mantis means it's easier to find relevant bugs
  218. <GhostOfMjn> indeed
  219. <chief1983> and it's still time and effort to clean up
  220. <chief1983> Trust me, I know all about that shit.
  221. <chief1983> We had a bug queue at work the other day over 120 issues long
  222. <chief1983> in a week we had whittled it down into the 30s, and then nearly less than 20 before it ballooned again :(
  223. <chief1983> But it's a fight just to keep on top of what's actually an issue or not.
  224. <GhostOfMjn> after 1153, 2713, 2165, and 2325 need coder attention.
  225. <The_E> Checking 1153
  226. <The_E> 1153 is closed
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement