[22:09] hey [22:10] we're calling you on stef's cell, the voip line is sucking [22:17] yay! [22:39] http://securityinnovation.com/ [22:50] So, I'm looking at the blog model definition, and I see the do: [:attack | attack riskFactor: 5]. at the end... that's setting all the risk values for the externalAsset, right? [22:51] how do I do the equivalent thing for the other assets? [22:59] ping ? [22:59] erp. [22:59] let's try that again. [23:00] ping/ [23:00] bah. :-) [23:00] i'm looking, i'm just not finding [23:00] i didn't make a convenient interface for that [23:01] basically you want to be able to get at each threat based on which action it is for, and its class, so you can set the values. [23:01] Action: Dymaxion grins. [23:01] yeah [23:01] how does the UI do it? [23:01] lessee [23:03] hee hee [23:03] i couldn't really say. i abstracted it all away. [23:03] heheh [23:04] i think the short answer is, it already has the threat it needs at any given point, because of the events that are triggered as you change the model [23:04] huh, ok. [23:04] there's a method on asset called actionAt: [23:05] is there a current image somewhere on zettai? [23:05] my image is a bit fux0red delux [23:05] no, you should get it from paul [23:05] or you could try the build script with a fresh image [23:05] oh, right...that's what I was trying to do [23:06] the build script is checked into svn, and the .mcz files are still on zettai in that one place [23:06] yeah, i have current mcz files [23:06] I'll update svn [23:06] theoretically you start the fresh image, browse to that cs file and file it in [23:06] worked for me [23:07] you have to set an environment variable. after you file it in i think it gives you instructions [23:08] Dym, you want to make a method like actionAt: on action, to get threats [23:08] what do you mean file it in? [23:08] oh [23:09] get a file list from the open menu, browse to it in the filesystem, and then i think there's a button like "file in" [23:09] that's a pretty poorly chosen name, huh? [23:10] no, it's called install [23:10] it was called "file in" a long time ago. like the 80s. [23:10] haha :P [23:10] chalk it up to smalltalkism [23:10] ok, it's running! [23:11] yay! [23:11] i'm sorry it still takes a long time [23:11] ok [23:11] or at least it popped up a progress bar that's not moving [23:11] P, so maybe attackInClass: [23:11] oh, yah [23:11] don't click on anything [23:11] the progress bar is better than it used to be [23:12] "could not resolve the server named: sqeak.krampe.se" [23:12] but still lame [23:12] fuck. :( [23:12] it's trying to update squeakmap [23:12] yeah [23:12] should I update sm first? [23:12] i shouldn't swear on a logged forum, sorry [23:12] sure, try that. it can't hurt [23:12] maybe it will give you a better error [23:12] hmmm...the progress window won't move, lol [23:13] yeah...it's all frozen :D [23:13] sometimes it isn't moving when it is doing something. [23:13] die die die! [23:13] you could press command-period [23:13] eh, screw that not swearing shit. :-) [23:13] and then if it is doing something you could resume or continue or whatever. proceed. [23:14] if it's not doing anything, you could abandon. but you would have the ui back [23:15] so paul, you could steal code for TrikeAction>>attackInClass: from TrikeAction>>addThreatIfNotAlreadyPresent: [23:16] oki, updating squeakmap [23:16] oh, I just killed the image and reloaded it [23:16] in fact you could call attackInClass: attackInClass:ifNone: and refactor addThreat... to call the former. [23:16] S, that works [23:16] and then deleted package-cache & sm in my image dir [23:17] right. i should figure out how to make that not be necessary [23:17] because I was getting a really weird error, that apparently means 'delete package-cache & sm in your image dir' [23:17] ;) [23:17] probably by filing a bug report on goran [23:17] yeah, i've seen that one :) [23:17] Dym, how's it coming? [23:18] ooh, can you patch him with a squeaker who will worry about backwards compatibility? [23:18] hee hee [23:19] i think he's trying, he just didn't get it all quite right. he works really hard & is a really helpful person, and there will be no goran-knocking on my beat. [23:19] so there. [23:19] Action: Dymaxion grins. [23:21] I popped debugger [23:22] damn [23:22] oh, by clicking on something? [23:22] Object class(Object)>>doesNotUnderstand: #subclass:uses:instanceVariableNames:classVariableNames:poolDictionaries:category: [23:22] it was all working so nicely.,. :( [23:22] Undefined object DoIt [23:22] whoa [23:23] had you started installing TrikeBuilder, or were you still working on SqueakMap? [23:23] no, it was installing TrikeBuilder [23:24] ok, that's my bug. but Paul, you tried it earlier & it worked for you? [23:24] es [23:24] yes [23:24] even [23:24] there were occasional weird errors, but it ended up working in the end. [23:24] in that debugger, can you find the stack frame that says Undefined DoIt? [23:25] oh, I killed it [23:25] paul said I could abandon [23:25] oh. [23:25] Action: ar4chne tries again [23:26] sometimes abandon did the right thing for me... it's been hard to figure out which I should do some times. [23:26] how about, get a working image from paul, and let's debug the build script when i am there in person? [23:26] it shouldn't be popping any debuggers, although it is occasionally ass slow [23:26] ok, deal [23:27] in that case, I think I'm going home [23:27] 4:30 is only 7 hours away [23:27] ok.. sleep well, happy flight and good luck at shmoocon. i wish i could be there [23:27] me too! [23:28] and thanks :) [23:28] sorry I wasn't more productive [23:28] 'sok. you have been awesome with the web site & all [23:30] oh my god. [23:30] there is a BalloonMorph, and it's a subclass of PolygonMorph [23:30] so all PolygonMorphs must look like that [23:31] weird [23:32] are there different types of polygon morphs? [23:33] like, maybe they pass in a constructor argument that tells it how to draw? [23:36] well, you can tell it to be "smooth" or not, but it looks like they all have jaggies like that [23:43] ok. we have some choices [23:44] the balloon spawning process is a collaboration between handmorph and whatever morph has the balloon text [23:44] we could make changes in any of 3 places: [23:46] handmorph - this might be where we need to deal with the cursor hints anyway, but i hope not; it is kind of a black magic part of the system since i think there are vm primitives to support it [23:47] balloonmorph - probably refactoring so it isn't a subclass of polygonmorph would fix it -- we'd just have bordered squares with text in them; however, it's a rude change to put in a package on squeakmap [23:48] the morphs that have balloon help - not so bad since we are already changing many of them; disadvantage is if we forget a change somewhere, users see inconsistent balloon help [23:48] btw, i have to go in ~ 10 min [23:49] or, we could fix the root cause, that being lousy polygon morphs [23:49] ned may be using polygons in connectors, too; can't recall. if he is, this would be a good fix [23:50] (P, are you still there? or am i talking to myself?) [23:50] yeah, I'm here, but talking to stef for a bit, haven't been reading closely for thesecond [23:50] sorry [23:50] 'scoo, just thought i'd shaddup if nobody was listening :) [23:57] ok. [23:57] i have to go. i am very tired & i can't figure this out right no [23:57] ok [23:58] i will see y'all later [23:58] sleep well [23:58] good night! [00:00] --- Thu Jan 12 2006