[10:58] gmorning [15:04] Action: Dymaxion waves [15:57] hey, dym :) [21:28] one of our criticals leads to a weird scenario. suppose you have this: [21:28] or [21:28] a [21:28] or [21:28] b [21:29] !b [21:29] you simplify the child or [21:29] the answer to the simplification is, of course, always [21:29] so now you have to say [21:29] or [21:29] a [21:29] always [21:30] but there is no way to store always in a rules tree [21:31] the right thing to do is simplify the top or, but if you do that, it is a little weird [21:31] because it's not what the user asked for [21:32] i think i am going to change the ui so you can only simplify the whole rule tree [21:32] not subtrees [21:32] there is of course a similar problem for and [22:17] yeah [22:17] That makes more sense anyway. [22:17] Now that we've fixed the parenting bug, we shouldn't need to simplify manually most of the time. [22:20] did you see i found the repro steps for your disappearing model sections? [22:20] unfortunately they involve a debugger [22:20] and i am having trouble finding them [22:20] blah [22:21] oh [22:21] i just figured it out [22:24] the radio is so funny. somebody just called the country station to report that 7 helicopters flew over her house [22:24] so the DJ is going to figure out what happened. the DJ requests anyone with information to call her [22:24] hehe [22:24] oh, I haven't added the xml export ui bug yet -- should I? [22:25] i mean, if helicopters flew over my van i'd call the country station for sure [22:25] naw, i've already partly implemented it [22:25] so i didn't forget :) [22:26] ok. :-) [22:59] no more criticals! whoo hoo! [23:01] wooohooo! [23:01] Action: Dymaxion looks at eir watch^Wtimestamp. Right on time, too! [23:02] well, i don't _quite_ have the file export ui working [23:02] there is a little missing glue [23:02] Action: Dymaxion nods [23:03] oh yah -- ara called. we should package up the windows release & put it on zettai, and when she is awake & has net again tomorrow, she'll release that. [23:03] then when she has time, she'll test for os x, and package & release that [23:04] cool [23:04] Action: asparag1 is waiting for unit tests [23:19] the files are uploading [23:20] i think i am starting to get excited [23:20] i like our debugging process. the part where we found & fixed that stuff before we shipped seems really good to me [23:21] Action: Dymaxion nods [23:21] ok. the files are waiting for you. [23:21] cool [23:21] i am going to go build an image as well :) [23:21] and yeah, I've been really happy with the past week or two as a whole, really, including the feature stuff [23:22] but mine will probably take the whole time to build that yours does to build & test [23:22] why? [23:22] slow net [23:23] oh, you're building from scratch, right. [23:23] the build process installs things off squeakmap [23:23] oh, i am so excited! [23:23] this is going to be awesome [23:23] you are going to build & find more criticals, and then I will fix them [23:23] i hope not [23:26] Action: Dymaxion grins [23:27] oh. i don't have 6718. i'm not going to have it for another 40 minutes & 43 seconds [23:27] so i guess i'm not going to build [23:27] Action: Dymaxion grins [23:27] Action: Dymaxion went and checked for a more recent build, but it 6718 is still current. [23:27] as soon as you have test results i'll stick the packages in SqueakMap [23:28] that should make it easier for ara to build on OS X, too, theoretically [23:28] ok [23:30] can you put it on squeak map in some kind of test mode so we can try a build from there, too, before it goes public? [23:31] nope. all public, all the time [23:32] hrm. [23:32] Well, we'll test quickly, then. :-) [23:32] but i yup [23:32] er [23:32] Action: asparagi 's head is full of half-formed sentences which drift together randomly [23:33] yeah. i am a little concerned about the dependency thing [23:33] Action: Dymaxion nods. [23:33] i never learned how to work it. [23:34] we have an abnormally fine-grained collection of .mczs... [23:34] maybe it will just work, ha ha [23:34] Action: Dymaxion grins. [23:47] built [23:49] debugger on Delete Actor [23:51] Create a new actor in Blog, then delete it. [23:52] Don't have another repro [23:52] Can't get it to happen in another TM, might have something to do with other rule state. [23:52] bah [23:52] Want me to file this? [23:52] lemme try to repro [23:53] worked for me first try [23:53] you did or did not reproduce it? [23:54] i reproduced it [23:54] then i fixed the bug [23:54] next? [23:54] gotcha [23:54] ;) [23:54] woO! [23:55] got any more bugs? [23:56] working on it, nothing yet [23:57] Hrm. [23:57] You can delete the external asset, thus leaving a model with no out of scope asset [23:59] that isn't supposed to work [23:59] i wonder if it only updates the ui. [00:00] --- Fri Feb 3 2006 [00:00] no, it's really gone [00:00] feh [00:01] debugger on negating an and with a negated text clause below it; working on confirming that as a minimal test case [00:02] no, there's something more going on, hang on [00:03] ok, go to the rule tab, open up Actor Reads User Account all the way, negate "It is Actor's own User Account", and then attempt to negate the AND above it [00:04] -> debugger, "Object is not in the collection" from WeakIdentitySet, called by TrikeActorRule removeParent: [00:06] uhh... not sure how to repro this one, but I was beating on something else, repeatedly negating it, and I'd swear something in the tree for another action changed. [00:07] hmm [00:10] If you have an action which has only a text clause as a child and you click on the child and say "add sibling rule->text clause", it doesn't. [00:11] ditto for clicking on the action and saying "also allow when->text clause" [00:11] ok, i fixed the Actor's own Account debugger [00:11] i think [00:12] i'm relying on the garbage collector a little more heavily [00:12] it makes sense that you couldn't add a sibling text clause [00:12] oh, no it doesn't [00:12] feh [00:13] oh [00:14] i think there's an automatic simplification going on when the rule is set [00:14] right [00:14] can we leave that one for next release, or do you think it's critical? [00:14] which then says "these are the same" [00:14] no -- annoying, but not critical [00:14] k. [00:14] like the thing with external asset, unless that's already fixed [00:14] how 'bout deleting the external asset? [00:14] want me to file? [00:15] no, not yet [00:15] file the auto-simplification [00:15] i'm going to take a quick stab at the external asset [00:15] ok [00:16] what do you think it should do if you try to do that? [00:17] the same thing it does if you try to uncheck it -- nothing. [00:17] ideally, we want some way of notifying the user of what's going on, but we don't really have that right now. [00:18] it's my hatred of modal dialog boxes [00:18] I think we'll eventually want some generic general purpose way of displaying help and feedback to the user, and once we've got that, it'll be fine. [00:19] I don't think we want a dialog box at all here. [00:19] In fact, I'll be very annoyed if one shows up for this sort of informational thing. [00:19] ok, fixed [00:19] cool [00:19] filed [00:19] yeah, that's why i didn't put one in [00:19] any more? [00:19] not yes [00:20] er, yet [00:20] I think we may want to schedule that for the next release if we can -- our UI is getting somewhat complex, and better feedback would be a big win for usability. [00:21] well, ara's suggestion was a little status line [00:21] i don't have a suggestion [00:21] and you & i apparently both dislike modal dialogs [00:24] http://bugs.octotrike.org/bugs/bugs/0074 [00:26] is there a missing piece before "that might make a good first pass"? [00:26] haha [00:26] yes. [00:26] i know how to avoid trouble with the external asset [00:26] the stupid auto text formatting crap did something evil. [00:26] don't put the remove item in the menu to start with [00:27] since we have menus with items that appear & disappear already, this isn't any weirder than any other situation [00:27] but wait, what if they add a new out of scope item and want to remove external asset/ [00:27] only let it show up if the asset could be deleted? [00:27] although greying it out and supplying a balloon help reason could be good [00:27] yeah, exactly [00:28] either one works [00:28] why tempt the user with the impossible? [00:28] well, i'm not going to do either right now.. [00:28] i'm starting to get tired [00:28] lemme know when you want the last drop, then i'm going to bed [00:29] ok, which attacks can be edited is not consistent. [00:30] You can edit ORs but not ANDs, and only some of the default tree is editable [00:30] uhhhh [00:31] oh [00:31] yah [00:31] Closing all the threat models makes Quit!, Save, and Save and Quit go away, and they don't come back when you create a new one. [00:31] hee hee [00:31] only some times though [00:31] sometimes it happens when you create a new TM after deleting the others [00:32] so remember when we were talking about how to tell if an attack node is an AND? [00:32] some times just some of the buttons go away. [00:32] can you fix it with the menu? [00:33] no, the menu doesn't change when this happen [00:33] i decided the way to tell, right now, was to have the ANDs always say AND [00:33] but, check & uncheck the appropriate box [00:33] ah, ok [00:34] do the missing buttons come back? [00:34] we should document that [00:34] hang on a sec [00:34] debugger! [00:34] I tried to turn off the section with missing buttons [00:34] re: only some attacks editable, can you give me examples? [00:34] hang on [00:34] damn [00:34] damn damn [00:35] ok, no, this is weirder [00:36] first, I could edit the name of a threat, but the name changed back [00:36] then, I went through editing names and they all worked, until I edited a name but clicked elsewhere without actually changing it. Now a bunch of them are uneditable. [00:37] hard to tell if there's a pattern [00:38] oh my [00:38] that sounds rather strange [00:38] silently truncates file name in the export box, but that's not bad behaviour [00:38] is the one you were editing still green? [00:38] nope [00:39] silently truncates file name? [00:39] if it's a few hundred characters long when you enter it, say [00:39] i see [00:39] (truncates to maxpath, I think) [00:39] seems good :) [00:39] file that threat name one [00:39] i don't know [00:40] Ok, I have to go hang out with alex -- I'll file stuff when I get back, and I'll email you if I find anything new. [00:41] bbl [00:43] ok. sounds good [00:56] i'm going to bed [00:57] i am working on the section debugger thing [00:57] it makes me sad that we can't release [00:58] i guess next time we need to get to stable features earlier so we have longer to pound [00:58] :( [00:58] good night