| 01:20.53 | *** join/#riftuidev Noshei (~quassel@c-24-9-171-99.hsd1.co.comcast.net) |
| 01:47.16 | seebs | According to Slashdot, the next NetBSD kernel will have a lua interpreter built in so some driver-like functionality can be written in Lua. |
| 03:51.45 | *** join/#riftuidev Cairenn (~Owner@MMOI/Administratrix/Cairenn) |
| 03:51.45 | *** mode/#riftuidev [+o Cairenn] by ChanServ |
| 05:17.24 | *** join/#riftuidev Imhothar (miranda@ppp-93-104-177-233.dynamic.mnet-online.de) |
| 07:37.18 | ZorbaTHut | that's sort of cool and sort of weird |
| 07:37.22 | ZorbaTHut | I admit I'm not sure as to why they'd do that |
| 08:23.36 | seebs | I am not entirely sure either, but they are pretty clever folks. |
| 08:23.44 | seebs | They were the first kernel I saw *usefully* optimize the idle loop. |
| 08:23.59 | seebs | Which became the "zero out free pages and move them to the zeroed page list" loop. |
| 08:24.43 | seebs | I once had to deal with Bad Customer Benchmarks. |
| 08:25.33 | seebs | One of the benchmarks involved a program which ran an order of magnitude faster if you called calloc() in its inner loop than if you called malloc(). As long as you didn't actually WRITE to any of the allocated memory. |
| 08:25.49 | seebs | I just killed a machine testing performance fixes. I'm pleased with that. |
| 08:25.59 | seebs | I'm also pleased with the performance fixes, I think. |
| 08:26.30 | seebs | I am pretty sure that, for real loads, I've managed improvements between 25% and about 70%. |
| 08:26.32 | ZorbaTHut | I'm always kind of amazed at how badly people take advantage of idletime. when I was on questhelper I had all sorts of crazy code that would attempt to flatten your framerate while using an average of around 7.5% CPU |
| 08:26.42 | seebs | Yeah. |
| 08:26.52 | ZorbaTHut | if your framerate dropped, it would quiet down and stop using CPU, and if it spiked, it would take advantage of the newfound wealth of CPU to get ahead a bit |
| 08:26.58 | seebs | So, cool thing to know, if you ever have horrible disk performance on Linux. |
| 08:27.12 | seebs | fsync() is *documented* as syncing writes to a particular file. |
| 08:27.34 | ZorbaTHut | hah, I know this one. It does sync writes to that file! And every other file! :D |
| 08:27.37 | seebs | But on many filesystems, the only way to reliably do that is to make sure that every write delivered to the filesystem code before the last write to that file is also flushed. |
| 08:27.41 | seebs | Yes. |
| 08:27.52 | ZorbaTHut | yeah, one of those functions you stay far far away from unless you *really* need it. |
| 08:28.05 | seebs | Well. A database might have a good claim on it. |
| 08:28.18 | ZorbaTHut | I think user-facing programs might have a reasonable use as well |
| 08:28.21 | seebs | The database that holds your package manager's view of the system has a very good claim on it. |
| 08:28.26 | ZorbaTHut | fsync() after a user-initiated save operation isn't the worst idea |
| 08:28.38 | ZorbaTHut | you just don't want it in inner loops or high-speed automated systems |
| 08:28.38 | seebs | Unless you're using that package manager to, say, hypothetically. Build filesystems. |
| 08:28.53 | seebs | And our build system does a LOT of filesystem assembly. |
| 08:29.38 | seebs | The beautiful thing is, everyone blamed pseudo because pseudo was spending so much time in IO wait, or because programs known to be hitting it fairly hard were in IO wait. |
| 08:29.48 | ZorbaTHut | hah, yes, I can see that be an issue |
| 08:29.50 | seebs | And it turns out, it's not at all pseudo's fault, but pseudo can fix it. |
| 08:29.58 | seebs | See. fsync, and fdatasync? |
| 08:30.13 | seebs | System calls. Calls made by loading a symbol from a shared library and executing code... |
| 08:30.36 | seebs | .../pseudo/configure --enable-force-async |
| 08:30.53 | ZorbaTHut | aha, clever :) |
| 08:30.55 | seebs | Oh, hey, fsync returned 0 in about 5 CPU cycles! |
| 08:31.04 | seebs | Not my idea, came from one of the other guys. |
| 08:31.04 | ZorbaTHut | yeah you get all kinds of crazy abilities once you start hooking internal code |
| 08:31.24 | ZorbaTHut | a while back I rigged up code to inject mouse events directly into scaleform internals, for example. need to go back and rejigger some of the rift frames to take advantage of that :V |
| 08:31.42 | seebs | Although I improved on it by modifying the wrapper generator so affected functions do it INSTANTLY, without even getting into the "about to do magic" mutex. |
| 08:32.08 | seebs | I also made pseudo use an in-memory sqlite database and only write it to disk on exit. |
| 08:32.13 | seebs | And! I changed the protocol. |
| 08:32.40 | seebs | Instead of client sends message, server responds, the protcol is now client sends message, server responds unless the message was the new "FASTOP" type, in which case it doesn't. |
| 08:33.06 | seebs | For cases where you're just notifying the database of changes, not querying it, this means the client continues processing without waiting for the server to get back to it. |
| 08:33.54 | seebs | So an untar of ~28k files, on my workstation, is about 1.1 sec outside of pseudo, ~15 sec in pseudo before changes, ~12 sec after I dropped fsyncs and went to an in-memory DB, and about 3.5 sec with the protocol change. |
| 08:34.02 | ZorbaTHut | nice :D saving roundtrips is absolutely a good thing |
| 08:34.22 | seebs | And scales roughly linearly if you run several at once, instead of rather worse. |
| 08:34.38 | ZorbaTHut | I may end up working on rift server performance after I've beaten the client into more efficiency. That'll be interesting. |
| 08:34.46 | seebs | Note, none of this is stuff I'm actually SUPPOSED to be doing. |
| 08:35.10 | seebs | But my manager is awesome, and accepts the occasional couple of days during which I perseverate on something that is not actually on my schedule. |
| 08:36.05 | ZorbaTHut | yeah, sign of a good a manager to recognize that and accept that |
| 08:36.18 | ZorbaTHut | there are a lot of people who need constant guidance in order to work |
| 08:36.26 | ZorbaTHut | and then there are people who, given constant guidance, get pissed off and annoyed :P |
| 08:37.25 | Noshei | geeze, i look away from IRC for a minute and you guys have a books worth of chat! |
| 08:38.06 | seebs | It's characteristic autistic obsession. Only hits occasionally, but when it does... No amount of trying to do other things gets me anywhere until I've figured it out. |
| 08:38.54 | seebs | And given how much suffering has been caused by "oh, and your entire machine becomes unusable during filesystem assembly", I figure they'll be tolerant. :) |
| 08:39.23 | ZorbaTHut | yeah, I do the same thing once in a while. Oh, there's a bug? It's behaving weirdly? MUST FIGURE IT OUT |
| 08:39.47 | ZorbaTHut | of course the end result is that I end up understanding our codebase and libraries better than anyone else, so now everyone comes to me to ask difficult questions that seem impossible :P |
| 08:39.54 | seebs | It started with "hey, what if I just do a quick test with no change other than telling sqlite to use the in-memory DB". |
| 08:40.01 | seebs | Result: Factor of three slowdown. |
| 08:40.37 | seebs | See, if it had been faster, I coulda reported that back and gone my merry way, not caring. If it had been marginally faster, I coulda written it off as unrewarding. |
| 08:40.53 | seebs | SLOWER? That is a sign that I am missing something serious. |
| 08:41.34 | seebs | Answer: sqlite 3.6 has some kind of severe performance bug I do not adequately understand. 3.7 is okay. |
| 08:42.20 | seebs | So I swapped to my new code, ran test builds... and they came out within a second or so of the time of unmodified builds. Which is bad considering that for real test cases I was seeing significant improvements. This was when I realized that my assumption, that my code was slowing things down, was unsupported. |
| 08:42.39 | seebs | And that was roughly two days ago and I'm just now regaining normal awareness of my environment. |
| 08:43.15 | ZorbaTHut | welcome back ^^ |
| 08:44.14 | ZorbaTHut | I did the same thing with the text issues |
| 08:44.29 | ZorbaTHut | I *think* they'll be fixed in 2.2, but I thought they would be fixed last time, so |
| 08:44.43 | seebs | haha. Yeah. |
| 08:44.46 | Noshei | whats wrong with test now? |
| 08:45.14 | seebs | pseudo 1.4.3 would sometimes double-prepend an emulated chroot prefix. 1.4.4 would sometimes omit it entirely. |
| 08:45.21 | ZorbaTHut | noshei, text :) seebs found a reproducible case where it would (as it turned out) clear html formatting |
| 08:45.39 | seebs | 1.4.5 compromised on one copy of the chroot prefix, which I think satisifies everyone. |
| 08:45.39 | Noshei | ah |
| 08:45.50 | seebs | Oh, good. So I can take my workaround out "soon". |
| 08:46.01 | ZorbaTHut | another of those bits where I end up unsure whether it's actionscript's fault or scaleform's fault, but it's *one* of their faults because that behavior is not intuitive in the least |
| 08:47.13 | Noshei | so ZorbaTHut, 2.2 on wednesday? |
| 08:47.30 | seebs | I really oughta sleep soon. Plan is, since I killed my workstation badly enough that any attempt to sync() would go into uninterruptible sleep, as would any attempt at a reboot, I have power cycled it. Now installing updates and deleting old build directories. After that, plan is to re-run two sets of tests (one with patches, one without) and see what I get. |
| 08:47.36 | ZorbaTHut | (for the curious: when you set HTML text, it parses the html and assigns formatting per-character, which makes sense. but if you then attempt to change the text format it doesn't apply the new text format as a default, overridden by the HTML formatting - it actually nukes *all* the formatting and sets it to whatever your new "default" format is) |
| 08:47.53 | ZorbaTHut | noshei, if it's not on the PTS yet, I doubt it'll be on Wednesday. I'm actually not entirely sure what the schedule is right now. |
| 08:47.58 | seebs | Ohhh! |
| 08:48.38 | seebs | And then the "that's what I already rendered" check prevented it from getting fixed until something changed. |
| 08:49.02 | ZorbaTHut | (solution: *all* text is now handled by HTML and I wrap the user-provided text in the appropriate tags before passing it in. I no longer use the global formatting functions whatsoever. Might reduce in slight slowdowns when changing text, but considering how slow that code is anyway, I doubt the parsing will be significant.) |
| 08:49.06 | seebs | Right now, I think, PTS is on 2.2 patch notes #4. |
| 08:49.21 | ZorbaTHut | (and yes if you haven't set the html flag it will carefully escape your text before wrapping it so you don't accidentally HTML stuff :V) |
| 08:49.27 | seebs | Hmm. I assume, then, non-... yeah. |
| 08:49.54 | seebs | I finally solved my HTML-escaping problem in printf. |
| 08:50.17 | ZorbaTHut | in retrospect I kind of wish I'd used markdown instead of html, but it turns out there isn't even an official markdown spec |
| 08:50.33 | seebs | See, I want printf("%s", foo) to be safe -- to actually show foo as-is, even if it has angle brackets. |
| 08:50.43 | seebs | But what if you really DO want to hand in formatting? |
| 08:50.44 | seebs | So! |
| 08:51.13 | seebs | The HTML-ready printf variants in my library scan the arg list for strings. Strings are subject to HTML escaping. |
| 08:51.27 | seebs | They also check for tables. Tables are subject to tostring(). |
| 08:52.03 | seebs | And there is an htmlok() which, given a string, yields a table with that string as a member and a metatable with a __tostring which yields that member. |
| 08:52.21 | seebs | I don't remember what I named it, but it works well enough for my purposes. |
| 08:52.26 | ZorbaTHut | ah, that's rather clever |
| 08:52.59 | ZorbaTHut | so you can do things like printf("%s (%s)", user_submitted_name, htmlok(formatted_stats)) and it just works |
| 08:53.00 | seebs | Markdown and textile are lovely, but markdown doesn't support color, which is pretty much the one thing you actually need. |
| 08:53.07 | seebs | Yup. |
| 08:53.15 | ZorbaTHut | I like the safe-by-default behavior, that's always a good feature |
| 08:53.22 | seebs | And it works even if the first field happens to have angle brackets or something. |
| 08:53.31 | seebs | Pikey uses it fairly heavily. |
| 08:53.34 | ZorbaTHut | yeah, markdown would probably need some extra love. I don't know how I'd even embed color into it |
| 08:53.53 | seebs | I should probably go rework SlashPrint to be smarter about this. And, possibly, to make it a little terser. |
| 08:54.09 | seebs | Although around 1.11, I revamped it significantly to add color coding. |
| 08:54.17 | ZorbaTHut | hmm, textile isn't bad |
| 08:54.23 | ZorbaTHut | maybe I'll try that out for my personal projects |
| 08:54.26 | seebs | RiftRC and SlashPrint are among my favorite tools for such things. |
| 08:54.28 | Noshei | seebs, I think you should make a nice scrolling combat text addon for me |
| 08:54.49 | seebs | I learned Textile for my blog (which runs Textpattern). I find it pretty friendly. |
| 08:54.54 | Thorarin | Wasn't there some initiative to work on an official Markdown standard? |
| 08:55.01 | seebs | Probably. |
| 08:55.09 | ZorbaTHut | thorarin, very recent, yes, though I don't know if they've yet made any progress |
| 08:56.10 | Thorarin | Haven't read anything about it since, I must say... I wouldn't know |
| 08:57.09 | ZorbaTHut | yeah, I'm liking textile. May end up using this. Thanks for the suggestion, seebs :D |
| 08:57.18 | ZorbaTHut | (with a bit of mangling because it won't end up targeting html, but) |
| 08:57.45 | seebs | Yeah, it's a respectable alternative to Markdown. Slightly different intent, but I'd have a hard time explaining how it feels different. |
| 08:58.21 | ZorbaTHut | I feel like markdown is an effort to make a readable formatting system that replaces html |
| 08:58.31 | ZorbaTHut | whereas textile is more of an effort to make a readable formatting system that is equivalent to html |
| 08:58.57 | ZorbaTHut | there's a lot of stuff in textile that looks to be "well, we did it this way because it's parallel to html" |
| 09:01.12 | Thorarin | Argh... too early to run into the most difficult issue in programming, bar none |
| 09:01.20 | Thorarin | Correctly naming stuff :< |
| 09:01.39 | seebs | Hah. |
| 09:02.03 | Ivnedar | Consistently naming is probably more important :D |
| 09:02.11 | ZorbaTHut | I'm glad seebs talked me out of sink/bubble :V dive/bubble works a lot better |
| 09:02.25 | Thorarin | Well, correct AND consistent in this case :) |
| 09:02.38 | seebs | I think... Markdown's goal is that if you don't do any conversion, it's still legible. Textile's goal is that you can understand what code it will generate. |
| 09:02.51 | Thorarin | I'm writing an extension to an IoC container, so have to be consistent with the existing stuff |
| 09:03.48 | seebs | So, I was muttering to myself about the need for "professionalism" in code that I ship to open source projects from my work address. |
| 09:03.51 | ZorbaTHut | yeah, whereas my goal is "text formatting that a designer or artist can do easily" |
| 09:04.05 | ZorbaTHut | (this is not for rift btw, this is for some future revision of my frames library) |
| 09:04.07 | seebs | And it came to pass that my brain attempted to kill me, by offering: |
| 09:04.37 | seebs | "Will smith don't gotta cuss in his comments to sell records!" Well, I do, so /* fuck him */ and /* fuck you too */. |
| 09:05.06 | ZorbaTHut | hah |
| 09:05.19 | seebs | Which has a fascinating impact because the timing only works if there is no verbal indicator of the comments, but it's only funny because of the placement of the comment tags. |
| 09:08.02 | seebs | Jesse was impressed. Which makes me all smug, because Jesse is an actually competent poet. |
| 09:08.45 | Thorarin | :P |
| 09:09.06 | Thorarin | Cursing in comments can be a very bad thing though :) |
| 09:09.16 | seebs | I am okay at funny writing and technical writing, but not so much at the art writing. |
| 09:09.44 | seebs | It's heavily contextual. Me, I really don't cuss in comments normally, but I do sometimes adopt a slightly silly tone. |
| 09:10.30 | Thorarin | Once ran into some code for MS Office templates full of "Customer wants stuff to work like this. Fucking morons!" |
| 09:10.45 | Imhothar | ./* The following code is autogenerated, any modifications get lost on every build */ |
| 09:11.03 | seebs | Yeah, stuff like that can be a bad sign. |
| 09:11.04 | Imhothar | I put that on code some tard keeps changing and by that makes things worse every time |
| 09:11.07 | seebs | I have... Well. |
| 09:11.20 | Thorarin | Apparently the programmer think that said customer might be able to view the macros :) |
| 09:11.23 | seebs | I have worked on a code base from which a *representative* sample made it to the Daily WTF. |
| 09:11.49 | seebs | Some of the comments in my code which had to do things to comply with this may not have lived up to the highest standards of professionalism. |
| 09:12.08 | Noshei | hey ZorbaTHut, any chance you could look into making the "screen messages" that show up for your character moveable, via the interface. |
| 09:12.45 | ZorbaTHut | noshei, I think it's on my list but it's pretty far down the list right now |
| 09:12.48 | Thorarin | I think we all run into some WTF-worthy code from time to time ;) |
| 09:13.07 | seebs | Well, allow me to share with you a representative sample. Imagine, if you will, a database. |
| 09:13.10 | Thorarin | "Only measure of code quality: WTFs/minute" |
| 09:13.30 | seebs | As you are of course aware, best practice is to store passwords encrypted, using a hash and a salt value. |
| 09:13.32 | Imhothar | We once had a fun project to jerk people off |
| 09:13.50 | seebs | ... I am betting you don't know the idiom there. |
| 09:13.55 | Imhothar | we made a tool that translated every code symbol to WTF001... WTF999 |
| 09:14.29 | Imhothar | our own little attempt at code obfuscation |
| 09:14.41 | seebs | So, the database designed by my predecessor on this project did precisely that. The three relevant fields of the user table in the database are "username", "password", and "encrypted_password". |
| 09:15.04 | ZorbaTHut | I guess, technically, they were storing the password encrypted. |
| 09:15.06 | Imhothar | saving cleartext password, yay for evil |
| 09:15.20 | seebs | Where "encrypted_password" holds the value of md5sum("COMPANYNAME" + password) |
| 09:15.28 | Ivnedar | maybe the password is a boolean! |
| 09:15.31 | Ivnedar | hopes |
| 09:15.32 | seebs | Oh, and the company name was misspelled. |
| 09:15.39 | seebs | No, it was the cleartext password. |
| 09:15.40 | Noshei | lol |
| 09:15.41 | ZorbaTHut | that's a security feature! |
| 09:15.47 | ZorbaTHut | makes it harder to guess. :D |
| 09:16.09 | Thorarin | trololol |
| 09:16.12 | seebs | In every record in the entire database, it was true that md5sum(salt + password) == encrypted_password. |
| 09:16.22 | Thorarin | How useful to store it encrypted like that :) |
| 09:16.29 | Imhothar | and convenient |
| 09:16.47 | seebs | The person I replaced had billed two thousand hours in a six month period. I have absolutely no reason to believe he worked even half of them. |
| 09:17.22 | ZorbaTHut | in fairness, never underestimate how slow an inexperienced coder can be |
| 09:17.25 | seebs | But... that was *representative*. |
| 09:17.36 | Imhothar | yeah, I'm sure he had 80 hour weeks for 6 months |
| 09:17.41 | seebs | Well, yeah, but... 2000 hours is pretty much a full working year. |
| 09:17.53 | seebs | And turnaround time on email was usually a day or two. |
| 09:18.25 | Noshei | yeah thats a lot to bill for 6 monthds |
| 09:18.34 | ZorbaTHut | not, admittedly, a good sign |
| 09:18.48 | seebs | It was... pretty amazing, really. |
| 09:18.50 | Ivnedar | maybe outsourced his job out to 10 other people and billed their combined hours? |
| 09:19.13 | seebs | The whole thing was built around one of the many basic-like things, plus a proprietary network database. |
| 09:19.25 | seebs | The database was just a bad network wrapper around sqlite. |
| 09:19.27 | ZorbaTHut | unrelated: my scheme is working! my cat kept coming over and nosing under the monitor and wanting to hang out next to me, but he couldn't because the monitor stand was in the way |
| 09:19.44 | ZorbaTHut | so I got a VESA desk-mount monitor arm and now he can curl up on the desk next to where I'm working and sleep ^^ |
| 09:19.48 | Ivnedar | Expecting to wake up to a crashing monitor later in the night? |
| 09:19.52 | seebs | The entire thing I did ended up being in PHP because it was the only language whose bindings I could get to actually work. |
| 09:19.56 | Imhothar | he/she will find a way to annoy you nonetheless |
| 09:20.03 | ZorbaTHut | oh I'm sure he will, he's a cat |
| 09:20.04 | Ivnedar | ooh, nice |
| 09:20.05 | ZorbaTHut | and a weird one at that |
| 09:20.19 | ZorbaTHut | but this way at least he'll be cuter and happier while doing so |
| 09:20.34 | seebs | The people who made it built their PHP bindings against a specific non-standard PHP build, then insisted it was a "linker bug" in Mac OS X that it couldn't be used with the system-provided PHP due to undefined symbols. |
| 09:20.35 | Imhothar | well 1/3 of all cats are sadistic murderers, so if your is just weird, be glad about it and shut up ;) |
| 09:20.56 | seebs | Also, they *did not know* how to escape special characters to prevent SQL injections in their API. |
| 09:21.03 | seebs | Their proprietary API that they defined themselves. |
| 09:21.22 | seebs | I was SO happy the day I got that !@#(!#( thing working with postgres. |
| 09:21.56 | seebs | Zorba, I've had similar issues. I recently moved things on one of my desks to provide a box with one of my old shirts in it so my cat could come sleep near me. |
| 09:22.15 | ZorbaTHut | I'm actually using my pants as a bed right now because I need to buy a cat bed for him |
| 09:22.21 | ZorbaTHut | but I suspect I'll be using old shirts now and then too |
| 09:22.21 | seebs | The trick is, when I first set up that box lid, I would put her near it, and try to block her from investigating it, or remove her from it. |
| 09:22.44 | seebs | Yeah, if you get a cat bed, be sure to try to keep the cat out of it a bit at first. |
| 09:23.06 | seebs | Most cats will not even CONSIDER sleeping in something if they think they're supposed to. |
| 09:23.14 | ZorbaTHut | our cats are actually pretty good about that |
| 09:23.20 | ZorbaTHut | we got a pair of cat beds and they just gravitated there instantly |
| 09:23.39 | Imhothar | cat: "animal psychology my ass" |
| 09:23.43 | ZorbaTHut | he'll happily climb up and down the cat tree and he even plays with toys that we bought him |
| 09:23.45 | Imhothar | dog: "food?" |
| 09:23.49 | ZorbaTHut | (you know, things that are actually intended to be toys) |
| 09:23.51 | seebs | I have occasionally succeeded in tricking them with catnip. |
| 09:24.16 | seebs | Yeah, my current cat is pretty good about that, mostly. But she's not actually a cat, she's some kind of... cross between a cat and a domesticated animal. |
| 09:24.39 | Imhothar | you into genetics experiments? |
| 09:24.49 | seebs | Not especially. |
| 09:24.54 | seebs | I thought she was a cat. |
| 09:24.59 | Imhothar | hmm though then you should know whtat kind of animal that other thing was |
| 09:25.04 | seebs | But I am reliably assured that cats do not behave like this. |
| 09:25.12 | seebs | So she must be something else. |
| 09:25.44 | ZorbaTHut | our cats actually sleep with us in bed, which is adorable. another of those "someone forgot to tell them they were cats" deals |
| 09:26.36 | seebs | Ours do too. |
| 09:26.58 | seebs | Mine occasionally comes and sleeps on me, mostly she carefully arranges to be JUST outside of arm's reach and then look sad and lonely because no one will pet her. |
| 09:27.06 | seebs | Oh, but! |
| 09:27.10 | Imhothar | having your pets in bed at night can have a bad effect on your sleep. just saying |
| 09:27.13 | seebs | She will only do this if I have a night light. |
| 09:27.40 | seebs | If it is dark she runs away and goes somewhere else. Usually the windows in front of the house, facing the street light. |
| 09:27.52 | seebs | See, that's the other way we know she's not a real cat. |
| 09:27.58 | seebs | Because she is *afraid of the dark*. |
| 09:28.09 | Imhothar | someone should tell her she's supposed to be nocturnal |
| 09:28.49 | seebs | Yeah, I think she's not having any of that. |
| 09:29.00 | seebs | And yeah, cats can be disruptive of sleep, but also soothing. |
| 09:29.03 | Imhothar | maybe she read a different book about cats than I did |
| 09:29.20 | seebs | I have a really, really, hard time sleeping in general -- my brain won't stop spinning until I'm utterly wiped out. |
| 09:29.34 | seebs | Cats have a nap-inducing effect on me, and once that kicks in I can sleep decently. |
| 09:29.48 | Imhothar | same here, that's why I stay up for as long as I can, until I know I drop asleep immediately |
| 09:30.42 | Imhothar | otherwise it can take hours |
| 09:31.05 | seebs | I also try to get some caffeine in me, because thanks to Weird Brain Chemistry, I have a much easier time sleeping when there's enough stimulants in my system for my brain to work. |
| 09:35.43 | Imhothar | I actually tru to get off caffeine, to get my brain back to normal |
| 09:36.31 | Imhothar | though my coke-only based caffeine consumption is nothing compared to some coffee abusers |
| 09:37.57 | seebs | I have pretty light (compared to serious coffee people) caffeine, but then, they also give me amphetamines for the brain. |
| 09:40.14 | Noshei | hey ZorbaTHut, are you aware of a bug with nameplates not always showing up for all of the mobs you are incombat with. |
| 09:40.47 | ZorbaTHut | huh, new one on me |
| 09:41.29 | Noshei | I've got Hostile set to incombat, but for some reason a lot of the time they only show up for my target |
| 09:41.52 | ZorbaTHut | strange |
| 09:42.59 | Noshei | I usually have to restart the client to get it fix itself |
| 09:43.17 | seebs | Oh, hey, Zorba, wanna see a cool bug in the Rift UIZ |
| 09:43.19 | seebs | UI? |
| 09:43.27 | seebs | (still getting used to my keyboard swap) |
| 09:43.29 | ZorbaTHut | sure :) |
| 09:43.41 | seebs | Okay, set the default bars to NOT always-show. |
| 09:44.02 | seebs | You will note that if you drag an icon, the empty button slots all appear. |
| 09:44.39 | seebs | Now, if you mess around a bit with clicking buttons at a high speed, while moving, you will occasionally be able to produce a fascinating behavior: The game thinks you're dragging an icon, but then realizes you were clicking. |
| 09:44.44 | seebs | The empty button slots do not go away. |
| 09:45.04 | seebs | You can make them go away by doing something which is unambiguously dragging, even if you end up just putting the button back. |
| 09:45.23 | ZorbaTHut | yeah, this does not surprise me, you're probably hitting a race condition or a state transition we forgot about |
| 09:45.32 | seebs | Also, it would be LOVELY to have an option similar to WoW's for the bars, wherein "locked" means "can be dragged only by using a modifier key". |
| 09:46.02 | seebs | Because as-is, unlocked is often-dangerous (too easy to move or lose a button), but locked is a hassle because you have to stop completely and go into the UI to change settings to make ANY changes at all. |
| 09:46.12 | ZorbaTHut | huh, not a bad idea, but probably a bit low-priority |
| 09:46.20 | ZorbaTHut | I think there's a lock icon on the main action bar that can be clicked |
| 09:46.22 | ZorbaTHut | I may be wrong on that |
| 09:46.23 | seebs | I think I reported this two or three times during beta, but at the time I couldn't figure out what was happening. |
| 09:46.55 | seebs | I think there might be. I wouldn't have noticed, I don't see the decorations. |
| 09:47.12 | Noshei | there is |
| 09:47.29 | Noshei | but that "main" bar is an ugly pain that most people hide |
| 09:47.43 | ZorbaTHut | true |
| 09:48.01 | ZorbaTHut | I'm hoping to get the infrastructure together needed to make custom action bars in the addon system, but that's still a ways away |
| 09:48.14 | seebs | My other complaint about the built-in bars: Even with show all off, empty buttons still eat clicks. |
| 09:48.19 | Noshei | I really wish they would give us an option to designate a "main" bar for things like the stealth and vehicle bars |
| 09:48.40 | seebs | I would love to be able to make custom bars. No one in WoW ever did quite what I wanted, although Macaroon was pretty close. |
| 09:48.54 | Noshei | it is really annoying having to enable the "main" bar every time i do storm breaker |
| 09:49.00 | seebs | Part of that depended on their much... hmm. "Smarter" isn't quite right. Their macro system. |
| 09:49.26 | seebs | Which is not NEARLY as situationally aware and reactive as Rift's, but a lot better about modifiers and showing the right tooltips. |
| 09:50.09 | seebs | Rift's would be pretty close if it had [mod:none], and the tooltip could show "the thing that will actually be triggered if you hit this button". |
| 09:51.45 | Imhothar | and the icon |
| 09:51.51 | Imhothar | the icon is more important than tooltip |
| 09:51.58 | seebs | Yeah. |
| 09:52.50 | seebs | For 5-player healing, I had a row of buttons such that the first button was @party1, and would cast different spells with nomod, shift, ctrl, or alt. Super handy. |
| 09:53.23 | seebs | Of course, the thing I most wanted a nomod setting for was a pet-healing macro which became irrelevant when necromancers lost their two targetable direct heals. |
| 09:54.33 | seebs | I gotta go sleep. My builds are running for testing on a freshly booted machine that hasn't got a *text editor* stuck in an unkillable wait now. |
| 09:55.20 | ZorbaTHut | 'nite :) |
| 09:55.49 | Ivnedar | wish the crucia doesn't reset bug would be fixed: ( |
| 09:55.57 | Ivnedar | sucks only getting 2-3 attempts before having to go do something else |
| 09:57.06 | Noshei | havent had that happen |
| 09:57.17 | Noshei | and we spent an entire night on here the other day |
| 09:57.24 | Ivnedar | we have it happen every week |
| 09:57.37 | Ivnedar | first platform is considered gone even though it isn't |
| 09:57.51 | Ivnedar | zone in or res and you die, crucia laughs at you, etc |
| 09:58.17 | Noshei | oh yeah we had that happen once, but we were able to engage her long enough to reset |
| 09:58.25 | Ivnedar | how? |
| 09:58.43 | Noshei | i think someone soul walked out to her |
| 09:59.43 | Ivnedar | ooh, may have to try that, thanks :) |
| 10:00.25 | Ivnedar | you can tell she hasn't reset because if you had her focus'd, she's still the same unitid |
| 10:00.33 | Ivnedar | instead of normally resetting and getting a new unitid |
| 10:00.51 | Noshei | yeah |
| 10:28.52 | *** join/#riftuidev Imhothar (miranda@ppp-93-104-177-233.dynamic.mnet-online.de) |
| 10:35.29 | *** join/#riftuidev Torhal|work (~torhal_wo@76.164.170.2) |
| 10:35.29 | *** mode/#riftuidev [+v Torhal|work] by ChanServ |
| 10:46.41 | Ivnedar | Huzzah, wasn't standing on the portal that Crucia kill |
| 10:57.38 | Ivnedar | And forgot to watch to see what happens ... |
| 10:57.39 | Ivnedar | oh well |
| 13:08.34 | *** join/#riftuidev g0urra (~g0urra@unaffiliated/gourra) |
| 15:23.17 | *** join/#riftuidev odine (odine@cpe-74-77-173-83.buffalo.res.rr.com) |
| 17:12.06 | odine | ZorbaTHut: add a PLAYER_ENTERING_WORLD event while ur at it with the event system... thus we can elimnate the need to always check if the player unit is avail, and it would signal all addons are loaded as well etc etc |
| 17:12.31 | Imhothar | and then strikes the loading screen |
| 17:12.32 | ZorbaTHut | there is one, it's hooking "player" with libunitchange |
| 17:12.54 | odine | hm, didnt realize that |
| 17:14.07 | ZorbaTHut | that said, it may not be as thorough as you might want |
| 17:14.21 | ZorbaTHut | rift's loading is kind of informal, things sorta show up when they feel like it |
| 17:14.21 | odine | indeed, but i suppose it works |
| 17:15.13 | odine | i mean i dont really mind checking if unit is fully avail, so ill prolly just stick with that, cuz i also dont wanna rely on libunitchange for everything |
| 17:15.43 | odine | cant wait to see this new event system in action |
| 17:16.06 | ZorbaTHut | yeah, I'm looking forward to it. the old one actually feels obsolete to me already |
| 17:16.13 | odine | hehe |
| 17:16.38 | odine | ya know i never really looked into this.. is there a way to tell the difference between a live client and pts client thorugh the API? |
| 17:17.15 | ZorbaTHut | hmmm. I believe you could check the shard name via Inspect.Shard(). it's also possible that Inspect.System.Version() has differences |
| 17:17.25 | ZorbaTHut | but besides shard name, I suspect you won't find any differences |
| 17:17.37 | odine | ya im gonna give that a shot and see what i come up with |
| 17:18.54 | ZorbaTHut | for hopefully obvious reasons, we try to test as close to a release environment as is practical :) |
| 17:19.28 | odine | ya |
| 17:19.56 | *** join/#riftuidev g0urra (~g0urra@unaffiliated/gourra) |
| 17:20.17 | odine | oh the joy of rewriting everything for the new event system... only downfall =P |
| 17:20.44 | Imhothar | how are native events integrated into the new thing? |
| 17:20.48 | ZorbaTHut | hopefully it won't be too complicated. non-frame-event names are still the same, and parameter ordering for events is also the same (with a single new parameter added at the beginning) |
| 17:21.12 | odine | goood question imho |
| 17:21.16 | Imhothar | native frame events that is |
| 17:21.35 | ZorbaTHut | imhothar, as in, "events on native frames", or "the events that frames trigger on internally"? |
| 17:21.38 | odine | ya i dont suspect it being difficult at all, just more of an....... cant quite find the right word |
| 17:21.51 | ZorbaTHut | odine, "timeconsuming aggravation"? :) |
| 17:21.57 | odine | there ya go ;) |
| 17:22.39 | Imhothar | the former |
| 17:23.22 | ZorbaTHut | imhothar, no changes, really. events on native frames never did anything particularly out of the ordinary |
| 17:23.50 | Imhothar | apart for Event.Loaded |
| 17:23.54 | Imhothar | which everyone wants to use |
| 17:24.31 | ZorbaTHut | well, yeah. this will help that because multiple handlers will now work totally fine |
| 17:24.52 | Imhothar | so the events are now lists? |
| 17:25.27 | ZorbaTHut | they end up being directed to rather complicated C++-side data structures. In effect, though, yes, they can store an arbitrary number of handlers. |
| 17:25.47 | ZorbaTHut | (there needs to be a bit of black magic to make sure it works properly when people inevitably add/remove events from a handler while the event is mid-trigger) |
| 17:26.20 | Imhothar | like, a copy-swap of lists? ;) |
| 17:26.47 | ZorbaTHut | too slow and the code is fiddly. plus it's really weird if you remove an event that hasn't yet fired and then it fires anyway |
| 17:27.09 | Imhothar | details |
| 17:28.16 | Imhothar | well, if appending is always at the end, all you really gotta do is re-evaluate end() at each iteration |
| 17:28.29 | Imhothar | then removing is a no-brainer as well |
| 17:28.31 | ZorbaTHut | unfortunately it's not - people can set the priority of events |
| 17:28.41 | Imhothar | basically iterating a list that ios mutable while iterating |
| 17:28.42 | ZorbaTHut | and of course you can always remove from the middle |
| 17:28.53 | ZorbaTHut | including the very event handler that is in the middle of being triggered :V |
| 17:28.55 | Imhothar | ah prioritues |
| 17:29.32 | Imhothar | as long as removing/adding doesn't invalidate the current iterator |
| 17:29.47 | Imhothar | (unless the object being removed is active ofc) |
| 17:37.22 | Baanano | ZorbaTHut: do you believe the texture coordinates stuff will be added before the old event system is removed? |
| 17:39.33 | ZorbaTHut | Hmm. No guarantees, of course, but probably. |
| 17:41.24 | Baanano | great news :) didn't want to rewrite all the GUI related code twice :P |
| 18:14.37 | Imhothar | wonder whether these claims about an arrest warrant against Benedikt 6 days before he resigned turn out to be true or not |
| 18:24.49 | Baanano | mmm, can't edit thread from last year, and i'd rather leave that horrible "BananAH isn't working" thread die, guess I'll have to create a new thread |
| 18:27.42 | Imhothar | hmm I need a recent book on Quantum Mechanics, I'm getting rusty |
| 18:30.40 | Imhothar | so much to do, so much to read, so little time for anything |
| 18:33.46 | Baanano | are you a physics major, imhothar? or do you just enjoy reading that? |
| 18:33.57 | Imhothar | the latter |
| 18:34.27 | Imhothar | though I was seriously thinking about adding an astrophysics major to my informatics |
| 18:35.45 | Imhothar | who needs a PhD these days |
| 18:39.59 | Imhothar | besides that, my curiosity is a real curse. I'd need several llifetimes to learn all the thing which are of interst to me |
| 18:40.35 | Imhothar | and possibly a brain attached storage or something |
| 18:42.08 | Baanano | http://xkcd.com/214/ ? |
| 18:42.50 | Imhothar | replace "three" with "twelve" and you get closer |
| 18:43.00 | Imhothar | it happens |
| 18:46.40 | Baanano | that's a major productivity killer |
| 18:46.52 | Imhothar | indeed |
| 18:50.40 | Imhothar | well besides that, work on a game set in space got me back into astrophysics |
| 18:53.46 | Baanano | is it a simulator? |
| 18:54.18 | Imhothar | celestial mechanics and relativistic effects are part of it, aye |
| 18:55.28 | Imhothar | but it's still a game, not a simulator |
| 18:57.13 | Baanano | sounds interesting, can you ellaborate it further? |
| 18:59.10 | Imhothar | not sure |
| 19:06.41 | odine | ne1 mind tripple checking something for me....go to http://biscal.org click forums and tell me if u can see Devs Only |
| 19:07.11 | Imhothar | negative |
| 19:07.15 | odine | i have joomla and phpbb bridged and im having a few small issues so i need to make sure of some tings |
| 19:07.18 | odine | thanks imho |
| 19:10.26 | Imhothar | astrophysics is awesome. it takes two neutron stars only 40ms to collide, eject massive amounts of energy and heat, and form a black hole |
| 19:11.56 | Imhothar | if we were unfortunate enough to cross that gamma burst, the entire athmosphere would just vaporize within seconds |
| 19:14.18 | Imhothar | oh the beauty |
| 20:03.37 | *** join/#riftuidev Imhothar_ (miranda@ppp-188-174-155-78.dynamic.mnet-online.de) |
| 20:29.17 | *** part/#riftuidev odine (odine@cpe-74-77-173-83.buffalo.res.rr.com) |
| 20:32.33 | *** join/#riftuidev odine (odine@cpe-74-77-173-83.buffalo.res.rr.com) |
| 21:18.04 | *** join/#riftuidev Noshei (~quassel@c-24-9-171-99.hsd1.co.comcast.net) |
| 21:21.18 | *** join/#riftuidev Noshei (~quassel@c-24-9-171-99.hsd1.co.comcast.net) |
| 21:32.51 | *** join/#riftuidev seebs (~seebs@67-5-22-174.rstr.qwest.net) |
| 21:32.51 | *** mode/#riftuidev [+o seebs] by ChanServ |
| 23:10.20 | *** join/#riftuidev seebs (~seebs@home.seebs.net) |
| 23:10.20 | *** mode/#riftuidev [+o seebs] by ChanServ |