| 00:03.13 | seebs | Ooh, neat idea. |
| 00:03.30 | seebs | I like that and may steal it. |
| 00:04.18 | Dramiel1 | please do, so I don't have to steal your idea ;p |
| 00:05.51 | Dramiel1 | even tho I probably don't want to see performance results of my framework |
| 00:35.37 | *** join/#riftuidev Adys (~adys@unaffiliated/adys) |
| 01:10.36 | ZorbaGama | seebs, if you haven't looked at Inspect.Addon.Cpu(), you should - it already times individual event handlers |
| 01:10.56 | ZorbaGama | that said, Inspect.Time.Real() is a high-resolution system timer and should be enough for any sane benchmarking |
| 01:11.26 | seebs | Why am I not surprised that you are ahead of me? Ah, well, a general-purpose time counter might still be useful. |
| 01:11.34 | seebs | And I did find a name for it. It shall be... perf-o-rate! |
| 01:11.40 | seebs | Spelled perforate. |
| 01:11.50 | ZorbaGama | it's the same data that's used to generate the system tooltip, but in substantially more detail :) |
| 01:12.34 | seebs | hmm |
| 01:12.49 | seebs | I am not seeing how to get individual event handler information out of inspect.addon.cpu. |
| 01:13.06 | seebs | Say I have two event handlers for two different events, how do I tell which time is going to which of them, assuming they are part of the same addon? |
| 01:13.07 | ZorbaGama | it's a table of tables - each addon has its own table |
| 01:13.14 | seebs | ahhhh |
| 01:13.24 | ZorbaGama | and each of the subtables includes a breakdown, with the time credited to whatever the identifier string in the event handler is |
| 01:13.27 | seebs | I see. |
| 01:13.31 | ZorbaGama | in fact, that's one of the main reasons that identifier string exists ;) |
| 01:13.45 | seebs | oh-hoh! |
| 01:13.46 | seebs | Very nice. |
| 01:14.35 | ZorbaGama | note that it only accumulates every frame, but for most uses that should be sufficient. |
| 01:17.50 | ZorbaGama | unrelated: I'm really excited about the 1.6 addon features :3 finally things are getting fleshed out |
| 01:18.03 | ZorbaGama | the item api was one of the few really big and scary APIs I'd have to deal with, and it's basically nailed down by now. |
| 01:20.35 | seebs | yay! |
| 01:20.43 | seebs | Your enthusiasm is very encouraging. |
| 01:21.13 | ZorbaGama | yeah, I've gone from "well, I totally think this is the right way to make an addon system, hopefully I'm not missing something huge and catastrophic" to "yep, this is totally the right way to make an addon system" |
| 01:21.30 | seebs | It certainly seems that way to me. |
| 01:22.03 | seebs | hmm |
| 01:22.21 | seebs | Okay, so I added an addon. And I copied it into addons, then started rift... and it doesn't show up in the addons window list. |
| 01:22.38 | seebs | Is there a log file anywhere that might have information like "found a directory but it was obviously malformed"? |
| 01:22.43 | ZorbaGama | unfortunately no |
| 01:23.05 | seebs | oooh, I bet I know what's wrong -- I didn't copy the embedded lib in. |
| 01:23.06 | ZorbaGama | something I've meant to add, but it's not entirely clear where it should go |
| 01:23.28 | ZorbaGama | huh, if that's actually the problem it should probably not cause a vanishing addon |
| 01:23.46 | seebs | Maybe that's not it. Hmm. |
| 01:24.16 | ZorbaGama | personally I just copy from an existing addon, then modify one line at a time while hammering "refresh" |
| 01:24.20 | ZorbaGama | it's not elegant but it works |
| 01:24.24 | seebs | hee |
| 01:24.39 | seebs | Well, let's see. |
| 01:24.55 | seebs | ohhh |
| 01:25.04 | seebs | heh, I didn't even come CLOSE on the syntax for embed = and dependencies = |
| 01:25.09 | ZorbaGama | that would do it :V |
| 01:26.17 | seebs | I had Dependencies = { "Foo" } instead of { Foo = { "required" } } |
| 01:43.06 | seebs | And yes, Inspect.Time.Real does what I want it to do, yay! |
| 02:34.29 | seebs | Is it safe/legit to *remove* an event handler, and if so, how? Say I want to insert something in Event.System.Update.Begin temporarily... Can I, or do I just have to insert something which then calls something else? |
| 02:41.26 | *** join/#riftuidev hp94 (4495e6ff@gateway/web/freenode/ip.68.149.230.255) |
| 02:41.50 | ZorbaGama | event tables are literally just tables. when an event triggers, it traverses the table with ipairs() and calls all acceptable events attached to it. |
| 02:42.06 | ZorbaGama | you're welcome to do whatever you like to that table as long as ipairs() will still work :) |
| 02:42.16 | seebs | hmm |
| 02:42.28 | seebs | So I could scan it for things that I know to be mine, say, that have my addon ID. |
| 02:42.42 | ZorbaGama | *nod* yup. This is another thing the event-hook-identifier label is for |
| 02:43.01 | ZorbaGama | if it's your identifier, and your label, and it's really someone else masquerading as you . . . I sorta don't have much sympathy for them |
| 02:43.01 | seebs | woo |
| 02:43.07 | seebs | I have now got: |
| 02:43.27 | seebs | <PROTECTED> |
| 02:43.35 | seebs | And after about 100 update events, it says: |
| 02:43.46 | seebs | Library.LibPerfORate.wastetime(): 0.00.001218/100 (stopped) |
| 02:44.16 | seebs | So that tells me that a hundred repeats of .wastetime() took about 13ms. |
| 02:44.43 | seebs | I'm doing the benchmark as an update hook that runs your code once per update rather than running it repeatedly in a single update, so it's less intrusive. |
| 02:49.58 | seebs | FWIW, a shallow copy of _G performed on every update is barely measurable in the addon CPU display. |
| 03:06.20 | hp94 | Has anyone here heard of bitcoins? |
| 03:12.28 | seebs | ... Apparently I'm having a stupid day. Programming In Lua seems to suggest that arguments after any named arguments are automatically stashed in a table named 'arg'. |
| 03:12.40 | seebs | But if I try to refer to this, I get "attempt to index global 'arg' (a nil value)". |
| 03:12.42 | ZorbaGama | I think that's 5.0 behavior, 5.1 doesn't do that |
| 03:13.17 | seebs | ahh, that would explain why there were references to the 5.1 "..." |
| 03:13.21 | ZorbaGama | lua> function extraparams(a, b, c, ...) print(select("#", ...)) print(select(2, ...)) end extraparams(1, 2, 3, 4, 5, 6) |
| 03:13.22 | lua_bot | ZorbaGama: 3, 5, 6 |
| 03:13.40 | ZorbaGama | er, dur, that's not quite right |
| 03:13.47 | ZorbaGama | wait no it is |
| 03:13.55 | ZorbaGama | I guess select returns everything after that member as well |
| 03:14.06 | ZorbaGama | anyway, the "#" one gives you the number of params |
| 03:14.25 | seebs | In this case, what I want to do is just stash the values. |
| 03:14.47 | seebs | What I am doing is basically foo.stashed_args = { ... }, and later foo.stashed_function(unpack(foo.stashed_args)) |
| 03:15.04 | ZorbaGama | well, that's valid code, so, there ya go :) |
| 03:15.10 | ZorbaGama | though keep in mind it won't properly stash any trailing nils |
| 03:15.43 | hp94 | unpack? |
| 03:15.46 | seebs | I don't actually care, because if they're omitted, and something was expecting them, it'll get nils. I think. |
| 03:15.59 | hp94 | I should probably google things instead of ask about them |
| 03:16.03 | ZorbaGama | not entirely true - lua will pad extra parameters with nils, yes |
| 03:16.13 | ZorbaGama | but it's possible to get the exact parameter count of a function, and that won't be preserved |
| 03:16.18 | seebs | ahh, true. |
| 03:16.27 | ZorbaGama | many of the rift API calls actually pay attention to trailing nils |
| 03:16.33 | ZorbaGama | well, not many |
| 03:16.33 | ZorbaGama | a few |
| 03:18.20 | seebs | Mostly this is so you can do Library.LibPerfORate.benchmark(function, 1000, arg1, arg2) |
| 03:18.32 | seebs | and you'll get timing of calling function(arg1, arg2) 1000 times. |
| 03:44.40 | seebs | http://www.riftui.com/downloads/info135-LibPerfORate.html <-- first pass up, yay! |
| 03:45.48 | lXl | Here's all the addons I've managed to write: |
| 03:46.01 | lXl | :( |
| 03:48.11 | seebs | That is because you are an inherently bad person. |
| 03:48.18 | lXl | Yes I am |
| 03:48.23 | seebs | I don't remember for sure why I wanted this one anymore. |
| 03:48.28 | lXl | Lol |
| 03:48.31 | seebs | There was SOMETHING I wanted to benchmark. |
| 03:48.38 | seebs | And then I got fascinated with how I could make a benchmark. |
| 03:48.52 | seebs | And then I COMPLETELY FORGOT whatever it was that I'd been planning to benchmark. |
| 03:49.38 | seebs | Might well be that it was something to do with my vague plans of writing a bag addon when there is an API for it. |
| 03:51.15 | seebs | I've already named it LootSorter, tentatively. So I can use /ls and have a properly intuitive bag interface. |
| 04:53.29 | starlon | Isn't the bags API on PTS? |
| 04:55.33 | seebs | I don't have a PTS setup right now. |
| 04:56.41 | seebs | And I'm not sure, but I think the bag/item API may not be quite completely published to PTS yet. |
| 05:02.02 | starlon | nods. |
| 05:02.12 | starlon | I don't have a PTS client installed either. |
| 05:02.15 | seebs | Anyway, I don't have time for a big project. :) |
| 05:02.21 | starlon | heh |
| 05:02.42 | seebs | Thinking about it. |
| 05:02.55 | seebs | Measuring real time during a function rather than CPU may actually be useful sometimes. |
| 05:03.14 | seebs | The addon API itself isn't multithreaded, but it may be that it can call out to things (like GPU actions) that take time but not CPU time. |
| 05:04.24 | starlon | if you're trying to get me to think, I have to warn you... I'm in unthinking mode. |
| 05:04.27 | seebs | So it is totally conceivable to end up caring more about real time anyway. |
| 05:04.38 | seebs | Quick! Queue as DPS! |
| 05:04.39 | starlon | hehe |
| 05:04.44 | starlon | lol |
| 05:43.28 | hp94 | =S |
| 08:26.21 | *** join/#riftuidev Kreiri (~user@193.108.249.252) |
| 09:04.40 | *** join/#riftuidev Dramiel (~Adium@134.140.broadband5.iol.cz) |
| 09:49.05 | *** join/#riftuidev Dramiel (~Adium@134.140.broadband5.iol.cz) |
| 09:53.54 | *** join/#riftuidev Snstr (~darki@p4FCDDEC5.dip.t-dialin.net) |
| 10:15.39 | *** join/#riftuidev Dramiel (~Adium@134.140.broadband5.iol.cz) |
| 11:19.45 | *** join/#riftuidev Kreiri (~user@193.108.249.252) |
| 11:20.10 | *** part/#riftuidev Kreiri (~user@193.108.249.252) |
| 13:59.28 | *** join/#riftuidev Mcbooze (~mcbooze@d51A4529D.access.telenet.be) |
| 14:16.08 | *** part/#riftuidev Mcbooze (~mcbooze@d51A4529D.access.telenet.be) |
| 14:32.32 | *** join/#riftuidev Dramiel1 (~Adium@134.140.broadband5.iol.cz) |
| 15:10.50 | *** join/#riftuidev newbie (~sylvanaar@tut.qvssoftware.com) |
| 15:12.19 | *** join/#riftuidev sylvanaar_work (~sylvanaar@pdpc/supporter/active/sylvanaar) |
| 16:51.50 | *** join/#riftuidev Dramiel (~Adium@134.140.broadband5.iol.cz) |
| 17:14.23 | *** join/#riftuidev BunnyHarbinger (~BunnyHarb@209.237.253.45) |
| 17:30.06 | *** join/#riftuidev Dramiel1 (~Adium@134.140.broadband5.iol.cz) |
| 17:37.57 | *** join/#riftuidev Marru_Laptop (~IceChat77@049-206.pioneernet.du.edu) |
| 17:38.58 | Marru_Laptop | sup guys |
| 17:41.00 | seebs | I made a performance metering thing. |
| 17:41.16 | Marru_Laptop | whats it do? |
| 17:41.35 | seebs | Provides timers, mostly. |
| 17:41.48 | Marru_Laptop | ah lol |
| 17:41.53 | seebs | And also a thing where you can wrap a function in a timer and get a count of real time elapsed doing that function. |
| 17:42.29 | Marru_Laptop | cool |
| 17:46.25 | Marru_Laptop | ugh |
| 17:46.31 | Marru_Laptop | stupid internet |
| 17:59.54 | *** join/#riftuidev Marru_Laptop (~IceChat77@049-206.pioneernet.du.edu) |
| 17:59.58 | Marru_Laptop | hey |
| 18:00.01 | Marru_Laptop | so if i have a table |
| 18:00.08 | Marru_Laptop | and i want to change one number in it |
| 18:00.22 | Marru_Laptop | say its just color = {1,1,1,1} |
| 18:00.41 | Marru_Laptop | i want to get or change each number individually, how would i do that? |
| 18:00.52 | Marru_Laptop | are they auto indexed with 1,2,3,4? |
| 18:05.19 | seebs | 0-3 |
| 18:05.33 | Marru_Laptop | lua doesn't index with 0 i thought? |
| 18:05.45 | Marru_Laptop | thought it all starts with 1 in lua? |
| 18:05.53 | seebs | ... maybe you're right. I could have forgotten, and that would probably explain why my lua code is so full of weird off-by-one errors. :) |
| 18:06.03 | Marru_Laptop | haha |
| 18:06.50 | seebs | it turns out I never need to know 'cuz of ipairs. |
| 18:07.11 | Marru_Laptop | ipairs? |
| 18:08.48 | seebs | for i,v in ipairs(table) do ... done |
| 18:09.00 | seebs | Since that works, I don't have to know the rules for table indexes. |
| 18:09.09 | Marru_Laptop | lol |
| 18:09.16 | Marru_Laptop | you do if u only want the first value :) |
| 18:09.19 | Marru_Laptop | well |
| 18:09.27 | seebs | I was raised by mathematicians. I don't know things I can derive trivially, in general. |
| 18:09.31 | Marru_Laptop | i guess u could do for ipairs and just break in first loop lol |
| 18:09.49 | seebs | that would be horrible |
| 18:10.34 | Marru_Laptop | i know lol |
| 18:10.46 | Torhal|work | Lua does indeed start indexing at 1 |
| 18:10.50 | Torhal|work | 0 is userdata |
| 18:11.02 | Marru_Laptop | userdate? |
| 18:11.06 | Marru_Laptop | *userdata? |
| 18:11.32 | Torhal|work | Stuff passed in from C/C++ |
| 18:11.48 | Torhal|work | Since Lua is embedded in the Rift client, which is not itself written in Lua. |
| 18:12.14 | Marru_Laptop | ah gotcha |
| 18:26.24 | starlon | tricksy Lua |
| 18:26.48 | Torhal|work | rewrites starlon in Lua. |
| 18:31.03 | Marru_Laptop | anyone wanna fix my addon for me? :) |
| 18:33.16 | Marru_Laptop | ? |
| 18:33.50 | Marru_Laptop | we need to rewrite lua_bot so that it can just write lua programs for us |
| 18:46.17 | *** join/#riftuidev Lyme (~Jason@S010600265ac734e1.du.shawcable.net) |
| 18:46.17 | *** join/#riftuidev Morv_ (~marv@playground.hostin.is) |
| 18:46.17 | *** join/#riftuidev ZorbaGama (tocohl@173.228.13.217) |
| 18:47.49 | *** join/#riftuidev RagebeastRT (44aaf4b4@gateway/web/freenode/ip.68.170.244.180) |
| 18:48.19 | RagebeastRT | Anyone around for a question? |
| 18:48.32 | Dramiel1 | depends on the question |
| 18:48.53 | RagebeastRT | Is it possible to grab a frame name and assign it to a variable? |
| 18:49.14 | Dramiel1 | local name = frame:GetName(); |
| 18:50.13 | RagebeastRT | does that retrieve the variable name or the name set in the Createname function? |
| 18:50.35 | RagebeastRT | pText[i] = UI.CreateFrame("Text","Name1"..i, windowlFrame) |
| 18:51.16 | RagebeastRT | I want to be able to pull a number value out of either the variable name or the interal Createfram name |
| 18:52.45 | Dramiel1 | GetName will return second parameter given to UI.CreateFrame .. |
| 18:52.56 | Dramiel1 | dunno what else you need and why would you want to do it like that |
| 18:53.17 | RagebeastRT | trying to set the text in a tooltip out of an indexed variable |
| 18:53.49 | RagebeastRT | by mousing over a text frame, I need to get the name, strip the # out of it and then reference the text string |
| 18:54.36 | Dramiel1 | why don't you just save the index somewhere else? :p |
| 18:54.46 | Dramiel1 | like pText[i].mightyIndex = i; |
| 18:55.56 | RagebeastRT | I did save the texts to an index. I am using dOxxx's library for widgets. THe tootip widget works fine if I use static text, if I try to mouse over an indexed frame, it doesnt know where it is |
| 18:56.01 | RagebeastRT | if the makes sense |
| 18:56.23 | Dramiel1 | nope |
| 18:56.26 | RagebeastRT | hmmm |
| 18:56.53 | RagebeastRT | ok, I have a display of pText[i] text frames |
| 18:57.30 | RagebeastRT | it only holds the Name of a rune in that frame |
| 18:58.13 | RagebeastRT | sigh, not sure how to explain this |
| 18:58.24 | Dramiel1 | pastebin.com |
| 19:04.25 | seebs | I think there should be an escalation mode for library of the runemasters called librarian of the runemasters where there's librarians running around silencing people and smacking them for fighting. |
| 19:19.54 | Dramiel1 | is it bad that when I read "library" I automatically assumed it was in reference to programming and that the rest of your sentence doesn't make sense? |
| 19:20.11 | seebs | hee |
| 19:20.28 | seebs | It's a warfront, I guess. <-- does not PvP, has only the vaguest notion of what people are talking about. |
| 19:37.21 | RagebeastRT | Got it working without having to try and retrieve the name. I thought his tooltip function only ran when setting the text but guess not. I took it out of the display loop, set it at the end with its own loop setting the text with my indexed variable and it works |
| 19:38.43 | RagebeastRT | Now. my addon uses lots of tables. Once I run the loop for accessing a table, do I need to somehow unload the table to keep from addon memory accumalating? |
| 19:39.48 | seebs | Not in general, no. |
| 19:39.53 | seebs | Lua has garbage collection. |
| 19:40.10 | seebs | It may be marginally more efficient if you keep and reuse tables, but marginal is the key word in most cases. |
| 19:40.40 | RagebeastRT | Was working on the mod last night over a few hours, refining tables and debugging code. The addon memory kept growing. Not real sure if it was my addon or not though |
| 19:40.48 | RagebeastRT | All the tables are preset |
| 19:42.49 | Dramiel1 | imo not.. reusing tables was "cool" in wow because creating new tables took additional memory because garbage collection was performed once in 30 seconds or so.. now I don't think that clearing table and using it again would be faster than collecting it and creating new (but feel free to measure it, you've got the lib for it! :p ) |
| 19:43.45 | Dramiel1 | on the other hand JITed code should be faster, so really - measure it please, for the community :P |
| 19:57.16 | seebs | My experience so far has been that stuff is Too Fast To Care. |
| 19:57.42 | seebs | I had to use a i=1,100 loop around creating shallow copies of _G before I could see whether it was actually happening. |
| 19:59.19 | seebs | I don't know how fast or slow things are, but I certainly get the impression that they are a LOT faster. |
| 19:59.31 | seebs | Which suggests that a lot of stuff which used to be worth paying attention to now really isn't. |
| 20:03.56 | Torhal|work | seebs: Shallow copy? Are you referring to a reference? :) |
| 20:04.04 | seebs | Not quite THAT shallow |
| 20:04.07 | seebs | a = {} |
| 20:04.14 | seebs | for k, v in pairs(_G) do a[k] = v end |
| 20:04.28 | Torhal|work | That's still just references. You aren't creating memory use. |
| 20:04.42 | Torhal|work | Just the table overhead. |
| 20:04.47 | Torhal|work | 100 tables is nothing. |
| 20:05.10 | seebs | Well, the table and the however-many references in it. Still trivial, obviously. |
| 20:05.14 | Torhal|work | Yar. |
| 20:05.38 | seebs | But I was always told that creating tables was "expensive", and I wanted to see how much, and the answer is "barely measurable". |
| 20:05.43 | Torhal|work | You should start using recycling when something is generating 10k of memory every time it fires. :) |
| 20:06.00 | seebs | This was the benchmark that caused me to add the ms case for things that took less than one second to run. |
| 20:06.22 | Torhal|work | But in general, if you're not churning through tables, it's safe to throw them away after use. |
| 20:08.58 | seebs | In Chatterbox, after a bit of messing around I eventually established that the cost of building a table of 10-30 items every time I was actually going to say something was basically unmeasurable. |
| 20:10.37 | seebs | Huh. I just realized. |
| 20:10.56 | seebs | I could start on Chatterbox now! I wouldn't be able to make it actually SAY things, but I could run it in local print mode. |
| 20:11.06 | seebs | ... Oh, wait. I was gonna wait on a native scrollbar. :) |
| 21:45.33 | Dramiel1 | http://pastebin.com/QaPhG47x what do you think about this api draft for addon that provides configuration window for addons? |
| 21:46.56 | seebs | Oooh, sorta cool! |
| 21:47.08 | seebs | Hmm. |
| 21:47.30 | seebs | I ... I think I would want the defaults/database/configwindow settings to be All One Thing. |
| 21:47.54 | seebs | Because as is, it seems to me that if I add a new value, I need to add it in three places, and they can conceivably be out of sync. |
| 21:48.19 | Dramiel1 | database is synced automatically to what defaults contains |
| 21:48.20 | seebs | Wait, two places. |
| 21:48.27 | Dramiel1 | :) |
| 21:48.36 | seebs | { path = "frames.player.position", renderer = FramePositionRenderer }, <-- this, though... |
| 21:48.52 | seebs | Seems like it might be sensible to structure things so that this information can be in there with the defaults. |
| 21:49.12 | seebs | Maybe make the final level be, not the default value, but { default = val, renderer = FramePositionRenderer } |
| 21:49.33 | seebs | So it'd be enabled = { default = true, renderer = CheckBox } |
| 21:50.07 | seebs | and stuff like description, etc., could be in there too. So one ginormous table that can be used to make the database and the window. |
| 21:50.17 | seebs | IT MUST BE AS I HAVE COMMANDED! |
| 21:50.24 | seebs | <-- best IRC programming advice EVER |
| 21:57.24 | Dramiel1 | this is exactly what I don't want to do :p |
| 21:57.37 | Dramiel1 | I just feel that metadata and actual db structure should be separated |
| 21:57.57 | Dramiel1 | it's like if you were putting css into database tables |
| 21:58.01 | seebs | I don't like having things separated, though, because then they are Two Things, and they can disagree. |
| 21:58.33 | seebs | See, what I'd think you'd want to do is generate the db object *from* the metadata. |
| 21:59.00 | Dramiel1 | db object sure, but the structure definition would still be mixed |
| 21:59.02 | seebs | But the metadata should know things like the conceptual-types of objects -- distinguish between two integers named x and y and a window position for instance. |
| 21:59.11 | Dramiel1 | later you decide my library is crap and you will have to clean it :p |
| 21:59.35 | seebs | I like to have a single complete definition rather than two parallel definitions. |
| 21:59.38 | seebs | It makes me happier. |
| 21:59.48 | seebs | Note the highly technical analysis. :P |
| 21:59.52 | Torhal|work | implements starlon in a mixture of CSS and DOM. |
| 21:59.54 | Dramiel1 | :p |
| 22:00.36 | seebs | argh |
| 22:00.45 | seebs | I'm still at work but I just realized a thing I want to do to PerfORate. |
| 22:03.19 | Dramiel1 | well ok, how would you define config table for the same example - 6 elements in DB, 4 elements in UI, position table (x, y) is considered one element.. I just don't see it |
| 22:08.59 | seebs | hmm |
| 22:09.17 | seebs | Okay, I think the obvious thing is, position table really wants to have some kind of definition of its mapping. |
| 22:09.32 | seebs | Maybe the right thing to do is to say that a position is ALWAYS an { x, y } pair, or something. |
| 22:10.02 | seebs | so { window_position = { type = position, default = { x = 0, y = 0 } } |
| 22:10.14 | seebs | Or something like that. |
| 22:12.38 | Dramiel1 | the value can be table with anything.. because you can write custom renderer for lets say colors |
| 22:12.41 | Dramiel1 | or anything else |
| 22:12.51 | seebs | hmmm |
| 22:13.08 | Dramiel1 | but yes, for position it will most likely be x and y.. maybe anchors but that would be probably too complicated for end user |
| 22:13.13 | seebs | Well, you could specify renderer still in a case like that. I guess... hmm. |
| 22:13.38 | seebs | Yeah, I see the issue. But I think you will be best off if you view each thing which has A Renderer as being a single object from a configuration standpoint. |
| 22:22.05 | Dramiel1 | so you would prefer this: http://pastebin.com/QaPhG47x (see bottom of the file) |
| 22:22.52 | Dramiel1 | I guess it could work.. the only issue I see is that it could be hard to determine what is just another config category and what is actually the final "metadata table" |
| 22:37.02 | seebs | hmm. |
| 22:38.04 | seebs | Yeah, because if you do it this way, you just KNOW that someone is going to come along and specify a thing which lets you configure values named default, renderer, and description. :P |
| 22:39.43 | seebs | And thanks to two-valued defaults like position = { x, y } you can't even just say "anything with subtables is not a leaf". |
| 22:39.54 | seebs | Hmm. |
| 22:42.32 | seebs | Okay, here's my thought. If something has values that you are looking for (default, renderer, description), it is ASSUMED to be a leaf. |
| 22:43.06 | Dramiel1 | well yea, but what worries me is fact that only "default" is mandatory |
| 22:43.36 | Dramiel1 | also I do like conventions, but not assumptions |
| 22:43.49 | seebs | Well, okay, if it has a field named default, that is a leaf, unless (condition). |
| 22:44.09 | seebs | Something like { default = { subtable }, field_named_default = true } |
| 22:44.42 | Dramiel1 | yea but what if someone will define { default = { default = true, description = "haha you fail" } } :p |
| 22:44.48 | seebs | I really do like the way it looks, but I see the issue with the ambiguity. |
| 22:45.11 | seebs | I think my answer would be "in the unusual case where you want to define a member named default, rather than set a default value, you must..." |
| 22:45.36 | Dramiel1 | well I will have to think about it little bit more |
| 22:45.39 | seebs | It's a fairly sane boundary, since the entire point of this is to make it so the user DOESN'T need to define something named default. |
| 22:45.54 | seebs | heyyy. Isn't there a metatable trickery you can do? |
| 22:46.29 | Dramiel1 | with this you will also have to handle "order" manually - which in the separated tables concept is handled automatically.. and categories |
| 22:46.35 | seebs | Like isn't there a way to assign a type-name to a table or something? You could offer a constructor that makes a thing which is definitely a leaf node no matter what its members are named. |
| 22:46.41 | seebs | Ohhh, hmm. |
| 22:46.55 | seebs | You're right, I forgot that Lua isn't a tables-in-order language. |
| 22:46.56 | Dramiel1 | I thought about metatables but that would make the config table really ugly |
| 22:46.57 | seebs | Curses. |
| 22:47.23 | Dramiel1 | which is one of the major features why I consider using one-table layout |
| 22:47.43 | seebs | Hmm. Well, you could do { order = 1 } |
| 22:47.45 | seebs | etc. |
| 22:47.50 | seebs | { default = true, order = 1 } |
| 22:48.04 | seebs | and then have a renderer called something like CategoryHeader which doesn't do anything. |
| 22:48.08 | Dramiel1 | yea I know but again - it is something user will have to manage manually |
| 22:48.14 | seebs | Yeah, it is. Ugh. Hmm. |
| 22:48.28 | Dramiel1 | so it is basically you define stuff on two places or more stuff on one place |
| 22:48.33 | seebs | well, there's an alternative to consider. |
| 22:48.51 | seebs | You do an index-type array for the objects, and use names. |
| 22:48.57 | seebs | So instead of windowposition = { ... } |
| 22:49.06 | seebs | it's just { name = "windowposition", default = ... } |
| 22:49.19 | seebs | and the top-level thing is then an ordered array with implicit keys. |
| 22:49.35 | seebs | *glee* |
| 22:50.00 | seebs | This is pretty much the thing I missed most about wow, hanging out in the wowi channel and watching people develop for love of developing. |
| 22:50.26 | seebs | addon devs, by and large, care more about quality and design than the bulk of the professional programmers I meet. |
| 22:51.14 | Dramiel1 | :) |
| 22:52.02 | seebs | I still remember the utter shock I experienced when I started working with programmers and discovered that some of them didn't actually enjoy programming. |
| 22:52.08 | seebs | I was all "WTF?!?!?" |
| 22:52.13 | Dramiel1 | I try to do the same at work, but sometimes (like when people force you to write IOS app) you just want to finnish it fast and move on :p |
| 22:52.39 | seebs | heh |
| 22:52.51 | seebs | I actually sort of enjoy iOS. I've certainly dealt with stuff I liked less. |
| 22:53.11 | seebs | I have a fractal app that I keep planning to get back to working on a rewrite of. |
| 22:53.21 | Dramiel1 | well I normally write for .net, mostly asp.net mvc so I'm little bit spoiled :p |
| 22:53.40 | Dramiel1 | but on IOS the most frustrating thing is xcode 4 |
| 22:53.41 | seebs | Heretic. :P |
| 22:53.51 | Dramiel1 | that thing is so, so, so, so bad |
| 22:54.15 | seebs | I am only marginally aware of it. I'm one of those old-school makefile/command-line people mostly. |
| 22:54.31 | seebs | I don't like it much, but I don't mind it more than other IDEs I've used. |
| 22:54.39 | Dramiel1 | oh :) |
| 22:55.37 | Dramiel1 | I hate it because it just does random things |
| 22:55.52 | Dramiel1 | sometimes debugger gets stuck so you have to restart COMPUTER to fix it :p |
| 22:55.57 | seebs | I thought that was the core function of IDEs. :P |
| 22:56.19 | seebs | I have had very few positive experiences involving software that claimed to make things easier to do. |
| 22:56.32 | Dramiel1 | or the most random thing I encountered is that if you update project from repo, it hides all the build configurations and you have to restart xcode :p |
| 22:56.42 | seebs | Charming. |
| 22:56.57 | seebs | BTW, I will say this much: I have actually gotten fixes by filing bug reports occasionally. |
| 22:57.14 | seebs | I suppose that ought to be a taken-for-granted thing, but in my world it usually hasn't been. |
| 22:57.23 | Dramiel1 | I'm not microsoft fan, but Visual Studio is something that I would describe as "pure awesomness" :p |
| 22:57.39 | seebs | I haven't touched any of their dev stuff in probably ten years. |
| 22:57.59 | seebs | Look at it this way -- last time I had to deal with it, we were having issues that turned out to have to do with the near/far conventions being used by a particular DLL. |
| 22:58.21 | Dramiel1 | hm :p |
| 22:58.32 | seebs | I'm mostly off in Linux land these days, where I do toolchain and build system stuff, which I am quite happy with. |
| 22:58.43 | seebs | I really *like* compilers and language specs. |
| 22:59.31 | Dramiel1 | hm I'm probably the opposite - I love high level programming languages such as c# |
| 23:00.07 | Dramiel1 | I did some low level stuff, but only little stuff where I needed to start up .net inside wow and eve online :p |
| 23:00.56 | seebs | I have mixed feelings about various HLLs. I quite like Ruby and Lua. I dislike Python and absolutely loathe PHP. |
| 23:01.24 | seebs | I have mostly ignored .net on the grounds that MS has this habit of introducing the Next Big Thing and then replacing it a while later. |
| 23:01.52 | seebs | So I figure by the time I care they'll be onto whatever replaced it. |
| 23:03.05 | Dramiel1 | well I'd suggest that you check out .net sometimes because ms is putting it everywhere |
| 23:03.41 | Dramiel1 | win 8 will have .net on phones and tablets too I believe |
| 23:03.45 | seebs | I'm still doing most of my coding using an API I learned in the late 80s. I am resistent to learning a new API without a real expectation that it'll be widely available in ten years. :P |
| 23:04.00 | Dramiel1 | oh |
| 23:04.05 | Dramiel1 | I might start calling you grandpa |
| 23:04.07 | Dramiel1 | :p |
| 23:04.13 | seebs | I also dislike single-source APIs and tools. I want real competition in the market... |
| 23:04.32 | Dramiel1 | what do you mean? |
| 23:04.35 | seebs | Yeah, I'm old. Not THAT old, mind. When I started using Unix systems I had to sit on a phone book. |
| 23:04.58 | seebs | Well, let's say I write in C. I can use Intel's compiler, Sun/Oracle's, IBM's, GNU's... I have a selection of choices there. |
| 23:05.06 | seebs | If one of the vendors pisses me off, I can just hop to another vendor. |
| 23:05.48 | seebs | Same for Unix-like in general. OS X, Linux, couple commercial choices. |
| 23:06.09 | seebs | One of the reasons my iOS app is a toy that I'm not hugely invested in is that if Apple annoys me, the app is dead. |
| 23:06.21 | seebs | ... I'd do Android, except I dislike programming for it too much. :) |
| 23:06.40 | Dramiel1 | oh I would actually exchange obj-c for java any time |
| 23:06.56 | Dramiel1 | I mean I dont like java.. but I dislike obj-c more :p |
| 23:06.59 | seebs | Heh. |
| 23:07.05 | Hyperi | What is wrong with you people :< |
| 23:07.18 | seebs | I don't much like java. I don't mind obj-c. |
| 23:07.26 | seebs | I've had to do some stuff in it a few times and it's not yet bitten me badly. |
| 23:07.46 | seebs | It's got a number of advantages, one of which is that it isn't C++. |
| 23:08.09 | seebs | I'm pretty biased, though. I like plain old C a lot. Like, enough that I spent about 10 years on the standards committee for it as a hobby. |
| 23:08.44 | seebs | It's certainly got design warts aplenty, but I basically like the language. |
| 23:09.18 | Hyperi | ANSI C > original C |
| 23:09.20 | Dramiel1 | now with ARC obj-c is not that big pain in the ass, but still it isnt something I like.. dunno I'm probably just really spoiled by Visual Studio and c# |
| 23:09.23 | Hyperi | Naturally _imo_ |
| 23:09.35 | seebs | I like C99, and wish more people had adopted more of it. |
| 23:09.43 | seebs | Has some really NICE idioms for solving common problems. |
| 23:10.01 | seebs | And actually, I do just use those now, since they're in gcc and the ones I care about are moderately widely available. |
| 23:10.20 | seebs | compound literals + designated initializers = love |
| 23:10.46 | seebs | select(..., &(struct tm *) { .tm_usec = 500000 }); |
| 23:10.52 | Hyperi | Dramiel1: C# is ...painful |
| 23:10.55 | seebs | Er, just (struct tm) |
| 23:11.17 | Dramiel1 | Hyperi: why would you say that? :p |
| 23:12.08 | seebs | http://forums.riftgame.com/public-test-shard/276864-1-6-progress-set-1-11-8-11-a.html#post3367348 <-- ooh patch notes |
| 23:12.28 | Hyperi | Dramiel1: |
| 23:12.48 | Hyperi | Dramiel1: Well I never had to do it professionally (mostly as I'm more of a UNIX geek :>) |
| 23:13.06 | Hyperi | Dramiel1: But my old guildm8 was still in school and had to do school learning projects in Java or C# and he rather chose C# |
| 23:13.19 | Hyperi | I ended up helping him with some of the tasks |
| 23:13.29 | Hyperi | And hence had to read C# documentation and google help |
| 23:13.41 | Hyperi | And noticed several really weird quirks that sounded really silly to me |
| 23:14.46 | seebs | ooooh |
| 23:14.49 | seebs | * Rift Guard: Absorption shield is now triggered whenever you use Guarded Steel or a Plane Shift ability. |
| 23:15.06 | seebs | That takes one finisher out of the OMG I NEED TO USE SIXTEEN FINISHERS cycle. |
| 23:17.11 | seebs | * You can now mine Earth Elemental, Oreling, and Golem corpses. |
| 23:17.11 | seebs | * You can now forage Treant and Bogling corpses. |
| 23:17.13 | seebs | yayyy |
| 23:19.07 | Hyperi | * The second tier of Planar Attunement is now unlocked! The window now allows you to preview abilities in tiers you cannot access yet. |
| 23:19.10 | Hyperi | ! |
| 23:19.38 | seebs | Yup. |
| 23:19.46 | *** join/#riftuidev Kraps (~chatzilla@69.114.203.17) |
| 23:19.46 | *** join/#riftuidev Marru_Home (~IceChat77@c-75-70-203-138.hsd1.co.comcast.net) |
| 23:19.51 | seebs | There's some nice stuff, like a temporary 120sp buff. Which does not stack with consumables. |
| 23:20.07 | Marru_Home | sup guys |
| 23:20.08 | seebs | So there's some cool stuff to be had, but it's not mandatory for raiding. |
| 23:20.28 | Marru_Home | raiding? huh? |
| 23:20.43 | Hyperi | Your cleaning closet |
| 23:20.46 | Hyperi | For some hoovering |
| 23:20.51 | Hyperi | ;P |
| 23:20.58 | Marru_Home | lol |
| 23:21.01 | Marru_Home | sooooo |
| 23:21.04 | Marru_Home | i realized something today |
| 23:21.27 | Marru_Home | i like writing interfaces, like menus n stuff |
| 23:21.51 | Marru_Home | soooo, if anyone wants me to write one for em in exchange for bug fixing my addons lemme know :) |
| 23:22.05 | Dramiel1 | seriously? interface is the boring part :p |
| 23:22.12 | Marru_Home | lol |
| 23:22.14 | Marru_Home | i like it |
| 23:23.31 | Dramiel1 | it's most of the time easy and repetitive whereas the internal logic is usually completely different every time |
| 23:23.41 | Marru_Home | i know |
| 23:23.51 | Marru_Home | i just like making cool looking interfaces |
| 23:24.20 | Marru_Home | so if u wanna bug fix riftcount and/or combat text for me i'll work on ur menus :) |
| 23:25.12 | Dramiel1 | I don't have time to finish my own addons, so not really :p |
| 23:25.41 | Marru_Home | lol |
| 23:25.55 | Marru_Home | it'd save you time on the menu part :) |
| 23:27.08 | Dramiel1 | I'm very specific on what I want from my addons and also I use quite different code style.. so far I haven't seen originally looking .. or even non-ugly addon for rift |
| 23:29.58 | Hyperi | I've the same, I love backend coding - hate frontend :< |
| 23:30.04 | Hyperi | In any language tbh, not limited to addons :P |
| 23:30.31 | Marru_Home | backend coding? like what? |
| 23:30.56 | Hyperi | In what? addon ? |
| 23:31.02 | Marru_Home | anything |
| 23:31.04 | Hyperi | libs and framework? |
| 23:31.07 | Hyperi | API systems? |
| 23:31.10 | Marru_Home | ahh ok |
| 23:31.18 | Marru_Home | sry im a comp sci major |
| 23:31.26 | Marru_Home | so if i hear a term idk |
| 23:31.35 | Marru_Home | i get curious |
| 23:31.40 | Hyperi | It's not even new term really :/ |
| 23:31.46 | Marru_Home | i know |
| 23:31.48 | Hyperi | You jsut haven't coded much ? |
| 23:32.00 | Marru_Home | just thought you meant like backdoor or somethin |
| 23:32.03 | Hyperi | Or by some luck just haven't ran into it :P |
| 23:32.10 | Marru_Home | just haven't run into it |
| 23:32.14 | Hyperi | Nah, I refer that as unprotected anal sex |
| 23:32.24 | Dramiel1 | for example this is how I write frame that shows aura (icon, progress bar, text and remaining time): http://pastebin.com/KbpDcJcn |
| 23:32.28 | Marru_Home | i know libs and framework n stuff |
| 23:32.28 | Hyperi | Gives a bad reputation to the industry, those who code backdoors |
| 23:32.31 | Dramiel1 | funzies |
| 23:33.33 | Marru_Home | import "core" kind of stuff, thats how you'd get multiple files without rifts toc thing? |
| 23:33.53 | Dramiel1 | no that is not possible |
| 23:33.58 | Marru_Home | oh |
| 23:34.05 | Marru_Home | whats import "core"; do then |
| 23:34.15 | Dramiel1 | all files must be in toc, but this imports functions from something that has line "package "core";" in it |
| 23:34.26 | Marru_Home | oh ok |
| 23:34.40 | Dramiel1 | it is totally unnecessary, but it was fun to write such system :P |
| 23:34.42 | Marru_Home | but isn't that already done with the toc stuff? |
| 23:34.55 | Marru_Home | ah ok |
| 23:37.07 | Dramiel1 | this basically allows me to put functions in "global" namespace without putting thing into global namespace :p |
| 23:37.44 | Dramiel1 | so for example I can extend table.xxx without affecting other addons |
| 23:37.57 | Marru_Home | not really sure what you mean exactly |
| 23:37.58 | Marru_Home | but ok |
| 23:38.30 | Dramiel1 | :) by the way was it you who was writing the clone table function the other day? |
| 23:38.43 | Marru_Home | was a while ago but yeah |
| 23:38.50 | Marru_Home | didn't end up writing one |
| 23:38.59 | Marru_Home | ended up using a example from a site |
| 23:39.31 | Dramiel1 | http://pastebin.com/z8mxrgrY this is the simplest and yet very powerful implementation I could think of |
| 23:39.49 | Dramiel1 | it ignores metatables tho, as I didn't need that functionality yet |
| 23:39.58 | Marru_Home | i dont either |
| 23:40.05 | Marru_Home | hey |
| 23:40.11 | Marru_Home | --[[ |
| 23:40.16 | Marru_Home | is that how u do block comments? |
| 23:40.18 | Dramiel1 | yep |
| 23:40.24 | Marru_Home | good to know lol |
| 23:40.37 | Marru_Home | i've just been throwin -- in front of multiple lines |
| 23:41.13 | Marru_Home | ok so the way that one is written |
| 23:41.21 | Marru_Home | clone = function(source) |
| 23:41.33 | Marru_Home | wats different between that and function source() |
| 23:41.52 | Dramiel1 | almost nothing, I just prefer this syntaxe |
| 23:42.22 | Marru_Home | can u do it with .Event:LeftClick() kind of stuff? |
| 23:42.44 | Dramiel1 | yes, but you need to know what ":" represents |
| 23:43.01 | Marru_Home | ? |
| 23:43.24 | Dramiel1 | .Event.LeftClick = function(self) vs function .Event:LeftClick() |
| 23:43.36 | Marru_Home | ahhh ok |
| 23:44.39 | Marru_Home | if i wanted to make a function that could either take 2 inputs or 3 inputs, how would i do that? |
| 23:45.50 | Dramiel1 | same, only you would put those parameters after the "self" parameter in the first case |
| 23:46.17 | Marru_Home | no i mean |
| 23:46.25 | Marru_Home | uhh |
| 23:46.42 | Marru_Home | okay, say im making a function to be a frame manufactory |
| 23:47.11 | Marru_Home | i want to call makeFrame(text, blah, blah) or makeFrame(text, blah, blah, 20) |
| 23:47.19 | Marru_Home | and one would make one frame, the other would make 20 |
| 23:47.24 | seebs | You just declare all the arguments. |
| 23:47.30 | seebs | If they aren't provided, they default to nil. |
| 23:47.35 | Marru_Home | ok |
| 23:47.39 | Marru_Home | so i'd just add nil checks? |
| 23:47.41 | seebs | So you'd declare function makeFrame(text, blah, blah, count) |
| 23:47.52 | seebs | and then if count == nil then count = 1 end |
| 23:47.58 | Marru_Home | kk |
| 23:48.05 | Marru_Home | thx |
| 23:49.54 | Marru_Home | ok this ones gonna be a bit weird |
| 23:50.22 | Marru_Home | i want to use 2 for loops, one nested in the other, to create frames |
| 23:50.34 | Torhal | Dramiel1: You use C-like syntax. |
| 23:50.42 | Torhal | I had to break myself of that when I learned Lua. |
| 23:50.54 | Marru_Home | here, this'll be easier if i explain the context |
| 23:51.21 | Marru_Home | i have 4 things, i want user control on the color for each one, and i want that control to be a slider for each r g b and a value |
| 23:52.00 | Dramiel1 | Torhal: that is quite possible :) |
| 23:52.17 | Torhal | Parens where they aren't required (in Lua) and semicolons |
| 23:52.18 | Marru_Home | so i want to go through each thing and make 4 frames and place them in a line next to it |
| 23:53.05 | Dramiel1 | oh yea.. I'm used to read code with parens and semicolons everywhere and I feel like the code is easier to read |
| 23:55.34 | Marru_Home | nvm got it |
| 23:58.30 | Marru_Home | zorba i need a new frame type, inverted mask :) |