**** BEGIN LOGGING AT Wed Jan 21 02:59:58 2015 Jan 21 03:00:08 hesperaux: it's a leaning tower of pisa, on top of a leaning tower of pizza. at times, it's easier to strap stabilizing rockets to the top tower to keep it upright, instead of laying a real foundation. Jan 21 03:00:24 lol Jan 21 03:02:31 hesperaux: if you ever find a cure, let me know. at the moment my bias is usually to build higher and higher level abstractions. not always a good idea, but it's where I am now :| Jan 21 03:02:49 it hasn't bitten me harder than other approaches yet, anyway Jan 21 03:03:32 my problem is that I plan out a nice foundation, but when I go to implement it, I realize there are small details/conditions which I've overlooked - so conditional code gets added here and there, in places I don't believe it should be (sometimes even in the UI) Jan 21 03:03:39 AbstractPisaPizzaFactoryImplementation Jan 21 03:04:09 ^ yeah, java does not make it easy Jan 21 03:06:18 basically, my goal is to have the UI thread be completely design flow oriented Jan 21 03:06:33 im looking into getting an android box tv can anyone recommend a decent one? Jan 21 03:06:34 everything else happens either in classes or on another thread Jan 21 03:07:13 hesperaux: God, yeah. I write pages and pages of code and my app barely performs a few functions. Jan 21 03:07:24 jc23: probably http://www.google.com/nexus/player/ ? at the very least it'll get faster updates than most others, because it's stock. Jan 21 03:07:32 I assume sony / others have bastardized it somehow Jan 21 03:07:40 And it takes me days to implement a new (big) feature. Jan 21 03:07:46 it's good to know i'm not totally off Jan 21 03:08:01 or at the very least, we're all mad here :) Jan 21 03:08:11 I'M NOT MAD! Jan 21 03:08:13 CHANGE PLACES! Jan 21 03:08:19 * JesusFreke cackles maniacally Jan 21 03:08:19 here's another question - do you guys freak out when you have to update your database schema on an update? heh. I fear that day Jan 21 03:08:21 I think you have to be mad to be a programmer. Jan 21 03:08:40 And if you're not then you'll soon become so. :D Jan 21 03:08:42 hesperaux: nope. done it too many times to be afeared of it now. Jan 21 03:08:54 every blown up your customers' databases? Jan 21 03:09:00 only temporarily! Jan 21 03:09:05 groxx: i have done everything listed in that remote-debugging link Jan 21 03:09:09 the nexus can be rooted aswell? Jan 21 03:09:10 so then I take it you back up their old one? Jan 21 03:09:33 For future reference, any suggestions/tutorials/etc on best practices for doing database upgrades would be appreciated Jan 21 03:09:33 hesperaux: haven't had to restore a backup, just brought the site down a few times because the changes took longer than we expected. Jan 21 03:10:12 I've been following the development of Android Studio recently. There seems to be a lot happening (especially tonight -- many merges are going on). I wonder what cool features it'll get in future. Jan 21 03:10:14 jc23: I'm pretty sure it has fastboot support Jan 21 03:10:14 then I ended up writing a load-aware helper. that saved us a lot of pain when it was time to migrate a 2.1 billion row table :| (it took months to complete) Jan 21 03:10:21 JVM unit testing is a great one that should be here before long. Jan 21 03:10:29 I can't imagine what else they have planned. Jan 21 03:10:51 It feels kinda cutting-edge. Jan 21 03:11:11 Here are tonight's, BTW: https://android-review.googlesource.com/#/q/branch:studio-1.1-dev Jan 21 03:11:21 ok thanks...just getting into this Jan 21 03:13:23 groxx: knapper_tech JesusFreke switching from chrome canary to just chrome did the job Jan 21 03:13:42 hesperaux: suggestions / tutorials... hm. a) grok sqlite.org, it's great stuff. b) use auto-increment keys unless you have a performance reason not to. c) I avoid database-level foreign keys, and bake them into an application layer. easier to migrate data later. database-level unique-constraints are very useful though. Jan 21 03:15:46 hesperaux: also, do you know about "insert into x select y" ? sometimes the easiest way to migrate data is to make the new table and copy stuff into it. Jan 21 03:18:23 Do any of you use the Gimp to make your own icons? Jan 21 03:18:24 one thing about sqlite - wish it had better 'alter table' Jan 21 03:18:41 MikeWallaceDev inkscape ftw Jan 21 03:18:47 I just forked this : https://github.com/MikeWallaceDev/script-fu-save-android-icons Jan 21 03:18:57 well, fork you too! Jan 21 03:19:02 :D Jan 21 03:19:44 I need some Java help Jan 21 03:19:46 http://pastebin.com/yGKNaBgC Jan 21 03:20:02 i am getting error: incompatible types: Object cannot be converted to String Jan 21 03:20:19 in maplist.get (key) Jan 21 03:20:47 so i'm not entirely sure what to do here Jan 21 03:21:08 Agamemnus simple, learn java ! :D Jan 21 03:21:15 Agamemnus: http://www.json.org/javadoc/org/json/JSONObject.html#get(java.lang.String) Jan 21 03:21:15 ,_, Jan 21 03:21:21 get() returns an Object, not a String Jan 21 03:21:42 If you know it will be a string, use getString() instead Jan 21 03:21:46 your iterator is the wrong type Jan 21 03:22:06 Just kinda trying to use Stack Overflow Jan 21 03:22:09 oh no wait, sorry, maybe i should learn java Jan 21 03:22:32 If I say "getActionBar().setSelectedNavigationItem(0);" and 0 is already selected, the event handler isn't fired. If 0 is not selected, the event handler is fired. Is there a simple way to make the event handler be fired in both cases? Jan 21 03:22:46 I'm guessing not, but I thought I'd ask. :D Jan 21 03:22:49 Ok, so that document is confusing Jan 21 03:23:01 not particularly Jan 21 03:23:07 pretty standard javadoc Jan 21 03:23:12 'produces the string {"JSON": "Hello, World"}.' Jan 21 03:23:15 shouldn't that be: Jan 21 03:23:24 produces the string '{"JSON": "Hello, World"}' Jan 21 03:23:48 Strings aren't denoted by a ' Jan 21 03:23:49 anyway. Jan 21 03:23:53 They used a different typeface for the string in question. Jan 21 03:23:59 Using getString helped Jan 21 03:24:15 Although maybe to be consistent they should've placed that second string on its own line. Jan 21 03:24:42 But it's not code, it's too confusing. It should be quoted for clarity Jan 21 03:24:49 The code has the same font Jan 21 03:24:57 It's not the same font. Jan 21 03:24:59 "{\"JSON\": \"Hello, World\"}" Jan 21 03:25:02 is not more clear, imo :) Jan 21 03:25:04 that's kind of dumb though, it's a typed list, why doesn't get return the correct type? Jan 21 03:25:16 the code right before the statement has the same font Jan 21 03:25:51 and the same style Jan 21 03:25:52 ANYHOO Jan 21 03:25:57 thanks for the help Jan 21 03:26:03 I think it's Times New Roman and Courier New there. Jan 21 03:26:16 the text is times, yes Jan 21 03:26:23 Oh, the code has the same font. I misread. Jan 21 03:27:32 Ok, one last question Jan 21 03:27:51 oh wait Jan 21 03:28:34 g00s, agreed on the alter table issue... Jan 21 03:28:34 g00s: yeah, more alter support would be nice. you can get a long way with fairly generic external scripts though. Jan 21 03:29:16 g00s: dunno what options there are with sqlite, but mysql and others have some pretty fancy ones in the percona toolkit. Jan 21 03:29:39 I saw a post on SO I think about modifying the schema in read write mode Jan 21 03:29:59 it is dangerous but it is very fast to rename tables Jan 21 03:30:06 no part of my app freaks me the hell out more than the scliteopenhelper onUpdateDatabase Jan 21 03:31:07 ohhh Jan 21 03:31:37 exactly Jan 21 03:31:51 brb Jan 21 03:38:06 hesperaux_: ha, I don't think I'd try that :) though with sqlite at least the formats are well defined. Jan 21 03:39:15 padding around my custom actionbar view... Jan 21 03:39:25 above and below Jan 21 03:39:29 about 6dp Jan 21 03:40:27 customize actionbar style? Jan 21 03:40:30 Okay Jan 21 03:40:38 So I'm still stuck a bit Jan 21 03:41:24 json_object.put without a throws Jan 21 03:41:45 Without " throws JSONException {"... but I can't add it to the function because it's an override function Jan 21 03:41:54 How do I handle this? Jan 21 03:42:27 i guess i could just call another function Jan 21 03:42:41 How do Android apps that get most of their data from the cloud go about showing things...do they save everything to a local database first and then use a loader, or do they skip the database altogether? Jan 21 03:43:09 Agamemnus: What are you overriding? Jan 21 03:43:21 something called "onPollfishSurveyReceived". Jan 21 03:43:52 oddalot: I don't know about anyone else but it is incredibly use-dependent. I save my users' data to my server when they log out. Jan 21 03:44:14 You'll need something to merge two save files Jan 21 03:44:31 you could also always just use the server only Jan 21 03:44:49 Well, that's the thing, I'm just using this app as a 'controller' of sorts, the server saves all the state and has a database etc... Jan 21 03:45:11 so there's no need for me to save data locally....but if it makes displaying it easier, i might do that. Jan 21 03:45:27 just think about it as a web app... Jan 21 03:45:49 exactly, and that's what i'm used to, a web app, javascript usually only saves minimal info in cookies Jan 21 03:46:40 hmmm....thanks Jan 21 03:47:03 Agamemnus: tl;dr: you can't, because Java won't let you. wrap it, handle the exception, and call the wrapped method. there's basically no other option (they're all some form of "don't call/override that function directly") Jan 21 03:47:59 Agamemnus: you could wrap the exception in a (subclass of) RuntimeException and throw that Jan 21 03:48:40 JesusFreke: how would that look? I'm confused by the terminology Jan 21 03:48:52 hm, yeah, there's a bit of ambiguity there. Jan 21 03:49:01 throw new RuntimeException(ex); Jan 21 03:49:17 Agamemnus: are you trying to override something that throws jsonexception, or are you overriding something else and you need to call jsonObject.get? Jan 21 03:49:18 so I should have try and throw? Jan 21 03:49:20 or .put Jan 21 03:49:28 RuntimeException and its subclasses don't require a throws declaration on the method Jan 21 03:49:45 guys, like this: http://pastebin.com/FCDPUVMZ Jan 21 03:49:58 but this doesn't work, gives me a compile error Jan 21 03:51:16 i mean try/catch Jan 21 03:51:26 Agamemnus: yeah. wrap it in try { .put code } catch (JSONException e) { throw new RuntimeException(e); }. Jan 21 03:52:35 all right, thanks Jan 21 03:52:49 that'll mean that, if it _does_ throw, it'll probably crash your app. but you're probably fine there. Jan 21 03:52:52 technically I don't need the throw, do I? Jan 21 03:53:18 should I have a throw? Jan 21 03:53:24 not really. you probably do want to at least call `e.printStackTrace()` so you see it in the logs if it happens while you're coding, but for that code it may never actually happen. Jan 21 03:53:28 It compiled without a throw Jan 21 03:53:50 I can't see how it could. Jan 21 03:54:40 the built-in json library is a little.... pedantic. it's probably correct that it could throw, but it is a bit extreme. Jan 21 03:54:43 crashing is a perfectly reasonable thing for an app to do Jan 21 03:54:54 if something unexpected happens, crash Jan 21 03:55:04 don't let your app continue on in an unknown, non-deterministic state Jan 21 03:55:11 it's just going to crash or misbehave later on anyway Jan 21 03:55:32 there's expected failures (IOException) and unexpected failures Jan 21 03:55:44 if you know the data can be encoded into JSON, rethrow the exception Jan 21 03:56:06 potentially with more data (such as the value you were trying to set) Jan 21 03:56:13 so to be "safe", would I need... Jan 21 03:56:22 e.printStackTrace(); throw new RuntimeException(e); Jan 21 03:56:27 or just the throw? Jan 21 03:56:30 just throw Jan 21 03:56:35 you either log or throw, never both, never neither Jan 21 03:56:35 will the throw print a stack trace? Jan 21 03:56:38 yes Jan 21 03:56:39 okay Jan 21 03:56:45 almost never neither* Jan 21 03:57:16 So I added the throw. It's like whatever, but it's cool. Jan 21 03:57:31 Better safe than the-nuclear-reactor-is-hijacked-by-a-terrorist Jan 21 03:57:36 it'll at least let you know if it happens, which is better than not knowing :) Jan 21 03:59:18 throwing exceptions and crashing the app is only viable when you have a crash reporting service Jan 21 03:59:31 at least in your release builds Jan 21 04:00:11 which you do want, because the built-in "report crash?" dialog almost never gets used Jan 21 04:01:52 indeed Jan 21 04:01:55 ok, great, it all compiles now Jan 21 04:02:36 Hi, i'm new to android dev, i'm sorry if this question is dumb: i'd like to know if an application can change the files stored in the apk package Jan 21 04:03:34 no, the application does not have write access to the actual apk file Jan 21 04:03:34 soulisson: literally in the apk package: no. it's read-only. if you want to modify and use it, you'll need to copy it somewhere else. Jan 21 04:03:49 Thank you very much :) Jan 21 04:04:27 your application will have a writable directory all to itself, that you can store files in Jan 21 04:05:02 JesusFreke: thanks Jan 21 04:05:34 groxx: I was considering relying on the "Report Problem" thing. Do users just not use it? Jan 21 04:06:10 not really. you'll get _some_, but you can't really rely on it for anything but big breaking bugs. Jan 21 04:06:31 Is there something better I should use? Jan 21 04:06:42 if that's enough: go for it. no permissions, it's built in, the web UI for viewing isn't all that bad. it has its perks. Jan 21 04:07:08 crashlytics/fabric is probably the most full-featured at the moment, or acra if you want to host it yourself. Jan 21 04:07:20 not sure if there's anything better out there Jan 21 04:08:01 TacticalJoke: as an example, we have something like 10,000x more crashes in crashlytics than in the play console. Jan 21 04:08:30 if you have a weirdly technical crowd, your odds will be better, but still. Jan 21 04:09:09 Does it show a "report crash" thing? I'm trying to figure it out with Google. Jan 21 04:09:44 yeah. it's the same thing you get with an app-not-responding -> click report. Jan 21 04:10:37 I see. I wonder why you get thousands more reports with that. Jan 21 04:10:57 oh, another downside to the google crash info: you won't get it for non-google-play devices. so all amazon devices, most chinese, most indian, etc. Jan 21 04:11:24 amazon probably has something similar, but I'm not familiar with it Jan 21 04:11:37 Ah, good to know. Jan 21 04:12:48 I _suspect_ it's just that people are used to hitting "ok" in dialogs. and if they don't, they get a black screen with a bunch of UI that they don't want to deal with to report things, and they're warned about private data, yadda yadda yadda. there's a lot standing between "app crashed" and "told developer about it". Jan 21 04:13:03 and if they hit back or home, it's gone forever. Jan 21 04:13:03 anyone know how to cast a Bundle to a PersistableBundle? Jan 21 04:13:29 n^izzo: they're not the same thing. you can probably _create_ a persistablebundle and copy the data to it, though Jan 21 04:13:56 Hmm, okay. Google doesn't seem to be giving me much info about this. Trying to figure out what's so different with regard to crash reporting. Jan 21 04:15:54 n^izzo: if it _might_ be a persistable bundle, you can check with `if (var instanceof PersistableBundle)`, but in general that will not be true. Jan 21 04:16:43 TacticalJoke: "what's so different"? Jan 21 04:28:57 groxx: cant seem to make a persistableBundle from a bundle Jan 21 04:30:47 hey guys Jan 21 04:30:53 www.twitch.com/babuloseo Jan 21 04:31:01 can someone help me find a way to reduce the code? Jan 21 04:31:13 I am fairly new to Java, so I don't know how to reduce the overhead Jan 21 04:32:13 n^izzo: from looking at the API, they're both instances of android.os.BaseBundle. there's enough info there to loop over all the keys, check their types, and copy the data (and recurse on sub-bundles). it's possible. odds are pretty good though that you'd be better off doing something else - why are you trying to use persistable bundles? Jan 21 04:32:49 is there a way Jan 21 04:32:57 to globally declare what I am trying to do on the stream Jan 21 04:33:06 so that I don't have to copy and paste the same thing for each method? Jan 21 04:34:13 groxx I need to use them because I'm using jobsheduler Jan 21 04:35:16 and to use setextras I need to cast my bundle to a persistable bundle Jan 21 04:42:52 Quick question. How do I restore Android Studio's layout. I have some broken tool windows and logcat isn't showing up Jan 21 04:43:13 matt_j: does cmd-6 bring it up? Jan 21 04:43:31 Yeah but the output is empty. Jan 21 04:43:32 or view -> tool windows Jan 21 04:43:50 but I also have a semi-transparent empty tool window that I can't close Jan 21 04:43:52 there's a square + circle-arrow button to restart logcat, that usually fixes it for me Jan 21 04:43:55 D: Jan 21 04:44:03 running on mac btw Jan 21 04:44:17 I need to, like, reset all user settings or something Jan 21 04:44:18 don't suppose you've turned on floating window transparency in prefs? Jan 21 04:44:51 probably ~/library/application support/something, if you want to just delete that folder for giggles. Jan 21 04:45:12 (not ~/library/application support/ in case you were about to do that. that would be bad.) Jan 21 04:45:14 yeah I"m gonna try that Jan 21 04:46:06 huh. there's not much in there. Jan 21 04:47:14 trying now Jan 21 04:47:18 hopefully my projects don't break Jan 21 04:47:56 ah crap I think it's installing the SDK. Oh well Jan 21 04:48:11 dunno where it stores its preferences :\ I'd _hope_ not in the app package... but it kinda looks like it might, since it's not in the normal ~/library folders Jan 21 04:48:19 yeah it is Jan 21 04:48:26 in ~library Jan 21 04:49:13 is it bad practice to put an Object into a listView tag? How to store 3-4 metadata items in each listView otherwise? Jan 21 04:50:53 matt_j: I see only "availables.xml", "extensions.xml", and some jar thing (closed the folder, names may be incorrect, meh). not enough to store all my changes to the settings. Jan 21 04:50:58 hesperaux: Can't you use a List or something? Jan 21 04:51:44 put a List in the tag? that would technically classify as an object. A hash map would be preferable Jan 21 04:52:25 No, I mean something separate. Jan 21 04:52:25 TacticalJoke: how is your reddit app going ? Jan 21 04:52:29 groxx: http://stackoverflow.com/questions/19384033/how-to-reset-android-studio Jan 21 04:52:33 there's actually a few directories Jan 21 04:52:48 shmoooz: It's going well, thanks. I'm finally used to Android Studio after a couple of days, so I'm back writing code. :) Jan 21 04:52:52 matt_j: https://intellij-support.jetbrains.com/entries/23358108-Directories-used-by-the-IDE-to-store-settings-caches-plugins-and-logs is what i found Jan 21 04:52:56 TacticalJoke, ok, so populate the listview parallel to the List, keeping the indices and Ids related? Jan 21 04:52:58 shmoooz: How 'bout yours? Jan 21 04:53:02 didn't think of looking in preferences, but that makes sense Jan 21 04:53:11 hesperaux: What is that you're storing? Jan 21 04:53:26 TacticalJoke: I haven't done much to it since you saw it, I started playing too much Tribes again ;) Jan 21 04:53:48 haha Jan 21 04:53:54 I went through years of that. Jan 21 04:54:21 shmoooz did you get an extra o in your nick ? Jan 21 04:54:35 TacticalJoke, each list item will have two strings taken from the UI, and possibly a third string (haven't decided yet) Jan 21 04:54:37 maybe i never noticed it Jan 21 04:55:13 hesperaux: What are these strings? Jan 21 04:55:16 g00s: yeah my alternate nicks for when it reconnects add another o for the second and 2 o's for the third alternate nick Jan 21 04:55:29 heh Jan 21 04:55:35 form data entered by the user, TacticalJoke Jan 21 04:55:37 One thing I need to do is move to material and Toolbar (and maybe RecyclerView). Jan 21 04:55:43 My project's code is so volatile. lol Jan 21 04:56:18 TacticalJoke i wouldn't bother rith recyclerview unless whatever you want to do is difficult with LV Jan 21 04:56:31 well - i wouldn't make it a priority anyhow Jan 21 04:56:33 hesperaux: Hmm. Would it make sense for this data to be part of the object that already backs the list? Jan 21 04:56:48 Yeah, I'm just gonna learn about it at first. Not a big priority. Jan 21 04:57:12 "1 star! - uses listview instead of recyclerview" Jan 21 04:57:14 TacticalJoke, sorry, I don't understand what you mean Jan 21 04:57:16 haha Jan 21 04:57:22 haha Jan 21 04:57:24 hesperaux: Your ListView is backed by some kind of list, right? Jan 21 04:57:35 I'll have to leave that review when you release it Jan 21 04:57:37 Would it make sense for those list items to hold the data in question? Jan 21 04:57:41 lol Jan 21 04:58:09 yes, of course Jan 21 04:58:15 (my personal favorites though are [1 star] "best app ever! \n ") Jan 21 04:58:26 the part I was confused about was getting it back out later. From the view object passed to onItemClick Jan 21 04:58:29 I've seen so many Amazon reviews like that. Frustrating even as a reader. Jan 21 04:58:41 Or the hilarious "Terrible. Do not buy" 5/5. Jan 21 04:58:48 lol i've seen that Jan 21 04:59:10 most of the dumb reviews i've seen on amazon are "5 stars, received on time" Jan 21 04:59:21 I hate the 3/5s that say "Just leaving this here as a placeholder until I read the book". Jan 21 04:59:22 I assume you've all read http://www.amazon.com/Tuscan-Whole-Milk-Gallon-128/dp/B00032G1S0 ? Jan 21 04:59:34 Gee, thanks for destroying the rating system for no purpose whatsoever. Jan 21 04:59:51 g00s: oh god yes. product/shipper reviews in the wrong place is so horrible. Jan 21 04:59:52 Yeah, they rate the delivery or whatever. Pointless. Jan 21 04:59:54 TacticalJoke, why would people do that? Jan 21 05:00:06 worst yet, it's like 50% of the reviews Jan 21 05:00:07 I'd not see that, groxx. Jan 21 05:00:29 just name your app 5-Star Jan 21 05:00:31 TacticalJoke: then you're in for a treat. note e.g. the top review: 21,665 of 22,007 people found the following review helpful Jan 21 05:00:32 hesperaux: Humanity always has another surprise up its sleeves. Jan 21 05:00:38 usually a stupid one Jan 21 05:00:40 and people will download it like zombies Jan 21 05:00:42 * hesperaux lols Jan 21 05:00:54 shmooz, that's a good idea Jan 21 05:00:55 shmoooz: that's _brilliant_ Jan 21 05:01:18 It's funny when authors are caught out giving their own books reviews. Jan 21 05:01:21 lol the poem one? Jan 21 05:01:26 Usually they're inept in some way. Jan 21 05:01:41 shmoooz: crap. you've got one competitor: https://play.google.com/store/apps/details?id=com.fivestars.FiveStarsConsumer (next to nobody uses them though) Jan 21 05:02:19 hmmm, maybe they need reverse subliminal mind control, name it 1-Star Jan 21 05:02:32 "Do Not Download this App" Jan 21 05:02:37 haha Jan 21 05:02:41 * hesperaux likes this conversation Jan 21 05:03:17 TacticalJoke, having not actually fleshed out the listview yet, I wasn't really seeing the big picture - keeping a List of thingies and making a custom adapter seems like a very easy way to do this, and just use the index of the item to retrieve the values from the list Jan 21 05:03:29 Right. Jan 21 05:03:35 * hesperaux admits that he forgot the listview would have received the list from the context Jan 21 05:04:15 i'm curious - is there anything from the framework that you don't have to extend to do stuff? Jan 21 05:04:33 hesperaux: startActivity? Jan 21 05:04:37 heh Jan 21 05:04:42 can't think of much else Jan 21 05:04:50 simple widgets I guess Jan 21 05:04:54 The adapter thing is all about lists being really flexible, being able to display all kinds of views for all kinds of data. Jan 21 05:05:13 TacticalJoke, i don't discount that - it's just the number of class files in my project has grown faster than I'd like Jan 21 05:05:31 java :| Jan 21 05:05:35 lol Jan 21 05:05:36 Yeah, subpackages are handy for that. Jan 21 05:05:44 But I hate that too. Jan 21 05:05:50 i never thought to use subpackages Jan 21 05:05:55 that would make sense though Jan 21 05:05:56 I have like six adapters already. Jan 21 05:06:06 And my app hardly does anything. Jan 21 05:06:23 i'm never sure if I should consolidate adapters into less with logic or keep them separate. so far, separate seems easier/cleaner Jan 21 05:06:29 you need a reusable adapter for all Jan 21 05:06:31 TacticalJoke, yeah i've got 3 so far for one activity Jan 21 05:06:33 Yeah, separateness is often the lesser evil. Jan 21 05:06:48 I need a meta-adapter. It adapters all the adapters. Jan 21 05:07:05 the mother of all adapters Jan 21 05:07:10 my theory is that throwing in extra logic into a single adapter will cause all of the adapter instances to slow down slightly. instead of just using the right adapter and having no extra junk in it Jan 21 05:07:24 * hesperaux comes from an embedded C background where one cycle matters Jan 21 05:07:44 i don't have a feel yet for how much CPU is a lot on Android Jan 21 05:07:47 I think that trying to make a single, flexible adapter could entail a maintenance nightmare. Jan 21 05:07:58 arguably any cpu is too much, since it's directly tied to battery life Jan 21 05:08:03 so use as little as you can manage Jan 21 05:08:10 dragorn, that's how I see it too Jan 21 05:08:31 that said you can over-optimize and get such minimal returns it's not worth the time investment Jan 21 05:08:31 Though it's hard to tell what's expensive and what's not, so don't go crazy with optimization. You can profile stuff later. Jan 21 05:08:33 i'm trying not to preclude my efforts for fear of CPU intense code Jan 21 05:08:50 yeah. do what oyu want to get done, first Jan 21 05:08:52 The biggest factor in general, IMO, is code changeability. Jan 21 05:08:56 then figure out what needs to get improved Jan 21 05:08:57 Also readability. Jan 21 05:09:06 yeah Jan 21 05:09:20 TacticalJoke: you sound like your therapist Jan 21 05:09:24 lol Jan 21 05:09:42 I was actually thinking that just then. I'm basically telling *myself* this stuff. Jan 21 05:09:49 and if your app only gets run once or twice a day and you only expect a small amount of time being spent in it, then who cares how much cpu Jan 21 05:10:01 so I have Jan 21 05:10:03 "Pending authentication: please accept debugging session on the device." Jan 21 05:10:03 dragorn, true Jan 21 05:10:09 so far this is the only reusable adapter we've used: https://gist.github.com/Groxx/c461634e65647f399e84 . makes multiple view types based on different types of data less painful (and we have that aaaaalll over the place) Jan 21 05:10:11 But it's authorized Jan 21 05:10:19 I just reset the authorization too Jan 21 05:10:19 hesperaux: Are you on Android Studio? That has some nice performance-profiling stuff. Jan 21 05:10:24 It has a really simple memory manager, for example. Jan 21 05:10:26 I don't know what's going on. Any ideas? Jan 21 05:10:31 I am. I've never profiled before, so.. Jan 21 05:11:24 Interesting, groxx. Jan 21 05:12:07 groxx, that's small code Jan 21 05:12:35 any ideas guys? anyone? Jan 21 05:12:41 I restarted my phone, didn't help Jan 21 05:12:44 * hesperaux has no idea Jan 21 05:13:12 i am debugging with chrome Jan 21 05:13:56 There is a good chance I'll switch to RecyclerView, because the API looks so much nicer. However, it does look kinda annoying in some ways (e.g., having to draw dividers manually), so I'm not sure yet. Jan 21 05:14:32 TacticalJoke: we have objects that bridge data and view-creation, so in the end we're left with a fairly-homogenous API. most of our adapters look like `new BaseAdapter(OneClass.class, TwoClass.class) { getView(etc) { return getItem(position).getView(context, convertView, parent) } }` Jan 21 05:15:11 adding a new type of data -> a new kind of view pool is trivial then. Jan 21 05:15:16 Yeah, that sounds nice. Jan 21 05:18:13 groxx: Totally random question, but I don't suppose you know of any JSON framework whose basic string type is mutable (e.g., a StringBuilder)? Jan 21 05:18:38 TacticalJoke: nope. it would be fairly un-java, I think :) Jan 21 05:18:52 I'm trying to avoid creating some kind of Frankenstein mutable-string monster class. Jan 21 05:19:11 Yeah, it definitely would be. Jan 21 05:19:20 also relatively hard for memory. if you e.g. insert a character, you'd potentially have to shift a crap-ton of data over. so you'd have to copy all the strings into string builders on the assumption that they _might_ be modified... which is wasteful. Jan 21 05:20:19 a while ago I was thinking about trying to build a trie-based "mutable" string thing that would keep the original around while it made sense, and copy when it didn't, and could make e.g. small prefix or suffix changes relatively efficient. but I never got around to it. Jan 21 05:20:25 I guess I'm thinking of something that basically says "Okay, have our original buffer. If you edit this, the entire object is invalidated and anything could happen". Jan 21 05:21:04 I had this idea of exploiting String.substring (which, on Android, returns a view into the original buffer {and not a copy}). Jan 21 05:21:06 TacticalJoke: in principle you could do that by buffering the entire input, and storing offsets in JSONObjects / etc. Jan 21 05:21:10 And JW gave me the idea of using GetChars with that. Jan 21 05:21:12 TacticalJoke: but then you have to buffer the whole input :) Jan 21 05:22:12 I have a very strange problem that I was hoping to get some pointers on. In short, an app fails to load libGLESv3 on Android 4.4.4. I have an identical phone with a different ROM that runs the app successfully. Jan 21 05:22:13 I investigated and found that libGLESv3 (a symlink to libGLESv2) is of permission 644 and visible to non-root user on the functional platform, but is permission 777 and visible to root ONLY on the troublesome platform. This goes against my basic understanding of the Linux permissions! What might I be missing? Jan 21 05:22:27 it'd be interesting to try, for sure. I don't know if it would actually have many benefits though. e.g. if you keep a single copy to a single reference, the whole thing is kept. might make sense in some cases though. Jan 21 05:22:33 groxx: How would I figure out the offsets? Jan 21 05:23:19 Honestly, I am half-tempted to write a custom JSON parser but it'd probably be a huge waste of time. Jan 21 05:23:30 It'd be so nice, though. Perfectly tailored to my use-cases. Jan 21 05:23:43 TacticalJoke: probably parse it by hand. json is pretty easy as long as you do it carefully (e.g. handle string escaping correctly), there's no crap like xml namespaces or schemas to worry about. odds are some existing project would have a decent framework you could build off. Jan 21 05:23:48 Okay. haha Jan 21 05:23:51 I thought you meant that. Jan 21 05:24:17 or maybe one of the tokenizers has a way to get the offset? dunno. Jan 21 05:24:23 Yeah, maybe. Jan 21 05:24:27 tokenizer in a pull-parser, that is Jan 21 05:24:58 maybe it's called something else in json-land. stream parsing, anyway. though I don't know _why_ you'd expose the offset, so I have some doubts. Jan 21 05:25:28 Jackson's stream parser exposes quite a lot. e.g., you can get the original char[] of a string value if you want it. Jan 21 05:26:18 The funny thing is that it says this: "Textual contents are not guaranteed to start at index 0 (rather, call getTextOffset()) to know the actual offset" Jan 21 05:26:24 Talking of offsets. Jan 21 05:26:30 not sure if that'll bite you with multi-byte characters, but yeah :) Jan 21 05:26:35 God, yeah. Jan 21 05:26:51 I wish java was utf-8 :( Jan 21 05:27:00 why would AT&T tell Samsung to use Dalvik by default, instead of ART, on the Note 4? Jan 21 05:27:04 Yeah, it's such a pain. Jan 21 05:27:08 I just noticed this and I'm switching to ART. Jan 21 05:28:08 "The requested URL /devices/tech/dalvik/art.html was not found on this server. That’s all we know." Jan 21 05:28:19 i guess some apps don't work with ART? Jan 21 05:28:20 first link on google is borked Jan 21 05:29:08 i see Jan 21 05:29:18 maybe conspiracy Jan 21 05:29:33 Agamemnus: it's pretty rare that that's the case. I think there might be JNI issues though? though that's largely just related to PIE afaik. Jan 21 05:29:42 The JSON RFC isn't long at all: http://www.ietf.org/rfc/rfc4627.txt Jan 21 05:30:15 TacticalJoke, try reading the SIP RFC Jan 21 05:30:40 or ldap! Jan 21 05:30:45 oh god Jan 21 05:31:14 HTTP/1.1 is really bad. Jan 21 05:31:17 I tried once. good god. Jan 21 05:31:26 "lightweight" my ass. Jan 21 05:31:54 lol Jan 21 05:33:04 sip does seem like it wins though, yeesh Jan 21 05:33:36 One thing that bothers me about the idea of writing a JSON parser is the encoding thing. There's no guarantee of UTF-8, for example. Jan 21 05:33:38 I much prefer http://tools.ietf.org/html/rfc1149 Jan 21 05:34:03 I tried to read SIP once. I barely got anywhere. Picked up a book instead. SIP is ridiculous for what it does... Jan 21 05:34:46