|
pokeyman posted:Beat you by over an hour! On the other hand, it's been an awful quiet chatroom since I joined...
|
# ? Sep 4, 2013 05:14 |
|
|
# ? May 17, 2024 02:25 |
|
Axiem posted:On the other hand, it's been an awful quiet chatroom since I joined... I was only there during work hours pacific time, then left to get dinner and stuff.
|
# ? Sep 4, 2013 06:58 |
|
Axiem posted:On the other hand, it's been an awful quiet chatroom since I joined... Then start chating! I'm on throughout the workday (EST) and I'll be there at night if I'm not busy doing other important things (video games, eating, in that order).
|
# ? Sep 4, 2013 15:12 |
|
I wonder if you guys could help me out on this. I posted a question over on SO: http://stackoverflow.com/questions/18602332/where-is-this-vertical-spacing-coming-from-in-uilabelview but I'm not getting much on it. I figured there'd be about a 95% chance the response would be "your constraints are probably wrong" when I'm almost certain that's not true and frankly it's easier here to expound upon further details if people need them. The main thing I'm getting at is why the gently caress does the UILabel behave in this particular way in the iPad layout? This may be an A/B problem but I'm convinced that's where things are going sideways.
|
# ? Sep 4, 2013 18:54 |
|
I literally just had an unassigned bool evaluate as YES over and over again no matter what I did. I finally just wrote out that it equals NO when I defined it. I had no idea that this could happen, and when you define a bool it would equal NO by default, but I guess not?
|
# ? Sep 5, 2013 06:08 |
|
If it's local to the function/method then no, its initial value will be whatever happened to previously occupy that same memory location. If it's an instance method, then yeah, its initial value should be NO. Something else has to be setting it to YES somewhere, so set a breakpoint on it and find out what. edit: for clarity: Objective-C code:
Doc Block fucked around with this message at 06:19 on Sep 5, 2013 |
# ? Sep 5, 2013 06:10 |
|
Thanks, it was the first thing you mentioned. (is allocated on the stack and the initial value will be whatever previously occupied that same memory location)
|
# ? Sep 5, 2013 06:17 |
|
LP0 ON FIRE posted:Thanks, it was the first thing you mentioned. (is allocated on the stack and the initial value will be whatever previously occupied that same memory location) This is actually something the static analyzer will check. If you choose "Analyze" in the build menu it should show an analyzer warning about the boolean being uninitialized and potentially garbage. Edit: and it looks like XCode 5 catches it as a vanilla warning at build time, which is pretty nice. ultramiraculous fucked around with this message at 10:22 on Sep 5, 2013 |
# ? Sep 5, 2013 10:19 |
|
File management makes me want to die. Not that it's badly implemented or anything, it's just so booooooring and tedious.
|
# ? Sep 5, 2013 16:31 |
|
LP0 ON FIRE posted:Thanks, it was the first thing you mentioned. (is allocated on the stack and the initial value will be whatever previously occupied that same memory location) Enable the warning for uninitialized variables in your build settings.
|
# ? Sep 5, 2013 17:12 |
|
ultramiraculous posted:This is actually something the static analyzer will check. If you choose "Analyze" in the build menu it should show an analyzer warning about the boolean being uninitialized and potentially garbage. Xcode 4.6 does too at my work. But I have the same Xcode at home I think?? Weird. Toady posted:Enable the warning for uninitialized variables in your build settings. I'd rather it just show the warning like it does in my project at work. I haven't tested it in ARC here, but that's the only difference I can think of.
|
# ? Sep 5, 2013 17:39 |
|
LP0 ON FIRE posted:I'd rather it just show the warning like it does in my project at work. I haven't tested it in ARC here, but that's the only difference I can think of. I must not be understanding you. Xcode 4.6 will display an inline warning and a Fix-It, often right after you type the offending code. The warning can be enabled in build settings (-Wuninitialized).
|
# ? Sep 6, 2013 21:31 |
|
How do you close a UIDocument without saving it? Am I correct that you just don't call -closeWithCompletionHandler? Related: this is my current plan to duplicate an open document into a new document, and leave the original file untouched: 1. Duplicate the open document's file (does not have saved changes) into a temporary directory 2. -saveToURL:forSaveOperation:completionHandler with the new document URL ("DocName Copy" or something) into the Documents folder 3. Move the duplicated file into the Documents folder Does that sound right? I was surprised that -saveToURL:forSaveOperation:completionHandler removes the file behind it if you save to a new URL.
|
# ? Sep 6, 2013 23:00 |
|
LP0 ON FIRE posted:Thanks, it was the first thing you mentioned. (is allocated on the stack and the initial value will be whatever previously occupied that same memory location) Also a good reason you should use calloc if you are doing any C stuff; it zeros out the memory as opposed to malloc leaving whatever garbage laying about. Fun fact: fools often claim that malloc is faster but in fact on most platforms calloc uses VMM optimization tricks that make it almost the same speed or even faster. On Apple platforms and Windows, it just allocates virtual address space. On Windows, it goes further... The reads all happen from the same zero page with copy on write semantics, and the system scrubs free'd pages in the background to keep a pool of zero'd out pages ready when you attempt the first write. Not sure if Apple platforms go that far, but I know they don't bother actually allocating the pages and zeroing them out until you use them.
|
# ? Sep 7, 2013 04:43 |
|
Ender.uNF posted:Also a good reason you should use calloc if you are doing any C stuff; it zeros out the memory as opposed to malloc leaving whatever garbage laying about. One thing I didn't know until I ran into it in practice, is iOS devices will zero larger mallocs but the simulator will not/did not as of around two years ago. It's basically for the VMM optimizations you're talking about, based on what Apple is saying. We basically had a bug for a short time where some image processing was showing up corrupted in the simulator and fine in device, because a malloc was used and some part of the image data wasn't overwritten.
|
# ? Sep 7, 2013 05:18 |
|
Ender.uNF posted:Fun fact: fools often claim that malloc is faster but in fact on most platforms calloc uses VMM optimization tricks that make it almost the same speed or even faster. On Apple platforms and Windows, it just allocates virtual address space. I highly doubt that, virtual address space comes in 4KB or larger chunks, which can be a lot of overhead. I know for a fact that Windows will only do that for large allocations, or if the page heap is being used (a heap implementation that places all allocations at the end of a page, followed by an inaccessible page. It turns heap block overflows into segmentation faults making them super-easy to debug) malloc is undeniably faster than calloc (zeroing out the allocation not only takes time, it also may thrash the cache), but it hardly matters
|
# ? Sep 7, 2013 13:34 |
|
From the why-can't-it-just-work department: my MFMailComposerViewController is FUBAR. No keyboard appears (except flashing when you hit Cancel), the layout is messed up (and I didn't block the email address - that actually appears that way): What it should look like: If you leave the app and return, the keyboard is suddenly there. But if you hide it, it's gone forever again. The same code in its own project works fine. What switch did I flip to mess everything up? I know it's not a lot to go on, but any clues?
|
# ? Sep 8, 2013 20:41 |
|
Now with the thrill of video! https://www.youtube.com/watch?v=6kp9_SSITeY
|
# ? Sep 8, 2013 20:58 |
|
lord funk posted:From the why-can't-it-just-work department: my MFMailComposerViewController is FUBAR. Are you presenting it from a background thread?
|
# ? Sep 8, 2013 21:11 |
|
bumnuts posted:Are you presenting it from a background thread? No, main thread. Edit fixed: ready for this? In iOS 7 beta 6 using drawViewHierarchy:afterScreenUpdates:YES randomly breaks layout and animations from that point forward. Thanks to someone on the apple dev forums for that, otherwise I never would have found it. lord funk fucked around with this message at 22:39 on Sep 8, 2013 |
# ? Sep 8, 2013 21:21 |
|
A disturbing amount of things in iOS 7 are still hilariously broken, especially on iPad. If it's any consolation, MFMailComposeViewController works fine on iPhone.
|
# ? Sep 9, 2013 00:51 |
|
Doc Block posted:A disturbing amount of things in iOS 7 are still hilariously broken, especially on iPad. Nah, seems fine to me: (Not taken during an animation, it just likes it that way.)
|
# ? Sep 9, 2013 00:56 |
|
Also, I hate the stupid "every view controller has to set its own tint color and status bar appearance" thing. Even though you're supposed to be able to set a global tint color, stuff like UIImagePickerController isn't smart enough to use a tint color of its own, which makes the Cancel button unreadable if your app has a light tint color. And Apple really really REALLY wants every app to have a white background and a status bar with black text. It's really getting on my nerves now that I'm trying to put the finishing touches on my app's iOS 7 UI, because the status bar style keeps getting changed back to default instead of the light style because reasons
|
# ? Sep 9, 2013 01:22 |
|
For me, UIImagePickerController (in camera mode) keeps the status bar (it doesn't tell the system to hide it), so on iPad it collides with the "switch to front camera" button. I'm presenting it as a full screen modal view controller. I cannot figure out how to get rid of this status bar. The usual status bar hiding methods don't seem to do anything.
|
# ? Sep 9, 2013 01:38 |
|
It wouldn't surprise me if a bunch of that stuff got fixed after merging in some top secret feature(s), so another beta wasn't doable. My Calendar app icon changes into one from another app. I thought it was random, but today I cracked the code: whenever I open a folder, Calendar immediately takes on the icon of that folder's second app.
|
# ? Sep 9, 2013 02:31 |
|
I got called into work over the weekend because XCode 5/Clang building in release mode breaks a library that we've used (and haven't touched) for the past 3 years. It's going to be a long week.
|
# ? Sep 9, 2013 13:56 |
|
lord funk posted:Nah, seems fine to me: I figure we'll get GM of iOS 7 for iPhone/iPod Touch only and then we'll continue on with iOS 7.0.1 betas for iPad.
|
# ? Sep 9, 2013 14:52 |
|
Doc Block posted:Also, I hate the stupid "every view controller has to set its own tint color and status bar appearance" thing. Even though you're supposed to be able to set a global tint color, stuff like UIImagePickerController isn't smart enough to use a tint color of its own, which makes the Cancel button unreadable if your app has a light tint color. 1000x thiiiiis. Do you want kind of hell I've been living in? I decided my app's colors would change based on the user chosen performance screen colors. You have to do soooo much manually that the 'hierarchy' system they set up is basically useless. Plus there are plenty of new problems with tinting, like whether the alpha component does anything, whether tintColor or attributes matter, and of course the multitude of beta bugs (including flying views when the tint color changes!). quote:And Apple really really REALLY wants every app to have a white background and a status bar with black text. It's really getting on my nerves now that I'm trying to put the finishing touches on my app's iOS 7 UI, because the status bar style keeps getting changed back to default instead of the light style because reasons What killlls me is that the bar tint color changes based on the background content to try and prevent this, but it never, ever, ever looks good. You get washed out gray colors all the time, and what's the point of changing the background color dynamically if the text / icon color doesn't change aaaaaagggggg.
|
# ? Sep 9, 2013 15:10 |
|
The entire status bar seems like a terrible hack. We were having a problem where we temporarily hid the status bar for one modal screen, but if you left and re-entered the app at that point it would silently re-layout the entire view navigation stack so that when the status bar was restored it would now overlap the top bar. There was no way to cancel or catch that process so we had to show the status bar while regaining focus and then hide it a moment later just to trick the OS into not doing that.
|
# ? Sep 9, 2013 15:42 |
|
Yeah, the whole "put stuff under the status bar" thing is really really stupid IMHO. Honestly, I think Jonny Ive is a great hardware designer but completely out of his league with UI design.
|
# ? Sep 9, 2013 16:00 |
|
Was iOS6 this wishy washy right before it was released? I'm dreading this
|
# ? Sep 9, 2013 16:12 |
|
Doc Block posted:Yeah, the whole "put stuff under the status bar" thing is really really stupid IMHO. I think the look is great, but yeah the execution is horrendous - especially regarding the status bar.
|
# ? Sep 9, 2013 16:20 |
|
Oh hey it's time for absolutely worthless "here's what to expect from iOS 7" articles: http://thenextweb.com/dd/2013/08/11/5-things-to-know-when-converting-your-app-design-to-ios-7/
|
# ? Sep 9, 2013 21:16 |
|
I'm having a really weird issue in the provisioning portal saying that a provision expired when I know it hasn't. On my iPad it says it expires next year. I have another iPad I'd like to put the provision on, so I was planning to grab it from the portal, but welp.
|
# ? Sep 9, 2013 21:22 |
|
So in what circumstance would you actually want content under the status bar or tab bar view? It seems to go against what users have been used to since day zero. I really pondered that hard while adjusting my layout to account for the fact that view bounds include altogether useless space, but I couldn't think of anything.
|
# ? Sep 10, 2013 03:23 |
|
So the bars can be translucent and look cool when you scroll up. True depth, deference to content, and whatever other BS they're on about. Oh, your app's UI doesn't use some descendant of UIScrollView? Tough luck.
|
# ? Sep 10, 2013 13:58 |
|
I get the idea behind it. You can see a bit more of your content (albeit blurred) so the screen should feel bigger. What bothers me is that they solved the contrast issue (what happens when the background content is extremely light / dark) for one use case, which is when the bar is white. If you want a different color bar, it will inevitably look like crap and be unreadable. Doc Block posted:Honestly, I think Jonny Ive is a great hardware designer but completely out of his league with UI design.
|
# ? Sep 10, 2013 15:30 |
|
lord funk posted:What bothers me is that they solved the contrast issue (what happens when the background content is extremely light / dark) for one use case, which is when the bar is white. If you want a different color bar, it will inevitably look like crap and be unreadable.
|
# ? Sep 10, 2013 16:14 |
|
lord funk posted:I think no one told him that designing a UI is not designing a finished product, it's designing a set of tools. This is probably closest to the truth. There's definitely more than one place where this comes up, where if you're not using the UI elements in a context very similar to how Apple uses them, then things get messy. Another example is using the UIPicker inline in a table: it's a cool idea, but your table better be white or it's not going to look right because you're not allowed to change the text color.
|
# ? Sep 10, 2013 16:22 |
|
|
# ? May 17, 2024 02:25 |
|
The new UIDatePicker can also go take a flying gently caress. It spills over its bounds considerably, begging the question of "why the gently caress did they give it those bounds in the first place"?
|
# ? Sep 10, 2013 16:51 |