return0

Members
  • Content count

    234
  • Joined

  • Last visited

Community Reputation

508 Good

About return0

  • Rank
    Member
  1. Local hash

    Seriously? I find this bizzare, have you never used a dynamic language like python, ruby, or lua? These use hash tables idiomatically lots.
  2. I mean live deployed crash reporting from NDK crashed in apps that users have downloaded from Play store. A common tool is Google Breakpad to parse minidump files and send them to a remote service, then use the unstripped debug symbols to get a useful trace. This is what HockeyApp does, for example. Mozilla have their own equivalent for Firefox. I've been experimenting, and can get good traces from some devices, but for others the dump looks weird - it appears truncated, as if it's not got all the stack frames from my app lib at point of crash. Was wondering if anyone had ecountered similar?
  3. Hi there, was wondering if anyone here has managed to get crash reporting working for Android apps using the NDK. I am using the HockeyApp SDK to try to get a stacktrace from a C++ app using Google breakpad. It seems to work great for some devices (samsung, asus), but doesn't work for other vendors. Interested to hear if anyone has got it or anything similar working - maybe crashlytics or similar?
  4. Can you write mobile games in C++?

    From my experience at tri-Ace, there are no Android NDK (C++) debuggers that work on all devices, only for a limited set of them.And according to the 2nd link you posted, you have to use quite a retarded hack to get breakpoints to work, assuming the device is supported in the first place.L. Spiro At work I have personally debugged a large C++ library on ten or so devices, mostly Asus, Samsung, Sony, LG and Motorola, without issue. I an aware that other engineers have debugged on other devices, including from smaller international Android handset manufacturers. We have a compatibility spreadsheet which indicates successful testing of both our library and development environment on 60 or so devices. All of these devices are Gingerbread or above; possibly prior to this support was less mature. I am sympathetic to you scepticism but in reality I've found support for debugging effectively ubiquitous. As for the "retarded hack", it is true that gdb takes a few seconds to attach. If debugging some user controlled event this is not an issue. If debugging application startup it is sometimes necessary to include an artificial delay (i.e., sleep(5) or similar). Not too onerous but yes, somewhat irritating.
  5. Can you write mobile games in C++?

    http://developer.android.com/tools/debugging/debugging-projects.htmlI see no mention of C++ there.L. Spiro Not sure what to tell you here, other than that it works. It uses the gdb version shipped with the NDK toolchain, and integrates with the Eclipse C++ view. See http://tools.android.com/recent/usingthendkplugin and http://stackoverflow.com/questions/17705109/how-to-debug-c-c-code-ndk-in-eclipse I guess you are unable to find more recent docs as ADT is being deprecated in favour of Android Studio.
  6. Can you write mobile games in C++?

    This is not true, you can use ADT and to 'visually' debug in Eclipse. It works fine.
  7. A hash code should be immutable and deterministic so is unlikely to be based on memory address?
  8. Well, you could triangulate the boundary, do an intersection test of all the grid points against the triangles, remove those which intersect then retriangulate with all remaining points? Wouldn't be particularly efficient but should work I guess... If using python, Shapely is a nice library for stuff like this.
  9. Construct a Delaunay triangulation from the vertices?
  10. Flash iOS Building

    Flash is free for commercial use, the Adobe toolchain is not. Free development alternatives exist. iOS will not run the Flash web plugin. You can build an Air executable and put it in the app store. You have to pay Apple for this, as a regular app.
  11. for and while inifite loops

    Threads... be careful with this.
  12. By "still need to sort opaque geometry" you mean to minimise statechange and not to avoid overdraw, right?
  13. [quote name='Hodgman' timestamp='1350973796' post='4993001'] [quote name='Waterlimon' timestamp='1350931178' post='4992843'] but it would be nicer to have it directly in the class like with everything else [/quote]This is the C#/Java way of thinking ([i]"corrupted OOP" IMO[/i]), where everything has to be in a class, not the C++ way of thinking. The first way of thinking leads to #1 below, whereas #2 is more C++ style (in my experience).[code]//#1 class Foo { public: int Bar( int ); static void UtilityHelper() { Bar(42); } private: int m_Baz; }; //#2 class Foo { public: int Bar( int ); private: int m_Baz; }; void UtilityHelper(Foo& foo) { foo.Bar(42); }[/code]The reason is that in #1, the private details such as [font=courier new,courier,monospace]m_Baz[/font] have been unnecessarily exposed to the utility/helper function. This means that if you're investigating a bug involving [font=courier new,courier,monospace]m_Baz[/font], then you have to treat [font=courier new,courier,monospace]UtilityHelper[/font] as a suspect, increasing the amount of code to read/maintain. In the 2nd style, the utility is known to only have access to the public interface, not the private details. [/quote] Does #1 actually work; on what instance is Bar invoked?
  14. Why not just some sets of functors implementing () or some invoke interface?
  15. I don't have any tooling suggestions, sorry; it just sounds really similar to the symptoms and profiler behaviour to a problem I have seen before. I guess you could skip your existing rendering abstraction layer and manually hack in hooking up your buffers? Good luck tracking it down!