Irrlicht Engine

Hi there,

I' starting & stopping lots of Irrlicht-Devices in our iOS App and this is causing a crash in libGPUSupportMercury.dylib



Is anybody abled to tell me how to solve this issue?

Regards,
Daniel

Hi folks,

solution was to make sure that our method to create screenshots of Irrlicht + UIKit Elements would pause the irrlicht update-Loop.

Now it's working great!

Hi there,

we’re currently receiving loads of crash-reports from our customers, who are using Apple’s iPad Pro (with iOS 10.2). The crash happens in COGLES2Driver.cpp in clearBuffers() method when it’s calling glClear(mask).

Is this a known bug/crash in the ogl-es branch and will there be a fix ...

I solved this issues by checking also for iOS instead of checking for OSX only.

CFileSystem.cpp - 28
#if defined (_IRR_WINDOWS_API_)
#if !defined ( _WIN32_WCE )
#include <direct.h> // for _chdir
#include <io.h> // for _access
#include <tchar.h>
#endif
#else
#if (defined(_IRR_POSIX_API ...

Hi there,

we're using irrlicht engine within out two iPad Apps and since this morning I'm running into problems while trying to compile CFIleSystem.cpp.

//! Creates an empty filelist
IFileList* CFileSystem::createEmptyFileList(const io::path& path, bool ignoreCase, bool ignorePaths)
{
return new ...

Nadro wrote:Hi,

Thanks for a report, I'll try to check that during fixing ES1 driver in ogl-es branch (probably near next week).

Hi ndro, thanks for giving it a try! Looking forward to hear from you

CuteAlien wrote:If this is a new problem (since last week) it might be because of the recent merge from trunk. Nadro is working on this and the priority was on ES2. If it's been a while longer then we need an example to reproduce the crash.

Hey, it's occuring since years.

mongoose7 wrote:The problem is unlikely to be in this code. Trying putting a printfs before and after the glClear: printf("calling glClear\n"); fflush(stdout); glClear(mask); printf("glClear called\n"); fflush(stdout);

What will this do to solve the problem?

The following Code in COGLESDriver.cpp is crashing sometimes randomly.

//! clears the zbuffer
bool COGLES1Driver::beginScene(bool backBuffer, bool zBuffer, SColor color,
const SExposedVideoData& videoData,
core::rect<s32>* sourceRect)
{
CNullDriver::beginScene(backBuffer, zBuffer, color);

#if ...

mongoose7 wrote:Do you know how big an image is? If it is R8G8B8, then the size is WxHx3, not WxH.
Do you know the format of the original data? Maybe it is not R8G8B8. Maybe it is B8G8R8 or A8R8G8B8. Maybe it is JPEG.

Therefore i'm using NSData's 'lenght' property, which is related to these factors.

So...after some struggel with this topic i'm at the Point, where i can load an image, but the resulting image is really noisy, black and strange...


// Start loading the image
void *data = frame->getCurrentFassadenBildData(&size);

if (data != nullptr) {
loadImage->loadImageForBackground(data ...

bool FAAppFrame::getCurrentFassadenBildData(void **data, core::dimension2d<u32> *size)

*data = malloc(len);
memcpy(*data, [imageData bytes], len);

I can't help you any more - you're going to have to learn to code before you can make any progress. As an old C programmer, I usually prefer
void ...

"data = <value>" only changes the local pointer. To change the pointer in the outer function you need "*data = <value>". You also need to change the function prototype.

Hi,

I've changed my function now to

bool FAAppFrame::getCurrentFassadenBildData(void *data, core::dimension2d<u32> *size ...

Do you get any warnings? Look, the problem is in *your* code so you have to be prepared to debug it. For example,
void* data;
bool success = frame->getImageData(&data, &size);
but
bool FAAppFrame::getImageData(void *data, core::dimension2d<u32> *size)
and
data = (void *)[imageData bytes];
In ...

Always no solution yet...