On Mon, May 21, 2012 at 10:04 PM, Javier Pimás
<[email protected]>wrote:

> Could you put a self halt in NBGLXContextDriver>>createContext: and step
> over line by line to see which one is the offending one?
>


ok the line is:
colormap := display createColormapOn: window visual: visualInfo visual
alloc: AllocNone.

and at that point the relevant local variables:

window is 634
display is a NBXLibDisplay( 0x896EFD0 )

it hangs when I touch visualInfo in the debugger, then eventually the VM
crashes with the following in the terminal:

out of memory

C stack backtrace:
./NBCog/NBCog[0x8095489]
./NBCog/NBCog(error+0x19)[0x8095709]
./NBCog/NBCog[0x8075a5e]
./NBCog/NBCog(createClosureNumArgsnumCopiedstartpc+0x54)[0x8076ab4]
./NBCog/NBCog[0x8066bed]
[0xb57ab5e9]
./NBCog/NBCog[0x808f694]
./NBCog/NBCog(main+0x38a)[0x809668a]
/lib32/libc.so.6(__libc_start_main+0xf3)[0xf75800f3]


Smalltalk stack dump:
0xffe7ac34 M Delay>schedule -1218533460: a(n) Delay
0xffe7ac4c M Delay>wait -1218533460: a(n) Delay
0xffe7ac64 M InputEventPollingFetcher>waitForInput -1247953104: a(n)
InputEventPollingFetcher
0xffe7ac90 I InputEventPollingFetcher(InputEventFetcher)>eventLoop
-1247953104: a(n) InputEventPollingFetcher
0xffe7acb0 I [] in
InputEventPollingFetcher(InputEventFetcher)>installEventLoop -1247953104:
a(n) InputEventPollingFetcher
0xffe7acd0 I [] in BlockClosure>newProcess -1218533756: a(n) BlockClosure

Most recent primitives
wait
signal
millisecondClockValue
millisecondClockValue
primSignal:atMilliseconds:
millisecondClockValue
wait
signal
wait
basicNew
millisecondClockValue
signal
primSignal:atMilliseconds:
millisecondClockValue
wait
primGetNextEvent:
at:
class
wait
signal
... (this cycle repeats)
Aborted

Reply via email to