chicken-boot.exe is an actual Chicken compiler. In fact, you are
reminding me that when building from the boot .c files in the
distribution archive, the two-stage boot process is arguably redundant.
I'll have to think about whether that can be conditionally shortened.
Anyways, if you are having build problems with chicken-boot.exe, you're
going to have the same problems with chicken.exe.
Too bad the stack size didn't fix it. Tomorrow I'll burn a CD of VS
2005 Express at a friend's house; I don't have broadband. I'll at least
see if I can duplicate the problem on my system.
Cheers,
Brandon Van Every
Mark Baily wrote:
Is the thrashing you're talking about occurring inside the actual
compiler or inside chicken-boot.exe produced by the compiler? The
infinite loop I get is inside chicken-boot.exe.
I tried the stack size thing, but it didn't fix.
On 12/24/06, *Brandon J. Van Every* <[EMAIL PROTECTED]
<mailto:[EMAIL PROTECTED]>> wrote:
Mark Baily wrote:
Hi,
I'm having trouble compiling v 2.5 on win32 using visual studio 2005.
After using CMake to produce the .sln file for visual studio then
building the Release (or MinSizeRelease) version, it gets to
using chicken-boot.exe to compile stub.scm and sits there on 100%
CPU usage in what appears to be an infinite loop.
I read in an earlier thread someone else had this problem running
inside a Parallels virtual machine, but I am not running inside a
VM and still get the problem. I also read that it might be that
Visual Studio 2005 Express Edition causes memory thrashing.
However, I am using Visual Studio 2005 Academic Edition which is
the full compiler. Also, the infinite loop does not occur inside
the compiler itself but rather inside chicken-boot.exe.
I also don't want to build using mingw as I need to link chicken
with other visual studio 2005 apps/libraries; I'm not sure if
this is possible with a mingw-chicken.
Has anyone else come across the problem and know a solution?
I guess I'll have to download VS 2005 Express and see if I can
duplicate the behavior. At least we now know it isn't limited to
the Parallels Virtual Machine, nor to VS 2005 Express.
On MSVC we used to set C_DEFAULT_TARGET_STACK_SIZE=300000 to avoid
some kind of thrashing problem. Something about Chicken emitting
so many C functions and MSVC handling that gracelessly because of
poor memory design. I was unable to observe any problem with my
Visual Studio .NET 2003 compiler, so at one point I removed the
code, but I left a historical footnote. Try changing line 863 of
CMakeLists.txt
SET(C_DEFAULT_TARGET_STACK_SIZE \(128*1024\))
to
SET(C_DEFAULT_TARGET_STACK_SIZE 300000)
and recompiling for MSVC. If that solves your problem then we
know the culprit. If it doesn't, you might try a larger value.
If that doesn't fix it, then we're back to head scratching.
Cheers,
Brandon Van Every
_______________________________________________
Chicken-users mailing list
[email protected] <mailto:[email protected]>
http://lists.nongnu.org/mailman/listinfo/chicken-users
------------------------------------------------------------------------
_______________________________________________
Chicken-users mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/chicken-users
_______________________________________________
Chicken-users mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/chicken-users
