[MLton-devel] FW: cygwin segfault
Mike Thomas
miketh@brisbane.paradigmgeo.com
Wed, 14 May 2003 15:46:41 +1000
Thanks.
| -----Original Message-----
| From: Stephen Weeks [mailto:sweeks@sweeks.com]
| Sent: Wednesday, May 14, 2003 3:26 PM
| To: Mike Thomas
| Cc: MLton@mlton.org
| Subject: RE: [MLton-devel] FW: cygwin segfault
|
|
|
| > How do I get that missing symbol "_Thread_returnToC"?
|
| It was a bug that only showed up on Cygwin, where MLton needs to
| prefix symbols with _. But we forgot to do it in one place. I've
| fixed it.
|
| > $ mlton -stop o ffi.c
| >
| > gives me "ffi.1.o" rather than "ffi.o"!
|
| Yeah, that was an accidental change in behavior recently introduced.
| I changed it back.
|
| You can get both of the fixes with latest experimental release,
| 20030513, available at
|
| http://www.mlton.org/experimental
|
| One other change in this release that is relevant -- instead of using
| 'include "mlton.h"' in the ffi.h, use 'include "libmlton.h"'.
|
| This release also includes the new MLton.Finalize structure, which may
| come in handy with C/SML programming.
|
| Also, I wanted to point out that the ffi function in ffi.c that you
| sent isn't safe because it holds two ML pointers, a1 and a2, live
| across MLton_callFromC, which (like any SML code), may do a GC,
| invalidating the pointers. One way to test for such problems is to
| compile main.sml with "-gc-check first", which causes a GC to occur at
| every GC point the first time it is reached. If you do that, you will
| (correctly) see a segfault.
|
| I hope things begin to work a little better.
|
| Now, back to FFI and OpenGL :-).
|
|
-------------------------------------------------------
Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara
The only event dedicated to issues related to Linux enterprise solutions
www.enterpriselinuxforum.com
_______________________________________________
MLton-devel mailing list
MLton-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mlton-devel