1. exw.exe access violation error
- Posted by Henri.Goffin at sbs.be Oct 05, 2001
- 401 views
Hi everyone, got a problem. A little Eu-tility of mine ran satisfactorily each time I started my NT4 workstation (you know ... Start Menu -> Program -> Startup). But a few days ago it began to crash because of a memory violation: "The instruction at 004177f4 referenced memory at 007d1014. The memory could not be written" It is really puzzling because: 1) it happens only with the bound version, i.e. exw prog.exw still goes in any case, 2) most surprising it happens only when I'm logged in an NT domain, not when i log as a local user. In that case the bound version works OK too. I bet it has something to do with the user environment but i have no clue into which direction i should look. Henri Goffin
2. Re: exw.exe access violation error
- Posted by Robert Craig <rds at RapidEuphoria.com> Oct 05, 2001
- 404 views
Henri Goffin writes: > But a few days ago it began to crash because of a > memory violation: "The instruction at 004177f4 referenced > memory at 007d1014. The memory could not be written" In 2.3 the interpreter will have trace(3) which will allow you to log the exact Euphoria statement that was being executed when a machine-level crash like this occurred. Until then, you can insert print statements to pin down where the crash happens. e.g. fn = open("debug.dat", "w") .... puts(fn, "reached point A\n") flush(fn) .... puts(fn, "reached point B") flush(fn) etc. It's important to call flush() after each puts() since Euphoria does buffering of I/O. With a sudden machine-level crash, some data could be left in memory. A crash like this will sometimes go away when you make a trivial change to your program, or the way you are running it, or the environment. Sometimes the crash depends on the contents of some random location in memory. Most likely this crash is due to poking into an illegal memory location, or maybe passing bad arguments to a C routine. It could also be a bug in the interpreter, but the interpreter is pretty stable, and I don't know of any bugs like that. Regards, Rob Craig Rapid Deployment Software http://www.RapidEuphoria.com