Just to be absolutely sure that my code has no obvious errors I have run every command file and eep through RCode today - after the crashes mentioned below.
Yes, there were a few mistakes, about 20 in around 300 files - too many, I agree - but none which struck me as _obviously_ likely to cause any problems and they had not been picked up by TRACE and they had been working as far as I know. I admit that I don't use RCode _every_ time I make a change, I usually only check with it if I've created a large command file or made a major change. During today's marathon checking session I've got it set up to a point where I'm happy with the way it and RStyle interact with _my_ code. Thanks Sami for making me react and do these checks (again). It will be interesting to see if there are any changes to the number of crashes I get. Regards, Alastair. ----- Original Message ----- From: "Alastair Burr" <[EMAIL PROTECTED]> To: "RBASE-L Mailing List" <[EMAIL PROTECTED]> Sent: Sunday, December 29, 2002 5:37 PM Subject: [RBASE-L] - Criteria For My Unstableness With R:Base6.5++ Build 1.862 > Let me play Devil's Advocate for a moment: > > Most things being equal, I would expect wrongly written code not to work and > give an error message but not to either crash the program and/or hang/crash > the PC. This is especially required when testing blocks of code - if you > can't test and get a "sensible" error message then it's very hard to work > out what's wrong. > > I removed CrashGuard this morning and since then R:Base has crashed twice - > more or less par for the course. Both times I needed to re-boot my PC and I > got no error messages to guide me as to why it crashed. It wasn't too hard > to work out why on these occasions as I was working on a new block of code > and I got the syntax wrong (see posting on using mdi forms). Both these > crashes came while using TRACE which, surely, is supposed to check the > syntax before it runs it.
