I probably have to admit to not thinking about it one way or the other but you have a very good point.
Starting out as you mean to go on is easy if you know something in the first place or a change takes place that you consciously decide to avoid at that time. I'd be more than happy to change _my_ default to off if it showed any improvement in the way things work. I'm sure that Razzak has been through this time and again but maybe it would be nice to have explained why the R:Base default is ON. I guess that many, if not most, programmers won't change any default unless they have a good reason - which means coming up against a problem which a change solves... I know that I have to change EQNULL for a particular report to work but that's (probably) the only one that I change other than the "user friendly" things - error messages and the like - that I intentionally change on a permanent basis. At some point in the distant past I must have spent some time on SETtings in general because I have a table with every one entered, it's R:Base default, my preference, where it's defined, the required value and a comment. My start-up app uses this table to ensure everything is set consistently every time I start R:Base. Regards, Alastair. ----- Original Message ----- From: Dennis McGrath To: RBASE-L Mailing List Sent: Thursday, January 22, 2009 4:39 PM Subject: [RBASE-L] - Re: Typecasting problem I suspect it is not a problem if you start a project with it on, and code everything that way. But, when you are maintaining a huge project that lived for years before WHILEOPT, OFF is the best option. Dennis ------------------------------------------------------------------------------ From: [email protected] [mailto:[email protected]] On Behalf Of Alastair Burr Sent: Thursday, January 22, 2009 10:30 AM To: RBASE-L Mailing List Subject: [RBASE-L] - Re: Typecasting problem Interesting... I have to admit I didn't know without looking what mine was set to but I appear to have left it as ON as the default is ON. As I have a start up file that sets all the defaults as I want them then I must have chosen to leave this as the R:Base default but I don't know why I did that. As far as I am aware I have no problems as it is. Regards, Alastair. ----- Original Message ----- From: [email protected] To: RBASE-L Mailing List Sent: Thursday, January 22, 2009 3:53 PM Subject: [RBASE-L] - Re: Typecasting problem Jason: I'd be tempted to set whileopt off in your startup program and leave it there. The funny thing about having whileopt ON is that it doesn't always create problems in the exact place where it IS a problem. The memory problems just keep accumulating and eventually your app bombs out somewhere else with no connection to the actual problem. We don't take "surveys" on this list, but if we did, I'll bet that over 90% of programmers set whileopt off for their entire app. It was introduced way back in some DOS version when code speed was a huge issue. Whileopt ON made the code run faster (supposedly) but along with it came some stringent coding "rules" that you had to follow or else you ran into these memory issues. So rather than asking programmers here who sets it on or off, I'd like to hear from programmers who purposely set whileopt ON for the entire app, have no problems with it, and have done it for a particular reason. Karen Razzak and Karen, Thank you. Setting WHILEOPT to OFF for just that loop fixed it. Jason ----------------------------------------------------------------------------
