thanks David. I like to be stingy with writing code, maybe ultra-economical. Seems to me that I use and reuse about 100 variables constantly, and that setting them to null instead of clearing and resetting them seems pretty efficient. If I don't hear any objections, I'll keep doing it until I get smacked down.
bill On Wed, Apr 20, 2011 at 11:42 AM, Fitts, David A. <[email protected]>wrote: > Greetings, > > In the past year and a half I migrated about 20 years worth command > files (100+) from our old Rbase version 2.11 database to version 9.1 (64). > > My oldest command files were pretty bad examples of coding, most of them > didn't even clear variables before running and many of them relied on other > command files to set up variables for use in the fashion I think you are > describing in your email. > > In converting my old command files of course I had to get rid of reserved > words that Rbase 2.11 had allowed and 9.1 didn't allow. I also had to change > a few commands like CONNECT for OPEN and SET POINTER with FETCH commands. > Aside from these changes, the commands created using 2.11 are all working > well in version 9.1 and it doesn't seem to matter whether they clear > variables before running or let the variables sit in memory for later use. > > Based on my experience so far, I think you can continue to do what you are > doing without any trouble in 9.1 but I would certainly defer to Razzak's or > any one else's recommendations. > > Regards, > > David Fitts > State of Maine > Risk Management Division > 287-3352 > 1-800-525-1252 > > > ------------------------------ > *From:* [email protected] [mailto:[email protected]] *On Behalf Of *William > Stacy > *Sent:* Saturday, April 16, 2011 12:40 PM > *To:* RBASE-L Mailing List > *Subject:* [RBASE-L] - setting variables > > Am migrating, ever so slowly, from 4.5++ (dos) to 9.1 (64). Have a large > ap that I've used to run my practice for many years, and use about 1000 cmd > files that use and re-use variables. > > I noticed that Razzak and others seem to have developed a careful set var > and clear var procedures in their cmd files, forms, etc. that seem to "start > over" with each use of vars. > > In earlier years I was aware of some memory leaks or something that would > over time degrade the performance of rbase, so I made a habit of declaring a > long list (about 100) of variables that I would use over and over again, > setting them only in this "setvar" cmd file that is always run right between > starting rbase and any workstation login file (password control, etc). > > This works well, and has enabled me (caused me?) to shorten my syntax to > the following example: > > 1. in the setvar file I have this kind of statement: set var t1 text=' ' > 2. in the command files I use things like this: t1='Personal health' > 3. at the end of each command file I have this: t1=null (unless I want to > keep "personal health" in the .t1 for the next cmd file, in which case I > just leave it as is. > > Thereby, in my cleverness, I have a minimum of coding and my variables > never get actually "cleared" until the ap actually closes. > > So my question is, am I just using sloppy programming, not making it clear > what is happening at each and every statement, or is this kind of > programming dangerous or even disallowed in 9.x? Has the memory leak of old > been fixed forever, or is it still happening? And does my system actually > help plug or eat least slow the (real or imagined) leak? > > TIA, and sorry if I asked this same thing a while back in a clumsier way. > I am now at the point of major or not so major rewriting of thousands of > lines of code. I would prefer the not so major, but can be convinced it > needs to be major. I am aware that with the placing of code within forms > themselves, many of my beloved cmd files may become obsolete anyway. > > Bill > > -- > William Stacy, O.D. > > Please visit my website by clicking on : > > http://www.folsomeye.net > > > > -- William Stacy, O.D. Please visit my website by clicking on : http://www.folsomeye.net
