Rings a bell (memory buffer size problem I seem to recall). Upgrade....
Otherwise . why? * BCI for foreign databases * UV-NET for UniVerse Regards JayJay -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: 31 May 2007 10:47 To: [email protected] Subject: [U2][UV]confusing memory problem Update: I posted it first to the wrong list [sbs] and got no answer. After playing around the last three day, I thought I found the problem: getpid() but after deleting the the debug-code I have the same problem: concatenation of strings doesn't work. the same code gives different results (I run memtest86 for 24 hours without problems) Hi All I am fighting for a week now with a problem: a variable in a BASIC program is changing itself. I work with Universe 9.6 on Windows XP I read a file, change the Field delimiter for each record and add it to an other variable und write this over a socket to another computer if the length is over 1024. The Important parts is after opening a file, doing a SELECT and reading the SelectList as idList For I =1 TO RecCount ImportID = idList(I) READ ImportRec FROM aFile, ImportID tResult=CHAR(19) : ImportID : @IM : CHANGE( ImportRec, @AM, @IM, 0, 1) rResult := tResult IF LEN(rResult) > 1024 THEN GOSUB SendResult: rResult = '' END NEXT I So far so good: With nearly all if my files it is working perfect (45000 Record with 450 Fields, or 200000 Records with 60 Fields ). The biggest chunk is 600kByte ( this database is not my own development and I can't change it) Now to the problem: A small file with 512 Records and 42 Fields is making my troubles. On the first run I get at least 2 faulty records (up to 10). If I survive this problems, I can read it 50 times over and over without any problems. For Debuging I write out with CRT: the ImportRec and the tResult and for safety I made a tmpVariable with the same formular as tResult and compare both: [EMAIL PROTECTED] [EMAIL PROTECTED] For example:ImportID=1162 ImportRec Dicht. Puchmotor~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~DE tmpVariable: 1162Dicht. PuchmotorDE tResult: Dict. PuchmotorDE62116 or another example: ImportID=2471 ImportRec Differential~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~DE tmpVariable: next 2 lines are one line 24712471Differential tResult: 2471DifferentialDE most of the time it hits the same record, but sometimes total others. If the program survives this first round, I can run it endless without any problem. If it is crashing, it gives me a memory error at 0x000801a4 or 0xffffffff or 0x000c01da, everytime with the instruction from 0x7c92103c. Oh I forgot to mention: it is a phantom, after the crash the pid in Windows is gone, but Universe has it still in LISTU and I can LOGOUT the pid. I changed the motherboard and the Ram to be sure to avoid a faulty memory or chipset errors. Another fun: if I change the sequence of the CRT commands I get a different output: once I tried to calculate tResult bevor I printed the ImprtRec and found ImportRec with an extra letter in the middle and the length was exacty as it should be (the last byte was missing) Sometimes the tResult is filled with CHAR(19) : ImportID : @IM (up to the real length) Another: all the variables are Ok, but rResult gets a different tResult added The faulty records can be everywhere, but I never had it at the beginning, the two examples are: 19 (the 3th after the SendResult) and 360 (the first after SendResult) out of 512 Does anyone has a glue what is happening with this program? -- Mani ------- u2-users mailing list [email protected] To unsubscribe please visit http://listserver.u2ug.org/ ------- u2-users mailing list [email protected] To unsubscribe please visit http://listserver.u2ug.org/
