|
I have definitely seen SYNON code do that.. IFM had a program that checked to see if PO's were received (3-way match).. It would read through the ENTIRE IMHIST file looking for completed receipts.. It didn't even use a Key for VENDOR or anything. It would take 20 minutes for the screen to appear and it would have exceeded the # of reads.. I think MAPICS put a fix in to limit the search for X number of reads, which wouldn't work since it wouldn't find the records.. I added a key to the lookup and the program went from 20 minutes to 1 second. I understand that tools are used to produce these programs, but there has to come a time where you just have to say, well.. The tool is done and we have to fix this problem by hand. Michael Franchino Custom Systems Corporation http://eax.cussys.com 973-726-0202 X214 -----Original Message----- From: mapics-l-bounces@xxxxxxxxxxxx [mailto:mapics-l-bounces@xxxxxxxxxxxx] On Behalf Of Konrad Underkofler Sent: Thursday, November 06, 2003 3:37 PM To: MAPICS ERP System Discussion Subject: You bet! Steve, It actually can and could be worse. With Synon being a brute force code generator it replicates routines whenever a file gets touched. A lot of the performance work they did to improve performance removed relationships that generated code. Can you imagine every program that references warehouse always having to check if the warehouse was still there, even after the record was edited the first time? Some record entry transactions were spawning 100+ ios per enter key. It was also difficult to have Synon use key access, it really does want to read the whole file at times. Since every program opened every file just about they ended up doing the shared open thing which is the big hit going into a COM program. They used to strip off a lot of the excess code but I don't think it is done anymore, for example, look at all the hidden fields on a display. Synon put every field on the display and then hides the unused ones. In an old version of client access it actually blew up the terminal emulator because hidden fields are transmitted to the client, overflowing the buffers. Some code is also due backwards compatibility to older versions of RPG. The multiple MOVEL's are an example of that. Not to mention that old routines still in the model but unused still generate programs and code even if they never get executed. It is a major accomplishment that COM works at all! The original six month rewrite of OEI turned into a three year slog and thousands of fixes since that time that we all still enjoy. Regards Konrad -----Original Message----- From: sjones@xxxxxxxxxxxxxx [mailto:sjones@xxxxxxxxxxxxxx] Sent: Thursday, November 06, 2003 1:39 PM To: mapics-l@xxxxxxxxxxxx Subject: This is what our ALF's buy?!? This is actual code from program AMBQGDFR. Notice field WUZ0NB. OK define the field. Nothing unusual here..... * Define work field &Chr C MOVEL *BLANK WUZ0NB 1 Now blank out the field.... Must you tell the program to move blanks 4 times? Actually if you look farther down it blanks out WUZ0NB a few more times. I really wish IBM would disipline the compiler so we only had to tell it to do things one time :-) MOVEL *BLANK WUAAGX MOVEL *BLANK WUZ0NB MOVEL *BLANK WUZ0NB MOVEL *BLANK WUZ0NB MOVEL *BLANK WUZ0NB MOVEL *BLANK WUCOCD MOVEL *BLANK WUDXCD MOVEL *BLANK WUZ0NB MOVEL *BLANK WUZ0NB MOVEL 'Y2U0005' W0RTN GOTO OGEXIT ENDIF Now what are they trying to do? OH wait maybe MAPICS is getting into the hardware business & are making sure everyone has to buy a really big computer.. C MOVEL ALCQTX WUZ0NB C MOVEL ALCRTX WUZ0NB C MOVEL ALCSTX WUZ0NB C MOVEL ALCDTX WUZ0NB The field WUZ0NB is on 205 lines of code in this program. I have seen the execute a subroutine that only has a BEGSR and a ENDSR statement in it, but this takes the prize!!! This is not the only field they do this with either in this program. I think we now have a better understanding why COM is so resource intense. Hope everyone gets as good of a laugh from this as we did.... Steve Jones _______________________________________________ This is the MAPICS ERP System Discussion (MAPICS-L) mailing list To post a message email: MAPICS-L@xxxxxxxxxxxx To subscribe, unsubscribe, or change list options, visit: http://lists.midrange.com/mailman/listinfo/mapics-l or email: MAPICS-L-request@xxxxxxxxxxxx Before posting, please take a moment to review the archives at http://archive.midrange.com/mapics-l. _______________________________________________ This is the MAPICS ERP System Discussion (MAPICS-L) mailing list To post a message email: MAPICS-L@xxxxxxxxxxxx To subscribe, unsubscribe, or change list options, visit: http://lists.midrange.com/mailman/listinfo/mapics-l or email: MAPICS-L-request@xxxxxxxxxxxx Before posting, please take a moment to review the archives at http://archive.midrange.com/mapics-l.
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.