View previous topic :: View next topic
|
Author |
Message |
VVRAVINDRA
New User
Joined: 04 Aug 2008 Posts: 46 Location: Chennai
|
|
|
|
Hi,
We are doing a major code changes, almost 200 programs are getting affected. The developers want a handy way of doing load module comparisions between two endevor staging libraries
For example:
If there is a load module AAAA in ENDEVOR.STAGE1.LOADLIB
and the same load module AAA is in ENDEVOR.STAGE2.LOADLIB, we want to compare these two libraries. Doing a 3.13 compare is not fetching the desire d results.
It would be of a great help if any one could assist of a differnt IBM tool or method to acheive this.
Please let me know if the question is unclear or more details are needed.
Thanks a many for the help, |
|
Back to top |
|
|
daveporcelan
Active Member
Joined: 01 Dec 2006 Posts: 792 Location: Pennsylvania
|
|
|
|
If the code is changing, why are you comparing Load modules?
Of course they are going to be different.
What useful information could a load module compare show?
Wouldn't comparing source code be better?
Wouldn't Regression testing be important?
Wouldn't Unit testing be important?
Wouldn't Systems Integration testing be important? |
|
Back to top |
|
|
VVRAVINDRA
New User
Joined: 04 Aug 2008 Posts: 46 Location: Chennai
|
|
|
|
Well, I agree to many of your questions. And that is what we are doing. We have completed UT and we are in integrated testing phase.
But the client's perception is totally different, even if we are showing the timestamps in the load module, they are not getting satisfied with that. They want to see a total compare load module.
So we are in this well of confusion. |
|
Back to top |
|
|
Akatsukami
Global Moderator
Joined: 03 Oct 2009 Posts: 1788 Location: Bloomington, IL
|
|
|
|
They want to see a total compare in what way? Do they have anybody who can actually read the machine code, or just a bunch of trained monkeys who'll drool on the listings? |
|
Back to top |
|
|
dbzTHEdinosauer
Global Moderator
Joined: 20 Oct 2006 Posts: 6966 Location: porcelain throne
|
|
|
|
as dave said, not much use in comparing load modules.
it would be better to know what requirement exists
such as what info is needed
then members can provide suggestions.
personally, i would rather compare amblists. |
|
Back to top |
|
|
daveporcelan
Active Member
Joined: 01 Dec 2006 Posts: 792 Location: Pennsylvania
|
|
|
|
This is a partial listing of a Load module compare.
Is this what you get?
What is wrong with this from the client perspective?
What is this good for from anyone elses perspective?
Code: |
LISTING OUTPUT SECTION (LINE COMPARE)
ID SOURCE LINES
I - IEWPLMH Ì Ì ð m h
D - IEWPLMH Ì Ì ð m h
I - å00 CEE { å00 qóÐ ¶ °Ö} 0 qÕ0<
D - å00 CEE { å00 qóÐ ¶ °Ö} 0 qÕ0<
I - { ¿ { s { ¿ {
D - { ¡ { s { ¡ {
I - K jxæK pxæK ¸xæK txæk zK ÐxîK ¼z K [z K ¨wAK BwAK GwDk
D - K jx¸K px¸K ¸x¸K tx¸k zK ÐxïK ¼z K [z K ¨wBK BwBK GwEk
I - z:å0§óì qþ °áL-N L-y åø©ÎK kwFK yÞå0§óì qþ ) { +&Lø1èL-LË
D - z#å0§óì qþ °áL-N L-y åø©ÎK kwGK y®å0§óì qþ ) { +&Lø1èL-LË
I - uÙ8í jL-8 « i85L jÜ-Lñ i³î »Lñ8í » »K.L8w K.MçL83vM ÎK M x K M
D - uÚ8í jL-8 « i85L jÜ-Lñ i³î »Lñ8í » »K.L8w K.MçL83vM ÎK M x K M
I - 2 L- 8 L-L-ì j 3 (L-ì J& 2ì j 2 L- ü9 L-uUå}© å0§ 2ÉL- ü| L-< ~
D - 2 L- 8 L-L-ì j 3 (L-ì J& 2ì j 2 L- ü9 L-uVå}© å0§ 2ÉL- ü| L-< ~
I - j 2! ô 8í ô ô2 LÇ ³ L½LÇ8 L½L½3 L½2 L- ü8 L-L-3 (L-K KHJÌ ©
D - j 2! ô 8í ô ô2 LÇ ³ L½LÇ8 L½L½3 L½2 L- ü8 L-L-3 (L-K KHJÌ ©
I - Ø Ø Ø &Ø & &K êzâ Ø &Ø è&- ì ø *K ;~ÜK Ã{ - Â - À &
D - Ø Ø Ø &Ø & &K êzä Ø &Ø è&- ì ø *K ;~ÜK Ã{ - Â - À &
I - LÀLøå ©Wì j nS åØ© N udåØ© N ugåØ© N uÑåØ© N uÃåØ© N uÄ
D - LÀLøå ©Wì j nS åØ© N uØåØ© N ucåØ© N uÁåØ© N uÂåØ© N u¬
I - B®å\· ì ¥Ü ° åØ· \ -K }8©Q ° &\}u&°}Ð(\© ì0}4nO0 åØ·ã ° K
D - ° åø· jØB®å\· ì ¥Ü ° åØ· \ -K }8©Q ° &\}u&°}Ð(\© ì0}4nO0 åØ
|
|
|
Back to top |
|
|
VVRAVINDRA
New User
Joined: 04 Aug 2008 Posts: 46 Location: Chennai
|
|
|
|
Well, I don't want to eat away your valuable time. I know this is a much of a pain. I am grateful to each one's responses.
I took a toll out of you folks, sorry for that.
Thanks,
Ravi |
|
Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006 Posts: 19244 Location: Inside the Matrix
|
|
|
|
Hello,
Quote: |
Doing a 3.13 compare is not fetching the desire d results. |
From a load module compare, what would be the desired results?
What Dave posted is exactly what should be expected when load modules are compared. . .
Often, simply showing the difference in module lengths shows that there has been a change. Unfortunately, this only works most of the time. . . |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
load libs compare is just a waste of time!
there are many factors which influence a LOAD MODULE image
for example
what tool are You using for change management...
some of them when promoting will use iebcopy
some of them will relink the LOAD..
at this point even if functionally equal the two load modules compared byte by byte will be different because of different internal timestamps
sometimes even iebcopy invokes the binder under the covers so also in this case the modules will non compare
given a reasonable sound change management environment and proper access to load libraries
the audit logs of the change management system should guarantee that a load is consistent across the different environments |
|
Back to top |
|
|
dbzTHEdinosauer
Global Moderator
Joined: 20 Oct 2006 Posts: 6966 Location: porcelain throne
|
|
|
|
amblists (Link here) are the correct formatted report to provide information on load modules.
STORY TIME ON
years ago, NCR source modules, when written with the SOURCE MODULE EDIT Routine, contained a up-arrow (X'5F') at the end of a block (actually it is BlockVL +1).
wrote a machine to machine data transfer utility (1970's, before internet) that could transfer any file from one mainframe to another.
But no way can you I/O input this 5F, because it was at the end of a Block + 1 byte.
we would transfer source files, and the customer would do disc prints and compare. they would see the 5F in the original, but not the copied.
So, I had to write the original block to data+1 byte (with a 5F at the end of the data) and then rewrite the block with a block vli of 1 less.
the customer could do the print-disc and we got our money.
STORY TIME OFF |
|
Back to top |
|
|
|