search

Taryck / zaplink

Automatically exported from code.google.com/p/zaplink
0 stars 1 forks source link

Public section of XXX has been regenerated because differences have been detected #121

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
Version: 0.7

What steps will reproduce the problem?
1. Download package from http://code.google.com/p/abap-di
2. Try to export ABAP package YDI

What is the expected output?
All objects exported succesfully

What do you see instead?
Public section of YDIABSTRACTBINDINGBUILDER has been regenerated because 
differences have been detected
Protected section of YDIABSTRACTBINDINGBUILDER has been regenerated because 
differences have been detected
Private section of YDIABSTRACTBINDINGBUILDER has been regenerated because 
differences have been detected

And lot more

#Please provide any additional information below.

If I import back from this pacakge and run transaction SCI over imported code I 
get messages about missing method implementations (a lot of them). In fact, 
when I check in the XML file, the implementations are there, but non in SAP. 
You can check this by comparing what you imported by SAP transport from abap-di 
page.

Original issue reported on code.google.com by dominik....@gmail.com on 30 Dec 2010 at 9:45

GoogleCodeExporter commented 9 years ago 
1 ) Could you provide us the file generated by transaction ZL_INFOS (or program 
ZAPLINK_INFOS). If framework is not working you could try with program : 
ZAPLINK_INFOS_STANDALONE
2) Could you provide us the XML File

The message "xxx section of yyyy has been regenerated because differences have 
been detected" is due to the fact that this section haven't been regenerated 
and will create a section with sligt difference after a new import and with 
lead to a new code signature. See issue 71 for more details.

The real issue you describe is : "If I import back from this pacakge and run 
transaction SCI over imported code I get messages about missing method 
implementations (a lot of them)"

This issue look like an other one I4ve already handled, but i'm not sure, I'll 
try to check this week-end. Sorry for the delay, but I do not noticed you issue 
before today.

Original comment by taryck%b...@gtempaccount.com on 7 Jan 2011 at 2:13

GoogleCodeExporter commented 9 years ago 
Generally I have a problem with the source code of the library I'm writing: 
code.google.com/p/abap-di. You can download SAP transport from library's page. 
I also provided you with the ZAPlink file generated by some older version of 
ZAPlink (the one that did not override the sections), if it helps. 

Generally what I did is I tried to import the ZAPlink file (that comes from old 
ZAPLink version) with the newest version. When I did that, some methods were 
not restored at all (transaction SCI reported missing implementations). So I 
copied/pasted the methods and generated SAP Transport (available from 
code.google.com/p/abap-di). 
When I try to export with the newest ZAPlink from my SAP system, I get warning 
about 'regenerated sections' and my code stops compiling.

The story I provided above is actually to state what you discovered from my 
last post: that I actually have two sort of problems (one is when exporting and 
one is when importing). 

Hope this helps.

If not, I can elaborate some more, but please tell me on what section would you 
like more comments.

Thanks

Original comment by dominik....@gmail.com on 7 Jan 2011 at 2:25

Attachments:

GoogleCodeExporter commented 9 years ago

Original comment by taryck%b...@gtempaccount.com on 7 Jan 2011 at 5:31

GoogleCodeExporter commented 9 years ago 
Hy,

I've installed Transport Request on my SAP system. I've got some errors like 
for class YDILINKEDBINDINGIMPL where check provide no errors (like SCI) and 
activation neither.
But there is, take a look to the provided screen shoot you've got two time the 
same interface : YDIBINDING & YDIELEMENT

Why it's not detected by SAP, i don't know, but this could be a reason why your 
loosing interface implementations.

I'll still investigate.

Original comment by taryck%b...@gtempaccount.com on 7 Jan 2011 at 6:23

Attachments:

GoogleCodeExporter commented 9 years ago 
Same trouble happend with CLASS (this makes SCI Short dumps) :
* YDIUNTARGETTEDBINDINGIMPL

I've run SCI after that it seem's to work.
I've exported and got the following changes (see Export log.zip).
Where :
* yyyy section of YDIxxx has been regenerated because differences have been 
detected : To ensure CODE SIGNATURE will be the same after import on target 
system. Those action are equivalent to menu Utilities->Regenerates Section in 
SE24. For info this might be the same problem your talking with you 
[http://forums.sdn.sap.com/thread.jspa?threadID=1839100&tstart=0 java anotation]
* ABAP of YDIxxx has been regenerated because differences have been detected : 
This clean up operation is removing all type pool not required, because SAP 
might add them sometimes. This action is performed in _CLEANUP_CLASS of object 
ZAPLINK_OBJECT. You could disable it by adding "RETURN." at the very first 
line. If you think this action is not good, you could open an issue for that to 
discuss the different points of views.

After exporting again I've no longer have any warning message. The XML 
container I've build is provided in YDI.ZIP

Original comment by taryck%b...@gtempaccount.com on 7 Jan 2011 at 7:19

Attachments:

GoogleCodeExporter commented 9 years ago 
After trying the way that I describe in 
[http://code.google.com/p/abap-di/wiki/Installation#commentlist your 
installation page] in the comment section.

I've got activation error due to Issue 120 which is corrected but will be 
released in Framework version 0.0.008. You might apply the fix manully on your 
current version.

Sorry for the trouble, Issue 120, was my mistake. I appologize for the time you 
loose on that Issue.

Let me know if after applying correction of Issue 120 you problem is over. I'll 
close the Issue then.

Original comment by taryck%b...@gtempaccount.com on 7 Jan 2011 at 8:39

GoogleCodeExporter commented 9 years ago 
As far as I understand what you have done is the following:

* IMPORT CASE: you identified that there was a problem with empty method 
implementations (Issue 120). This fixes the problem that when importing ZAPlink 
file some methods were not properly imported and therefore SCI reported errors.

* EXPORT CASE: you identified that for a couple of classes there is a case that 
the interface is included TWICE on the interfaces tab. Actually I don't see why 
you regard this as an error because SE80 clearly shows that one interface comes 
from the superclass and the other comes from the class itself. I'm not saying 
that I cannot remove one interface reference (I probably will) so that this 
weird situation is not happening, but since SCI does not report any error, I 
would not regard this as ABAP bug. 
But the above does not fix the exporting error, which is that the code gets 
regenerated. From what you write you did the following:
- ran SCI over the package (successful)
- ran export over package (reporting warning with regeneration)
- ran export again
You claim that the last export did not report any warnings. I'm sure it didn't 
because the previous export regenerated some sections. But what you did not do 
is RUN SCI AGAIN. 

Could I ask you please to run SCI on exported code again and report to me 
whether SCI reports errors? I'm sure it will.

I cannot see in your description how you fixed the EXPORT case. For me it looks 
like there is no change from your side so I bet it will not work as in the past.

Original comment by dominik....@gmail.com on 10 Jan 2011 at 8:02

GoogleCodeExporter commented 9 years ago 
"Actually I don't see why you regard this as an error because SE80 clearly 
shows that one interface comes from the superclass and the other comes from the 
class itself"
May be it's like :
"for testing duration short risk level harmless" it's a difference between EHP1 
and EHP2. But in our system adding more than one time a interface leads to 
compilation error, that why I have to fix it. So depending on your target, you 
should "think" EHP1 or EHP2...

For export there is nothing to fix, for now. Regeneratin section is done by SAP 
function module each time difference are detected. This is required to ensure 
same code signature after import on an other system.

For investigation, you could manully do it yourself by doing 
Utilities->Regenerates Section in SE24 on class that are reported to change by 
ZAPLink. You'll be able to investigate changes done to understand why SCI 
reports errors and provide us the section source code before and after 
regenerating sections.

I'll try to also test SCI errors on my EHP1 system as requested.

Original comment by taryck%b...@gtempaccount.com on 11 Jan 2011 at 9:21

GoogleCodeExporter commented 9 years ago

Original comment by taryck%b...@gtempaccount.com on 11 Jan 2011 at 11:24

GoogleCodeExporter commented 9 years ago 
We do not have the Same SAP version, this might be the reason of the 
differences.
When I use SGEN on your transport request i've got the following result (no 
errors, 78 object generated).

When I use SCI I've got :
* 3 implementation missing error (which is a warning object activated)
* 26 Dead Code Error in SCI

When I use SCI for the first time I've got :
* 1 Interface statement more than once (which is an error)
* 121 Dead code Error in SCI

I think that your sections contains pseudo code that is lost after SAP standard 
Section regeneration. Please try Utilities->Regenerates Section in SE24 on :
* CLAS - YDIBINDINGBUILDER
* CLAS - YDIBINDINGIMPL
* CLAS - YDICONSTRUCTIONCONTEXT
* CLAS - YDICONSTRUCTORINJECTOR
* CLAS - YDIDEFAULTCONSTRUCTIONPROXY
* CLAS - YDIDEPENDENCY
* CLAS - YDIINJECTIONPOINT
* CLAS - YDIINJECTORSHELL
* CLAS - YDIKEY
* CLAS - YDIMEMBERSINJECTORIMPL
* CLAS - YDISINGLEPARAMETERINJECTOR
* CLAS - YDIUNTARGETTEDBINDINGIMPL
You should get the same trouble with SCI.

Provide me sections before and after regeneration to understand the problem.
I do not have the same result as you on SAP Trai 7.01 EHP1.

Original comment by taryck%b...@gtempaccount.com on 11 Jan 2011 at 7:00

Attachments:

GoogleCodeExporter commented 9 years ago

Original comment by taryck%b...@gtempaccount.com on 28 Mar 2011 at 9:40

GoogleCodeExporter commented 9 years ago 
Same as Issue 120

Original comment by taryck%b...@gtempaccount.com on 29 Mar 2011 at 10:11