Magical Data Item Disappearing Act causes package build failure

[Crazy_About_JDE]

Hello, list! I successfully built a full package on October 8. One month later when I tried to build another full package, R9621 stopped after about one hour and gave me a gazillion syntax errors in buildlog.txt. We traced the problem to a semi-missing data item RPPC. To fix it, I had to delete data item RPPC (mnRenewalProbPercent) from DD7334/F9200, then add it back in via Work with Data Dictionary Items (P92001).

In a nutshell, the build process couldn't find the numeric data item, so it inserted the default (an empty string, probably in quotes), which caused all subsequent code to be flagged with syntax errors.

Has anyone else experienced this? There is an Oracle Global Support Solution ID 200812849 for this, titled E1: Package: Full Build error - n15l0002 - n15l0003: "When trying to build a full client and server package it fails on N15L0002 & N15L0003." To me, that is a red flag.

Any ideas how in the world my data item got corrupted or semi-deleted?


>>>

Here's the back story (plus relevant screenshots, attached):

The syntax errors in buildlog.txt looked like this:
<ul type="square">[*]\\...\include\n15l0002.h(64) : error C2061: syntax error : identifier 'INVALID0'
[*]\\...\include\n15l0002.h(68) : error C2059: syntax error : '}'
[*]\\...\source\N15L0003.c(412) : error C2065: 'DSD15L0002' : undeclared identifier
[/list]

The two most significant culprits appear to be:
<ul type="square">[*]N15L0002 - AREF Calculate Blended Amount
[*]N15L0003 - AREF Calculate Existing Lease Revenue
[/list]

Global Support's first response was that I need to apply a bunch of ESUs. Har har.


When one of our developers used OMW to open D15L0002 (the business function template), she got these errors:
<ul type="square">[*]"JDBRS_GetColumnSpecs of data item RPPC failed. See JDE.LOG for details." [OK]

[*]"The data item size for the structure member, mnRenewalProbPercent (alias - RPPC), has changed since this data structure last changed. The length has been fixed and you need to resave to fix specifications. Any static type definitions of this data structure needs to be regenerated." [OK] [Cancel]

[*]After clicking OK, the Data Structure design window appears. Structure member mnRenewalProbPercent shows nothing in the Data Item column.
[/list]

When she saw RPPC missing, she tried adding it, but got this message: "The Alias field name entered already exists as an Alias."

I poked around and found an RPPC record in DD7334/F9200, but no child records in F9202 or F9203.

Global Support sent me screenshots of what the data item should look like, so I deleted the RPPC record from F9200 and re-added it. After that, I was able to successfully build an update package with these two business functions.

 

[Crazy_About_JDE]

Hello, list-- I found a resolution to this problem, but I would still like to know if that has happened to anyone else OR what I might be able to do to find out how it happened in the first place.

It turns out that my package build problem was not entirely fixed. After re-inserting the RPPC data item, the syntax errors caused by N15L0002 and N15L0003 went away, but I continued to get similar syntax errors caused by B0900008 (Return Account Balance) and B0900090 (Trial Balance Calculate Balances) -- though there wasn't a data item problem with these.

The syntax errors looked like this:

\\...\PD7334\package\WPF061115\include\b0900008.h(126) : error C2061: syntax error : identifier 'KEY13_F0911'

\\...\PD7334\package\WPF061115\include\b0900008.h(127) : error C2059: syntax error : '}'

\\...\PD7334\package\WPF061115\include\b0900008.h(246) : error C2146: syntax error : missing ')' before identifier 'lpDS31107Keys'

\\...\PD7334\package\WPF061115\include\b0900008.h(246) : error C2081: 'LPDS31107KEYS' : name in formal parameter list illegal



Global Support had me try some things and then build update packages against the new, "broken" full package (11/9); nothing worked. However, when another analyst built the update package over the 10/8 package, it worked fine. This led me to search for everything on the deployment server that changed between 10/8 and 11/9. There were four header files modified in \jdedwardsoneworld\pd7334\include; one of them was F0911.h -- modified 10/11.

I compared this version of F0911.h to the previous version (4/17/06) and found the KEY13_F0911 did not exist in the new version. (Under what conditions would it have been deleted or removed?) When I copied the KEY13_F0911 typedef section from the old to the new, I was immediately able to build update packages again. My full package ran last night without a hitch.


Now my question: Is there a way to know what happened? OMW Logging shows the F0911 being checked out, designed, and checked in on 10/11, but it doesn't say anything specific. (The developer says she probably added an index.) What's weird to me is that the F0911.h on her workstation still shows a datestamp of 4/17. I would think it would say 10/11. So now I'm just confused.


Thank you in advance to anyone who can offer a lead. And merry Christmas!