Corrupted Run-time library

[dabney]

I have an issue with a corrupted run-time library. I have been left in a precarious posistion of fixing an issue on an OS I have little knowledge of.
This is the error I am receiving.
NRCAVB::Theresa's:::>run odin:[menu]mainmenu

%OTS-F-INTDATCOR, internal data corrupted in Run-Time Library

I need to know how to fix a corrupted Run-time library. I also need to know any possible resulting issues.
Let me know if you need more info.
Thanks for any guidance you can provide.

[Richard Brodie_1]

If it's happened recently, without a system update or similar, then it's likely a buffer/array overrun of some sort. Something has overwritten some bookkeeping data.

HELP/MESSAGE INTDATCOR, gives you the basic advice on the error.

For preference, rebuild from source with any bounds checking on. Have a look round for obviously bad input data. Try to narrow the fault down to a particular set of changes.

[dabney]

This is a pretty old issue. Nothing new has been installed to the best of my knowledge. I have been doing some research will installing the Fortran RTL kit resolve this issue?

[Dean McGorrill]

anything done recently to the system. another thread Dean


http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1117141

[dabney]

Nothing new done to the system in a while. Left over issues.

[John Gillings]

dabney,

This is not an OS issue. It doesn't mean the RTL itself has been corrupted, it means a data structure owned by the RTL has been corrupted at run time, within the address space of your process.

There are no on-going issues, or effects on other processes, it's just a coding bug in your mainmenu program.

Likely candidates are an array index out of bounds writing memory it shouldn't, a buffer overflow, or a pointer miscalculation.

Hand the problem back to the developer of the program. Debugging can be tricky because the code signalling the INTDATCOR is victim rather than perpetrator.

[Jon Pinkley]

John,

I agree with your assessment "Debugging can be tricky because the code signalling the INTDATCOR is victim rather than perpetrator." but I know we have a vested FORTAN program, LIST_TV.EXE (the LIST utility originally written by someone at ICR and submitted to the DECUS library) and porting it to Alpha was not a simple recompile. There was a change in the 6.2 to 7.x Alpha VMS timeframe that the new FORTRTL_TV broke some existing executables. If you do a Google groups search for "OTS-F-INTDATCOR" you will see that is the case.

For example, this one:

http://groups.google.com/group/comp.os.vms/browse_thread/thread/168a149f6337b872/97c155fb88973680?lnk=st&q=&rnum=1#97c155fb88973680

The patch fixed the problem in LIST_TV. Is it possible that the original .EXE was incorrect and using some unallocated memory? Yes, that is a possibility. Perhaps all the patch did was to move its private data a page up in virtual space, I really don't know what it did, but I do know that installing the patch makes the program work as it did before.

I still think dabney should read my response dated Apr 12, 2007 10:18:42 GMT in the following thread:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1117141

That may not be the problem, but if it is a translated Fortran image, there is a high probability that applying the patch will solve his problem. I doubt he can "Hand the problem back to the developer, as he stated that this was part of "Left over issues"

dabney,

My note in the referenced thread had a pointer that was valid as of April 12, to a patch kit that may fix that problem.

Apply the patchkit, and please report back as to whether applying the patch did fix the problem.

Jon

[dabney]

Thanks very much I will let you know how it works out.

[dabney]

The link to the patch kit does not work. Do you or anyone else know where I can find it?

[Jon Pinkley]

http://mvb.saic.com/freeware/vmslt00b/vu/forrtl_tv-t0600-1.zipexe

grab it before it disappears :-)

BTW, Google is great for finding things like this.