Another question about forcing a checkout

BBritain

VIP Member
Wow, I find myself on this side of the question. I NEED HELP! Aaaaugh!

Ok. I'm in E900 and I created a business function. I was on a remote system and had to log out, so I checked in the business function and left. ***Unfortunately, I did not compile the business function before checking in, so I get an error when I try to check out the BSFN with a message saying "....was not able to check out the object, but you now have the token....."

Now I tried Saving the object to a .par file / deleting the object on the machine / restoring the object, just to get the checked out status, but I find that because the BSFN was not compiled, the delete function ALSO errors out and never completely deletes the object.

So can anyone help me either A) Help me invoke Object Librarian and force check out, B) Help me completely delete the object or give me the steps to tell my CNC how to... or C) Something totally different.

Thanks in Advance,
Ben again
 
Ben,

In older versions, checkout could be "forced" by updating the F98222 to checked out status, there are several columns involved. I can't remember how this is done and it may be different in E9. You could get a debug log of a checkout to see what is updated when an object is checked out, or you could get a before and after checkout to see what has changed for the appropriate F98222 row.

The risk with this is, of course, that, if the object on the server is different to the one the client, then the object will be changed upon checkin.
 
Hi Ben,

I've seen 'similar' when I attempted to check-in a function, at a client's site - and did not have read/write authority to the Source and Include folders on the server. The Checked-Out 'flag' would get dropped and I would get an error that the check-in failed.

To validate the status of the check-in, I would use Beyond Compare to compare my local Source & Include folders against those on the server. If they were different, I would 'attempt' to copy the function, using BC, from my machine to the server. If I didn't have authority - BC would tell me I wasn't authorized (why couldn't E1 have done the same???)

Additionally, since I didn't have authority to the folder, the Check-Out was failing - too. Again, not authorized (maybe to read the contents of the folders)...

First step - make a physical copy of your functions - with the dos command, to some safe location. Include your .C and AND .H

Then make sure you do have access to the folders on the Deployment server.

If the issue that is occurring on your system is similar, you'll be able to complete the Check-Out, then copy the 'Saved' function (.c & .H) back to their correct locations on your developer machine.

I hope the resolution is this easy, for you!

Regards,

(db)
 
Ben,
Is this an NER bsfn? This would make sense if you never compiled it because the .c and .h files would never have been generated. I wonder if the errors in your OMW operations are because it can't find either of the source files.

If this is an NER, are the event rules still on your box? If so, you can try doing a build without it being checked-out. This will create the source files which you can copy to the deployment server and then retry the check-out.

If this is C bsfn, then disregard everything I've said
smile.gif


Good luck.
 
Thanks all,

First of all, yes, this is an NER.

Second, in trying to move the .c and .h files onto the server, do I need to get CNC to do that? and will they know where they go?

Thanks,
Ben (again)
 
You could try it yourself...you should have read/write access to the server.

The deployment server will have an E1 install with a folder structure much like a fat client in that there will be separate folders for each of the environments. If you're working, on your fat client, in an environment named: DV900, there will be a corresponding folder on the deployment server.

The (.h) file will go into the E900\DV900\include\ folder
the (.c) file will go into the E900\DV900\source\ folder

If you are unable to access the server yourself, CNC should know where they go.
 
If you are on the same machine that you created the function on, and you can still see your NER code, try this:
- Build the function.
- Save it to a .par file.
- Check out the function.
- Restore it from the .par file.
- Build the function.
- Check in the function.
 
Ugh - nothing like hijacking an old post...


I'm hitting the same issue as defined. NER was checked-in, there was an error on check-in. I go to check-out and I get the "....was not able to check out the object, but you now have the token....." message.

If I were to adjust the F98222 record for the object to match a checked-out object in the same project - the object now shows as checked-out, but I get an error on check-in that the object is owned by another project and the project ownership is blank in the Object Librarian Object Status window...

What is new, that I am missing?

ps... it has been a long time
 
Can you export the object out to a par, delete the "blank" record in the F9861 (the one without the dep server mkey), check out again, restore from par, check in and....profit?
 
Can you export the object out to a par, delete the "blank" record in the F9861 (the one without the dep server mkey), check out again, restore from par, check in and....profit?
I do have a healthy .par file. Since I no-longer have CNC authority - I'll have to see if the CNC is willing to try.

There is not a blank record in the F9861. It is a spec count mismatch error that kills the check-out.
 
Hi Daniel,

I am seeing the exact same problem with checking out an NER. The OMW logs are saying there are record count mismatches. This is the second time I have seen this problem. The first time, I solved it by deleting the NER and recreating it from scratch. That approach would be cost prohibitive for this second case.

How did you solve this problem?

Adam
 
I figured out how to fix this:

1. Delete the object. Check "From ALL Locations"
2. Checkout
3. Get from PY or PD (or restore from save).
4. Check in.

Voila.
 
Back
Top