>>>Nadya,
>>>
>>>I feel that this is getting beyond control.
>>>
>>>Why not start with the basics
>>>
>>>(1) if you are on a network, verify that all network adapters are 100Mbit and full duplex. The hardware Default sometimes results in Half Duplex
>>>
>>
>>Sorry for being dumb. How can I check this?
>I really do not know. Recently, one of my customers complained (startup time app over 5 minutes compared to 10 sec or so. I do a lot of tables open on startup. It turned out that he had his pc reinstalled and the network was half duplex on his site. At least that is what the hardware man said.
>But I think this is not very likely in your situation
>>

Haven't yet checked this.

>>>(2) next, change a record or 10, activate the coverage profiler and save them.
>>>What routines are taking up time and why ?
>>>
>>
>>Good idea. I'll try that, after our Add method would be fixed by my colleague.
>>

My colleague fixed Add method, so now it works fine.

>>>A couple of candidates are record validation, triggers.
>>>Maybe set reprocess is the guilty one. If it is -1 or -2 the system is spending its time trying to lock a record that is used by another user
>>>
>>

Our baseform doesn't set SET REPROCESS. So, we're using default 0 (indefinite) setting.

>>I think, I was the only user of this table in that moment, so SET REPROCESS should not be in affect. Anyway, I'll check our Basform class settings.
>>
>>>Also, 10 minutes for 4734 records is not that bad over a network
>>>
>>>just my two EuroCents
>>
>>Using Vlad's idea the saving time was decreased to ~1min. I need to run couple of more tests to confirm it, though.
>
>I am surprised. I had no idea that the pending window events took that much time. In fact what he is doing is avoiding the vfp process of getting interrupted. When I saw the code I would have given it 5 or 10% speedup.
>I find it really hard to believe that this has such an impact.
>I wonder whether the problem causing the delay disappeared overnight.
>

I probably should not trust my judgement about time without actually checking.

Ok, with Vlad's idea - Saving time = 13.74 sec.
without Vlad's idea Saving time = 15.99 sec.

Well, 13 sec. seemed to be a whole minute, while waiting :)

>Once you have cracked it, would you pls do a test with and one without the autoyield and let us know the difference ? Thanks
>>
>>Thanks for your ideas. Much appreciated

So, it's not that bad at all. However, nobody except me was working on this server at this moment. I need to run couple of more tests when Network is busy.

I'll try more tests on Monday.

Thanks again to you and Vlad.