Replies

Unbelievable! It’s the USB!

Following MerC’s suggestion, I used Tclock to continuously monitor CPU usage, and noticed that when I plugged in any USB device, the CPU usage returned to normal levels, and then sky rocketed again as soon as I unplugged it. If I disable USB in this profile, the CPU usage remains normal.

There was an additional feature that I had forgotten about, which is when I move to battery power, the CPU clock speed halves, but this effect is nothing like the USB effect.

I will see if I can get any joy out of IBM, but in the meantime if anyone knows enough about Win98SE USB implementation to be able to shed any light on this, I’d be very interested.

Mark

Unfortunately IBM Thinkpads haven’t heard that memories cheap – I’m looking at about 200US for another 64MB (max expansion) where I am (although that’s a very long way from the centre of civilisation).

Performance on the whole isn’t an issue – except for CPU intensive tasks as mentioned above, which is why I’d like to understand what’s going on.

Is that the south african TClock or the japanese TClock?

While we’re on the subject of useful apps, is there anything that will do the equivalent for Win98SE that windows task manager does for win2k – ie list all processes together with their CPU usage. I’ve found plenty of process monitors, but none that show CPU usage for each process in real time.

Thanks again

Norton AV2001. I had already tried eliminating all suspects by ending tasks through task manager, with apparently no effect. I then stripped everything except systray from my startup, hklm run, hkcu run, hklm run services, leaving only systray and explorer in the task list.

When I start system monitor under these conditions, the initial CPU usage is 3%, but as soon as I so much as move the mouse, this goes up to about 25% and stays there, if I then open a simple app (eg notepad), the usage goes to 75-95% and stays there. Sometimes it fluctuates by about 10%, and occasionally (for one reading) dips towards 0, but then always bounces back again.

I don’t understand how system monitor assess CPU usage, and maybe this is my problem, because it seems that an awful lot of nothing is using my CPU time – but only as soon as I start any app.

BTW, I am running this on an IBM thinkpad 240, so do’t know if that makes any difference.

I guess my only thoughts are that either I have a very difficult to find nasty, or that a complete reinstall might help, althought that’s an awful lot of hassle for very uncertain gain.

It is probably as simple as checking in advance whether the individual exists in the table underlying the main form, but I can’t work out how to use a SELECT statement to create a temporary recordset whose members I can count (and if the count equals 1 can know I’m in trouble).

Might as well answer my own question as I’ve found DCOUNT, which seems to work quite well, testing if DCOUNT(“IDParent”, “ParentFormTableName”, “IDParent= ” & Me.ID) is greater than 0.

I don’t always save every new or changed object because I may throw those away so not saving is simpler than deleting

That’s exactly the point – I can’t seem to find a way to not save without causing chaos. Every time I want to save an object, the multiple object save dialogue box pops up, and if you deselect any object, that’s when the trouble starts. It seems to be worst if you choose not to save changes to a module, but any form with attached code can cause it. I can’t find any way to specifically save one object.

BTW, does /decompile work with A2K SR1a? I couldn’t seem to get it to do anything (it certainly didn’t help).

Sean

Gulp!

Found one post through google referring to a KB article (which I can’t find), suggesting deleting everything that was changed before the save operation that created the problem – so only had to delete 2 forms and import from backup.

My question is, though, given that access doesn’t allow you to save things individually, how are you supposed to prevent this happening? Saving each and every change would be a complete pain, especially if changing the name of a variable (for instance) throughout a project.

Anyone else had this problem?

Opera 5.12 – chokes all the time. Not it’s fault maybe, but severly limits usefulness. chokes on hotmail, chokes on this lounge, chokes on MS knowledge base.

Can be very fast, but also, over my slow connection, has a habit of just hanging on a slowly loading page and not recovering.

Download manager also managed to stuff up a lot of files – abotu 10% in a recent extensive pdf download, compared to none for getright. I know getright’s a dedicated download manager, but the point is you REALLY need it for IE, which frequently hashes up downloads, and I hoped Opera would do away with it.

The way Opera saves web pages is also not as useful as IE, particularly if you subsequently want to move those files around.

Tried 5.12, may keep it as a second browser, but doesn’t free me from IE, unfortunately.

Parent and child forms based on separate tables. Parent table has primary key (study number), and second unique key (individual ID), child table has primary key (individual ID – same value as parent). Parent and child forms are linked on non key values (name or village/hamlet depending on the subform).

For various reasons, the child table and parent table are separate, and the child form passes some information from it’s record back to the parent on the press of the button, and it is on this event that key violations can occur if the individual entered in the subform has been entered previously in the subform. However, it is possible that that individual has been entered before through another subform, in which case they would have a record (as determined through individual ID) in the parent table but not the child table, or entered through this child form but subsequently deleted from the parent table, in which case they would have a record in the child table but not the parent table, or the individual is being reentered, in which case they will have records in both.

If a record already exists in the parent form, then the user must start entering all data again, as something has gone dramatically wrong.

If there is a record in the child form but not the parent, things might not have gone dreadfully wrong at this stage, so this is when I want to renumber the existing record in the child table and allow the passing of information back to the parent form as normal.

For a more detailed description of the whole process, see my previous post:

http://www.wopr.com/cgi-bin/w3t/showthread…&vc=1#Post14017

Thanks

Just when I thought I had it.

There are 2 tables involved that could suffer a key violation as the result of the data entered into the subform – the table underlying the parent form, and the table underlying the subform, or both. My plan was to test the parent form first, as if the individual had already been “used”, the subform selection process would have to start afresh, then to test the subform table, in which case I would renumber the existing entry (with a number identifiable as having been changed in this fashion), and allow things to continue. This, then, was the code:

NotInParentFlag = 2
Parent.Refresh
NotInParentFlag = 1
If Me.Dirty = True Then Me.Refresh
NotInParentFlag = 0

and the error handling routine:

Select Case Err.Number
Case 3022
If NotInParentFlag = 2 Then
MsgBox “This individual is already in the database for this survey.” & _
“You must choose again”
Parent.Undo
Me.Undo
Parent.Hamlet.SetFocus
Me.Visible = False
ElseIf NotInParentFlag = 1 Then
DoCmd.RunSQL “UPDATE [Household Data Corrections] SET ID=500000+ID ” & _
“WHERE ID=” & Me.ID ‘Run update query to change number of existing entry
Resume
Else
MsgBox “ERROR, value of NotInParentFlag is ” & NotInParentFlag
End If
End Select

(Sorry, indentation structure doesn’t seem to show in preview of message – not sure why)

Only problem was, even when I chose an individual who already had a record in the table underlying the subform, but definitely not in the table underlying the main form, the error always happened at the Parent.Refresh statement, presumably because that also forced a refresh of the subform.

Any suggestions as to how I can do this? It is probably as simple as checking in advance whether the individual exists in the table underlying the main form, but I can’t work out how to use a SELECT statement to create a temporary recordset whose members I can count (and if the count equals 1 can know I’m in trouble).

Thanks

Dear Charlotte,

Forcing a save on dirty does cause a 3022 error. Not sure why moving the focus never did, but I have at least a workaround (though it will require some recoding).

I have 2 supplementary questions:

1) At present I have my error handling routines in separate procedures, as they are the same for many procedures. I cannot find anyway to resume the procedure that caused the error unless the error handler is in the same procedure. Am i being stupid?

2) If I am, and there is a way to resume a named procedure, is there any specific way to pass or find out the name of the procedure that caused the error? I could put it in a global variable before I call the error handler, but global variables have a habit of being empty when the code breaks.

Thanks

PS have now ordered some books from overseas, so maybe won’t have to bother you so often!

The trouble is Charlotte, I never get error 3220, whatever the circumstances, only error 2110. If I then end the code, and try and save the form, then I am informed of the key violation in access. I have tried resuming if error is 2110, but then get error 2165 (as I hide the subform when the focus returns to the main form), and if I resume on errors 2110 and 2165, the focus stays with the subform, and nothing else happens. I never get error 3220.

So, how can I tell that the reason I cannot move the focus away from the subform is because there’s a key violation?

Thanks again

I hadn’t thought of using a property of the form to hold the information. The Forms(strVariable) was exactly what I needed. Thanks.

really? share please.

Slightly off message, but I have been trying to find a way to expand 2 drives – I would ideally like the root of C and D expanded when I open explorer, but %SystemRoot%explorer.exe /n,/e,c:, d: ( with or without after c and d) doesn’t work. Have tried powerdesk – no good.

Any suggestions?

Whilst we’re on topic, the first window of explorer I bring up after boot often has my computer expanded showing only c: as available, and has to be collapsed and reexpanded to show others. Why?

Thanks Charlotte,

>1) As you have discovered, the button’s click event happens before the exit event of the

>control. However, the afterupdate event of the control will still fire, so you could try

>using that.

Actually, that’s not been my experience at all. All the events I referred to are exit

events, eg if Control1 has the focus, and it has an OnExit procedure of If Control1=1 Then

Me.Control2.SetFocus, when I click my FormClose button while Control1 has the focus and has

value 1, the focus passes to Control2 and the form does not close. I have also confirmed

that the OnClick procedure of the button does not even start to run. Sorry if I had not

made myself clear. I suppose I could try and do things with mouse move and finding out if

the last key was a tab, because most of these OnExit events are just helpfully tabbing over

a lot of secondary questions if the answer to the primary question is Negative, so only

really need to fire when tabbing rather than mousing out of the control. Ho hum.

>How does the beforeupdate event of a form or subform force refreshes and what are

>”conditional” comboboxes and subforms?

Just for information (as my nomenclature is not good), I refer to a conditional combo box on

one that varies the contents of its list dependent on the value of another control. The

obvious instance in my case is that once a village has been entered, the hamlet combo box is

populated with hamlets only in that village. I found I had to refresh the hamlet control on

entry in order for this to work. Similarly the subform seemed to need to be refreshed on

entry in order for it to display the correct data if the master field is changed in the same

record.

In fact, after further research, it is the very act of leaving the main form to enter the

subform that forces the main forms Update event – as this happens before the record is

completed, I could not therefore use the BeforeUpdate event to check that the correct

procedures have been followed.

>don’t use it because it would be very, very bad programming

– thanks, I’m learning. I’m also lazy, however, and I was thinking of the situations where

many procedures call the same subroutine, and I want to stop execution of all the parents

when certain conditions are met in the sub, which would mean testing for that flag in each

parent. Yawn (especially with the limitations of the replace function in VB editor –

suppose I could always cut and paste into a text editor). But point most definitely taken

and I won’t even think about it again.

I’m afraid that your method of accessing help topics resulted in a blank grey pop up screen for me. Boo ho. If this problem is so widespread, why is there nothing in the MSKB about it? The only issues dealt with there are terminal help errors.

Oops, am I being naive again?