Svenson

-- Yesterday -- Previous week -- This week -- Next week -- Tomorrow --

 

1999-12-21     Tuesday

 

MIM-ccclv

 

 

Contrary to expectations there was no ice on my car. I guess this is what they call micro-climate. The thermometer, which is in the most exposed part of the garden, was at -0C, with fresh ice on the ground. My car which stands about 20m further was totally ice free. The mist clearly had not decided yet where to freeze. The day turned out very gray, with clouds hovering about 50m above the ground.

There was a problem in Belgium with invoices that were not transferred to the financial system (OFI) trough the general ledger. They were passed to OFI for the Accounts Receivable side. This is a technically impossible situation under normal conditions, there had been a crash (a program looping in the daily housekeeping had been killed). A few days ago the situation was mentioned but I assumed that they had killed the whole series of programs than so I advised to re-run the transfer. However this did not solve the problem. Surprise.
After some brain storming, well a one man blizzard really, I think I found the problem and the possible solution. Normally orders are archived after all processing is done, in this case, thanks to continuing the normal sequence of programs after killing some, the archiving was done before the orders were posted. And posting afterwards did not pick up the 'missing' orders. Recovering is not simple because the temporal invoices (-OM3 members) get deleted by the posting interface even if the orders aren't found. So they have to reconstruct the temporal orders, based on the difference between the General ledger and the actual Invoices (line by line comparison). Then they must make a copy of the posting interface (ivc025r) and adapt it so that it will search the order-archives (QH). They'll be having fun.
I did put down some documentation so I'll know what to do when I actually have to make an adaptation. I set the stage but am keeping the curtains down.

And another problem was mentioned by Loek . He is one of the functional people from the logistics module and he uses the OMSI-3 programs only as input for his module. What he comes up with is something strange. He makes an order and gets a specific messages about an item that cannot be delivered from a specific warehouse, he claims this messages should not come. He then changes the warehouse to one from which the item can be delivered so he doesn't get the message anymore. When he changes back to the original warehouse he doesn't get the message again. He is right that either the message should come each time or never.
When I tried the same it worked but that doesn't mean a thing because there are so many possible variation that I may just miss the mark. After an initial inspection of the code I think he is not describing the situation as it happens. I sense a great black hole here.

One for our black holes specialist :

What is the difference between a great black hole
and a small one ?

As I expected the target problems had nothing to do with the use of frames. And it works in Netscape as well. I just wonder why it looked different yesterday. Posting went good enough. I only produced one error. I (finally) placed a link at the bottom of the daynote page, pointing to the Daynote gang. BUT, capital offense, the link is wrong. O dear, o dear. It linked to //www;daynotes.com/daynotes.htm
I am waiting my trial. Maybe if I plead guilty but unaccountable (isn't that a daynotes requirement?) I can get away with it.
Oh no, Bob frowns. Send shivers down my spine.

A real problem with frames is that scrolling in a frame only works when that frame has focus. No problem when you open the window because the last frame that is filled will have focus. As coder you can select it, I fill the actual daynote (main frame) as last so that scrolling, for example with a wheel mouse, scrolls the text. The problem occurs if the user opens the page and then goes off to some other place or program. How does he return? Most likely by clicking somewhere inside the browser window. The frame he happens to click on becomes the active one, if it is the top scrolling doesn't work. Another method to get the browser back is by selecting it from the task manager or task bar or docking bar (or whatever it is called in your favourit OS). In this case it seems to be random, well I have tried it in Win98 and Win95 and it seems random, while OS/2 seems to be random but with a slight preference for the top frame. The back and forward buttons in Opera and Netscape keep the focus where it was (most of the time).

Well Brian, go ahead and check day ccclxv. And don't forget I live in code.


Adios

 

-- yesterday -- tomorrow --
-- Home -- Links -- Sjon -- Calendar -- The Gang --

Mail Me
Svenson © 1999

Every day is too short.