Week 10, 2001 ,Svenson

---

| Previous week | Most recent | Next week | 05-03 to 11-03

e-Mail Sjon Calendar    Today Lair The Gang   External Links

Top <  Mon | Tue | Wed | Thu | Fri | Sat | Sun  > Bottom

Cool and dry but cloudy. Almost the whole day, just a bit of weak sunlight in the afternoon.
Current temp is 11.9C in and 0.3C out.

Busy looking up stuff in TeleSales and in TOS. That is a combination of Progress and HTML with some Java script thrown in. Talking to and AS/400 DB/2 back side with RPG programs.
My Progress understanding is moving in leaps and bounds (leaps forward and bouncing backwards again <g> ).
I am getting better in Progress but I am not getting finished with that stupid Validation-10. I have got to redo that. (it is indeed re doing a validation, and I am not good a revalidation, not in hospital, not out of it :-)

 

Of course the 'greens' will be against it, and when it is leaked to the general public under a headline like 'Genetically manipulated oranges' by some by some nutty green journalist you can count on a fiasco. I am enthusiastic about it. http://www.newscientist.com/dailynews/news.jsp?id=ns9999478 .

---

Hey, you don't want to mix up with me, I am genetically modified
Just came back fromm sunning on the beach.

 

Top <  Mon | Tue | Wed | Thu | Fri | Sat | Sun  > Bottom

Nicely cold with loads of frost as a start into a bright sunshiny day. With only late at night some clouds paying a visit and leaving their rainy mark.
Current temp is 12.1C in and 1.2C out.

Sorry, I lost the day.
But I keep looking.

---

 

Top <  Mon | Tue | Wed | Thu | Fri | Sat | Sun  > Bottom

Wet morning. It started to rain, hard, yesterday evening (around 22h00) and it only stopped this morning at around 9h00. But then it stopped to make way for bright untainted sunlight. Till sunset when we get some bright untainted moonlight.
Current temp is 16.9C in and 8.1C out.

Well, I have been searching but I didn't find anything. Another day lost in the mists of time.

I did some real work (on the validation-10) to day but again most time went to fire fighting.
The biggest timesink (^tm of Dan Bowman ) was a problem in Norway. Some time ago Iwan had a problem with the W0 file that grew completely out of proportion, he had to cancel the job. Of course cancelling a job destroys much of the possibilities for tracking the problem so I couldn't help him then.
Now the same problem occurred and he mailed me. The problem is an endless loop in a program (prm034r) for the maintenance of conditions. Now I know hat to look for and I started to analyse the program. I will (hopefully) finish that tomorrow.
Apart from that I got a call from Robert, in Belgium, about the sequence in which the invoices get printed. Usually the batch invoices are printed at night followed by the cash-sales invoices that are printed the moment an order gets status 'shipped'. This morning some cash-sales invoices were printed before the batch invoices (even though their 'batch' number was later). I have no explanation for this without seeing the system logs and even then there is probably nothing I can do.
A one off event I suspect.
And I have been waiting on a phone from Italy. But they never called.
And I have been looking in on TOS for a time.
And I ....

...well just a standard day in the mines, I'd say.

 

Hard to ignore http://www.theregister.co.uk/content/2/17419.html . And hard to swallow.

Is this a high flier or what? ( from Zannah )

---

Making something foolproof is difficult,
making something childproof is very difficult,
making something Syroid proof is close to impossible, but
making something Landon proof is impossible.

 

Top <  Mon | Tue | Wed | Thu | Fri | Sat | Sun  > Bottom

After a wet morning we survived a mixed sunny / cloudy day. With lots of soft rain and mostly warm. With a nice setting sun framed by glowing clouds.
Current temp is 17.3C in and 7.3C out.

We had to solve a small problem with 'tool tips' on a HTML page. The HTML is generated by Progress programs and I assisted Ronny with the HTML part while he was typing the Progress code. After a relative short struggle we got the stuff working.
The problem with the tooltips is that they use the onmouseover method (that is javascript) to display the tip and onmouseover doesn't work on straight text. You need an image or a link to hook the method on. So we had to convert some text fields in a table into links. Of course the links don't link to anywhere and that won't do so we added targets as well (the "progress boys" -Sasha Bol - didn't put targets in so clicking on a dummy link jumps to the top or bottom of the page). We added the target to the same row in the table as the link so that the page doesn't jump up and down when you click.
Ronny tested that and noticed that when he clicked a link he got an hourglass cursor. Until he clicked a real link. He is using MSIE 4.0 with SP1. I tried it with NS 4.7 and MSIE 4.0 and I didn't get the hourglass cursor. So technically the solution is ready.
While testing on my machine I noticed that opening an existing order reacted as if I had just made a new order. All the heading data was missing.
That was not the problem we set out to solve but it was most definitive a problem that needed solving. So we did (by changing a trim command by a trim-right command). It took us the best part of six hours and the same problem possibly occurs in other places as well.

And then I spent some hours un-knotting the conditions program for Norway. I know a lot more now but I still didn't locate the exact spot where things go wrong. The big problem is that the program contains a lot of do-while loops that are constructed from goto statements. And goto's are the best way to make sure your spaghetti doesn't stay straight.

I didn't even have time to read my internal mail. Theo reminded me about one of the mails that needed an answer. The problem in this case was with code-pages. In Italy the # (alt-035) gets replaced by a £ (alt-163) character (for example). Of course the # is used in HTML and XML in defining special characters so the Italian item descriptions came out with unexpected formatting. Though nut to crack because the ( IBM supplied ) FTP program on the AS/400 does this conversion completely by itself.
This need more work (and discussions because other applications suffer the same maltreatment).

---

Why is everybody happy that the days are lengthening nicely?
I will get happy when they really lengthen beyond that meagre 24 hours!

 

Top <  Mon | Tue | Wed | Thu | Fri | Sat | Sun  > Bottom

Again a mixed day with lots of cloud and rain. The rain never lasted long and never was heavy. It was never far off either. And there were sunny spells as well, never far off either.
Current temp is 17.4C in and 8.0C out.

Sorry, see tomorrow. My eyes are protesting and I have to increase the voltage to my brain, over clocking it, just to get it revving beyond sleep level.
Won't do, so I will cut it short.

---

 

Top <  Mon | Tue | Wed | Thu | Fri | Sat | Sun  > Bottom

Mostly dry but cloudy in the early morning. After 10:00 we got rain in bursts of an hour interspersed by outbreaks of dryness lasting an hour as well.
Current temp is 17.2C in and 10.4C out.

What did I do yesterday?
I dug in and stared my eyes red on the Conditions maintenance program.
And I found the problem, solved it too.
The bug survived since 1996 without, seemingly, ever tripping someone up. The program got into a loop is a record in the item file was NOT found at some point, and stayed in the loop as long as the item was NOT found. Obviously both 'not's should be dropped. This means of course that the program never did what was expected from it, and nobody ever noticed. Of course this also means that the Norwegians have some contracts without items allocated to them which is probably a data error.
Now I got to explain of course that I searched about two days before finding such an 'obvious' error, a "do-while *on" in stead of a "do-while *off" .

 

My brother gave his 'old' (Celeron 400, 64Mb) PC to Piet (hid brother in law). He got an even older defect PC in return. A 386/25.
I plugged the PSU in a spare motherboard that I had lying around (from one of my firewall attempts), plugged the old display adapter (ISA) in and connected up the old power supply. And I switched it on. And I got a puff of real smoke (smelling like burning plastic) off the display adapter!!! The amazing thing was that, when the smoke had cleared, the adapter worked normal, I got a good BIOS display on the screen so the display adapter and the power supply (and my motherboard) worked. Of course without drives or keyboard attached I didn't get far. Far enough though to prove the PSU is all right so it is the old motherboard that somehow got cooked.
I disassembled that box further.

It is a DTK box and motherboard (PEM-2530) from 1988 in a reversed tower (drives at the bottom, expansion cards at the top) case. It got memory in four SIP modules, totalling 1MB (I think, it has got 9 Panasonic chips per module with NN41C1000ASJ-07 markings) and an empty custom memory expansion slot in-line with an 8bit ISA slot. And loads of small surface mount chips. A lot socketed as well. And of course a true Intel i386 chip in a definitively not ZIF socket (marked with "A80386DX-25 IV, SX218, L1141800, 1985" on top and "90515671EK, BZ 08" on the bottom).
Video was provided by a nameless Tridend based 16bit ISA card (trident TVGA8900B) with 256KB of memory. While the drives and parallel and serial ports were done by a typical multi-I/O card (TDK pti-217).
Someone worked on that PC because it's got two hard drives. Both Quantum Prodrive LPS, one a 52MB thing an the other a 270MB that got jumpered as slave and most likely added some time after the sale.

Hey, I love archaeology.

---

Archaeologist and palaeontologist learn most from the things discarded long ago.
The current 'green ness' obsessive 'recycling' mania is depriving future archaeologist of their most valuable data.

 

Top <  Mon | Tue | Wed | Thu | Fri | Sat | Sun  > Bottom

Way too warm (for the season) all day. And way to wet (for my taste) with showers, some serious, distributed throughout the day. No sun whatsoever.
Current temp is 21.4C in and 11.3C out.

Running. To warm basically and dry, but there is a breeze cooling things down sufficiently. Just like last week my right leg, ankle and calve, remains stiff and a bit painful the first few kilometres.

In between the showers I do some pruning until the green box is filled. I'll continue next week.

And I swap components around for my brothers new PC. I want to put a CD-RW in it for backup purposes but it doesn't fit the case so I decided to use another case (as I explained last Sunday ).
Both systems have the same motherboard so I swap the processors (AMD Duron 700 and 850) around, struggling with these ridiculous large heat sinks, and I swap the memory around as well (128 and 256 DIMMs). I also move the 52x CD ROM drive out and the Plextor CR-RW drive in. After booting I find out that the reset button was wired up wrong. So back open the box to fix that.

And then I struggle to install a bunch of software. Oh dear!
This is a box for Peter but it will also be used by Martine so it needs Windows in Dutch. So I start to install Win98 and have to stop halfway trough because I don't have the codes here, they are at my brothers place.
Then I attempt to install Windows ME but that complains that it isn't an upgrade version and simply refuses to install on a system that got a partial and non functional Win98 install on it. So I start loading Linux because that has the only partition manager I have handy at the moment. I shrink the windows partition and add a second partition. I make the mistake of making that 20GB in size and telling Linux that it must be FAT16 type.
Later I install Win98 (English version) on the first partition and I get all sorts of problems with the second partition until I remember I made it FAT16. Which of course doesn't support the size. Well, in the end I get everything installed.
Next weekend I will reinstall everything again. Or rather I wil let Peter do it so he knows the procedure.

---