Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - ddan

Pages: 1 ... 3 4 [5] 6 7 ... 12
61
General Discussion / Re: A curious case of case.
« on: 2008.10.09, 19:49:22 »
Yes, I LIKE case-insensitive on the command line, wouldn't have it otherwise.

Let me clarify: the BUG is that text as I happened to type is passed to the Rexx function, rather than the ACTUAL directory name. I ought to be able to rely on a function call getting the ACTUAL text off the drive. That this isn't so is the mystery. -- I was in fact relying on the program call RATHER than any typing, and in the application it's definitely an annoying anomaly.

Fortunately, except for the one case where it locates the highlight one place off if I start in an "os2", "Os2", or "oS2" (but not the "OS2") directory, will probably be tolerable, so doubt I'll ever do the rather extensive uppercasing.

62
If after two weeks, you've gotten nowhere, maybe it's time to re-install? It's not that big of a hassle, and by now, would have saved many hours. Also, I doubt that figuring out the actual problem here will ever be of value, as it's become clear is a unique crash.

63
General Discussion / Re: A curious case of case.
« on: 2008.10.09, 05:02:52 »
No, demonstrating the effect doesn't SOLVE any mystery. When you logged into a directory, it should show the actual text of the directory name as stored on the drive, not as you HAPPEN to have typed it in.

And, the problem is the effect on my program, which needs to match the directory name while preserving case for readibility, and it will likely be started IN a directory that I've used only lower case to get to. Rexx directory(), then, gets a non-actual name of the directory! Looks as though to dodge the buggish effect I'd have to kludge in an upcase of everything to compare, which ought not be necessary, wouldn't be if the actual text off the drive were used for the path.

64
General Discussion / A curious case of case.
« on: 2008.10.08, 18:24:49 »
In looking at an apparent bug in my file selector, I found a strange difference in case:

OS/2 Wed 10-08-2008 10:59:16.38 C:\>dir os*

The volume label in drive C is OS2.
The Volume Serial Number is A9A7:D414.
Directory of C:\

 4-15-08   1:07p     <DIR>       11177  OS2      <<< NOTE DIRECTORY NAME IS UPPER CASE
 4-15-08   1:37p     <DIR>           0  os2tk45
        2 file(s)          0 bytes used
                   3761091 K bytes free

OS/2 Wed 10-08-2008 10:59:20.21 C:\>cd os2

OS/2 Wed 10-08-2008 10:59:22.68 C:\os2>       <<< AND NOW IT'S LOWER CASE!

Can anyone explain HOW this can even be?

It SEEMS to apply only to the \OS2 directory; haven't found any others with case different depending on where you look at it from. My thought is that it's somehow related to the volume name (OS2).

65
Setup & Installation / Re: Dual Xeons bring no joy.
« on: 2008.10.08, 02:09:50 »
Whew. The RAID controller seems to be flaky too; won't reliably recognize even the two drives which I KNOW to be good because have installed on them, though will assume that the two which flash continually and the one which has a steady green are bad. For now, I refuse to believe that I don't know how to plug drives into an SCA rack and have them show up on the scan list, but that's the symptoms I get. Tried all combinations of racks and channels with same results, usually only one of up to 3 drives is listed, whether re-scan is started from software, hard reset, or powered down for a minute. But the on-board LSI controller recognizes and spins up the 3 drives, and WSEB on one still boots.

Anyhoo, title of this thread is still apt. I'm sure you're all breathlessly awaiting my next report, but at the moment, I'm a little low on humor.

66
Setup & Installation / Re: Dual Xeons bring no joy.
« on: 2008.10.07, 21:11:48 »
Still ambiguated. Turning off hyperthreading in BIOS leaves me with 2 processors, only changes what BIOS reports just before going to the HD. "/P=4" on the OS2APIC.PSD line didn't help, either. Since I don't have anything later, guess I'm stymied for now, will just console myself that I have twice as many CPUs over there as here, and each is probably twice as fast as this 1.6G core AMD, too.

Haven't tried a new kernel either (have the latest SMP 104 around somewhere), but will wait while I dither over whether to use the Adaptec 3000s controller, that I assume is RAID, but doesn't seem to appear on their site.

67
Setup & Installation / Re: Dual Xeons bring no joy.
« on: 2008.10.07, 06:11:32 »
Hmm. No change with "/SMP" or "/APIC" switches.

This from The Google: "you get thread-level parallelism on each processor" implies that it doesn't count as double the number of CPUs.

But the Wiki has "A processor with Hyper-Threading enabled is treated by the operating system as two processors instead of one."

Then almost contradicts it:
"Hyper-threading relies on support in the operating system as well as the CPU. Conventional multiprocessor support is not enough to take advantage of hyper-threading."

So still don't know whether WSEB is counting "all" processors; the 30% increase isn't in line with OS/2's direct scaling, given that specialized apps are implied.

Meanwhile, much else to do. -- By the by, original problem seems due to TWO of four memory sticks going bad at once, which I just couldn't believe, and somehow didn't hit on the right combination in swapping. Otherwise, an exceptionally nice box, with good control of fan speed by the BIOS, lowers their speed soon after power-on.

68
Setup & Installation / Re: Dual Xeons bring no joy.
« on: 2008.10.07, 02:04:45 »
AHAH. Refer to philosophical point of previous. Apparently I turned APIC off in BIOS while flailing around.

But have only TWICE the joy of before, so my question ends up as: does HyperThreading count as processors? That's what I thought, so I feel cheated! (Even though I got this box for free!) BIOS screen says 4 CPUs; OS/2 reports only 2 processors initialized, as CPU monitor shows also.

Can anyone field this? (Haven't found anything explicit on The Google, murky area...)

PS: I'm not confused between ACPI and APIC, just found the latter disabled going through BIOS again.

69
Setup & Installation / Re: Dual Xeons bring no joy.
« on: 2008.10.07, 01:45:00 »
Thanks for the replies. Problem SEEMED to be memory, though I tested it after PCLinux installed, and same symptom. SO then I went with the Client version, which WORKED.
<LONG pause to meditate.> Then, tried WSEB again, and went fine, is running.

By the by, it's SCSI HD (320s!), but using the internal LSI controller (160) for now. So the Dani drivers are no help.

BUT still have only one-fourth joy! Now it doesn't want to install SMP. Tells me to point it to path with "manufacturer specific" .PSD, and I have none (as yet). The Google found me a probably similar Primergy F260 (mine is F250) that IBM CERTIFIED to work with OS2APIC.PSD, which of course I have.

Tried setting Multiprocessing to 1.1 in BIOS. Same. Tried turning off 2nd CPU and HyperThreading. Same. So I'm going to look into this ACPI thing, thanks Saijin -- or at least I HOPE...

<philosophical aside> Some problems can't be fixed until you've yelled for help. Perfect example here: whatever memory problems there were seemed to just disappear the second I posted the first here.

70
Setup & Installation / Dual Xeons bring no joy.
« on: 2008.10.06, 21:12:55 »
I'm attempting to install OS/2 on a Fujitsu-Siemens server with Dual 2.4G Hyperthread Xeons.
Had some problems with memory, which MAY still be present, though PCLinuxOS installs and runs just fine (15 minutes from CD boot to HD boot!), and MemTest86 didn't find any errors.

Problem with OS/2 isn't subtle: CP2 CD doesn't even get to the boot blob! W4 from floppies gets only to 2nd diskette, then a trap D.

I've tried all standard BIOS conflict items and specific to Xeons: another CD drive, turning off everything non-necessary in BIOS, including Hyperthreading and disabling 2nd CPU. Nothing affects symptom (with CP2), just doesn't get anywhere. I'm looking for more options such as turning off CPU cache, but may not be able to.

<a bit later> Hmm. Is there some difference between CP2 client and server? Because the client gets all the way to normal partitioning stage (haven't gone further because went to get another drive to experiment on...).


From the Linux tests, this system looks really fast, isn't too noisy, and doesn't even run very hot, so I'm hoping to use it for my main system with OS/2.

I can fairly easily clone a drive if anyone thinks that's the route.

71
Jep, have to nitpick on this one: it bails out on me at
     5 +++   Call RxFuncAdd 'SysGetKey', 'RexxUtil';
REX0040: Error 40 running C:\temp\inputkey.cmd, line 5: Incorrect call to
routine

I already have the whole rexxutil.dll -- it's only 68K -- loaded from a
program that runs on boot-up (and of course once loaded, it stays resident
unless one goes to some trouble to drop its functions), so think that specific
function loading is best skipped, even if your purpose is education as here.
-- On the other hand, it's easy to forget that EVERY program needs to specify
loading the library because one can't be sure that it IS already loaded. REXX
just skips over if it is, so I start a program by pasting in the WHOLE .dll
loading.

No problems with the rest. Tested it because wanted to be sure that SysGetKey
includes effective sleeping while waiting for a key; however, I usually need a
loop around keyboard polling so am forced to test charin() and use syssleep.

72
Internet / Re: MySpace.com updated Flash version check
« on: 2008.09.30, 18:02:11 »
You can't be SURE that it IS Adobe checking version, nor that anything at all is "checked". May be simply trying to get you to allow software to be installed. If you're using OS/2, you may be blissfully unaware of how common are attempts to exploit the gaping holes that Windows leaves open. Every time I see someone allowing all web "features", I'm amazed that they can stand the non-stop bombardment of flashing ads, redirection, and exploits.

73
Setup & Installation / Re: Only SATA, any options for DANI?
« on: 2008.09.27, 16:00:21 »
Well, I don't know of any reason TO turn IDE off in BIOS, and interested to know whether the problem is present when it's enabled but without any IDE drives connected, or is nailed down to just being turned off. Don't see any major implications if the latter, but needing an actual IDE drive could be annoying.

Also, have any benchmarks for the SATA HD?

74
Internet / Clickjacking!
« on: 2008.09.27, 03:47:53 »
The newest FEATURE EXPLOIT: Clickjacking! Seems to be a hidden frame, enabled
by Flash:
http://blogs.zdnet.com/security/?p=1972

But this blog seems to have the most info:
── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ──
http://www.webadminblog.com/index.php/2008/09/24/new-0day-browser-exploit-clickjacking-owasp-appsec-nyc-2008/

zmjjmz  |  September 26th, 20088 at 2:28 pm
Ok guys, so, supposedly the NoScript dude has a fix to this. (Assuming you
have FF and NoScript installed) You would go to Tools, Add-ons, Extensions,
NoScript, Preferences, and then the Plugins tab.
Check off "Forbid " and (according to the NoScript maintainer) you should be
100% protected.
── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ──

Oh, rats! I'd already done that! Turns out that my policy of forbidding
plug-ins and avoiding "features", especially Flash which is said to have the
flaw here, made me safe from attacks that hadn't been invented.

Darn! Now I don't get that warm and fuzzy glow as Windows types get when after
fixing one flaw, they await the next to be discovered. (To be fair, even with Windows,
MOST security problems are due to UNNECESSARY FEATURES.)

75
I now find my reply to 'What is the point of OS/2ers bashing "Windows" when
from all appearances "Windows" is now comfortable hugging "Linux" rather that
it's "Half Brother" IBM's OS/2.' to be too glib.

That's actually a deep philosophical and even religious question which of
course has been answered many times. Rephrased, it's: What is the point of
resisting evil? (Especially when it's SO easy to go along with it as most
people do, and they visibly profit from it, too.)

Well, IF you recognize an evil, then you're pretty much obligated to resist
it, or you're complicit. Yielding to it as a practical matter should be done
only as necessary to continue the struggle. But as Crimosoft itself sneers, if
they're not stopped at SOME point, then THEY are NOT stopping short of
complete monopoly. ── You should learn from history of how hard CHOICE has
been fought for, how easily it's lost. ── Even if the struggle were futile ──
it's not ── then to be at all moral, you MUST protest evil when you see it.

I assume that no one here endorses Crimosoft's business practices, regardless
what is thought of the technical merits ── and "ease" ── of their products.

Crimosoft has a stated goal of eliminating ALL competition. Just download:
http://www.groklaw.net/pdf/Comes-3096.pdf
Everyone should read it. (You can skip most of the last half, duplicative at
best.)

You can't hope that being friendly to a ravening beast will even cause it to
eat you last. Just try to imagine life IF they achieve their stated goal.

Pages: 1 ... 3 4 [5] 6 7 ... 12