|
anyone come acrosss a problem where they cannot increase the size of the virtual memory. I can set it to any number but the one displayed is still 20 megs but the odd thing is if I check the actual swap file, I get a size of whatever I allocated it..
Every time I reboot, I get the same increase virtual memory bla bla bla
thks,
eddie
w2k server
|
|
|
|
|
Hello all,
I need to perform some operation on some old 16 bit application (compiled under borland). Operation requires some text drawing (custom text control) and there is a problem. I use script which calls about 5000 such operations. After circa 500th there is a GDI problem. The result is that I receive dialog box "GDI cannot execute operation" and content of every window is not painted correctly (looks like global GDI failure). I tried to peridically closing this app, but I didn't notice any improve.
I wonder is there any solution of this? Maybe I should try to change OS to Win9x, which has different old Win app emulation? Or maybe there is some app, which 'cleans' unused GDI resources? Any anser will be appreciate.
I use Win2000, 128MB RAM, 32MB Video.
Regards
Bartek
|
|
|
|
|
If don't know if this is possible to do from a script. But, if you could kill the "NTVDM" process (Windows NT Virtual DOS Machine) process, the GDI resources used by 16-bit programs should be free'ed...
- Anders
Money talks, but all mine ever says is "Goodbye!"
|
|
|
|
|
Yes, thank you - I found this method useful yesterday. Instead of killing NTVDM process - I kill wowexec (all resources are freed, too). I wonder which method is more correct?
Regards
Bartek
|
|
|
|
|
GREETINGS
My internet doesn't work properly under Win2K. I can't view different sites(for example - CodeGuru or mail on Yahoo). Who knows how to solve this problem?
Modem - Creative Modem Blaster V.90 PCI DI5655
Thank you!
LOVE & PEACE to everyone
LOVE & PEACE to everyone
|
|
|
|
|
Anyone know off the top of your head what the max length of a domain name is defined as under NT? Thanks in advance. I'd rather not spend an hour searching MSDN.
|
|
|
|
|
Hi,
I think I asked this question before in "the lounge" a while ago but this forum is probably a better place for it.
Because we cannot tell which machine our customers our going to run our software on I got myself an "Me" machine. It all works fine except for when I start up the machine or release the communication port after having used it the serial ports on it go into a very strange state. In all other windows versions I have used (3.1,95,98,Nt 3.51, Nt4.0, Nt2000) releasing or not using a serial port just brings the RS232 port in rest condition (RTS,DTR and TX line go negative: between -3 and -12V) which is what is supposed to happen. An unused is "not used" and therefore inactive. The RS232 standard says the port signals should be at rest in that case. That is not what my "Me" machine does. it sets RTS and DTR (positive voltage) and puts the TX line in "break" mode continuously.
I now have the problem that I do not know whether the problem is a one off case or whether it is typical for "Me". Anyone got any ideas or has heard about that sort of thing ? At microsoft nobody has a clue and when it comes to giving specs about what their product is supposed to do it is equally hopeless (believe me, I have tried extensively)
Any ideas/help greatly appreciated.
Filip
|
|
|
|
|
Recently I encountered a weird problem: when running debug mode in Microsoft VC++ 6.0, passing the first breakpoint (or second press of "Step Into/F11") produced blue screen of death saying "A fatal exception occurred in ... VXD TRIOVXD..."…
I found that this behavior was independent of application being debugged, and after wasting some time trying to find something related in DevStudio settings - but everything looked fine - I started to think what have had changed in the system since the last time I've been working with debug. The only change that occurred to me is that I installed Matrox G450 DualHead card instead of old ATI 128 Rage. So I went back, removed Matrox, uninstalled drivers, put ATI, but the problem still persisted. What was suspicious though, that at some point of dealing with old/new card drivers (I do not remember what stage it was exactly) I got exactly the same blue screen message… So, it convinced me even further that the problem originated in graphic card, presumably Matrox, and then I did all kinds of things - downloaded updated drivers, uninstalled, reinstalled, physically removed and installed new/old card, etc. - with no result. Finally, I searched system files for any reference to TRIO, and found the line device=trio.386 in [386enh] block of system.ini file. I deleted it, partly expecting that the machine will not boot, but it did, and this fixed the problem, without any other noticeable effect…
It is still quite obscure to me where this line came from, should it be there at all or not, and what does it mean, so if anybody can shed light on this issue, I'll be very grateful.
Regards,
Igor Kagan
http://igoresha.virtalave.net
|
|
|
|
|
I keep running into this problem in the vc++6 editor where the copy and paste functions dont always work. Whether I use the edit menu, context menu, or ctrl-c/ctrl-v, sometimes it just wont copy and paste. Sometimes it does.
This only happens on one particular machine of mine. I've never had this happen before but its damn annoying.
Any ideas.
Josh
josh@that-guy.net
|
|
|
|
|
This is a known bug in NT. See MS KB article Q231652. Fix is in the following .reg file:
Windows Registry Editor Version 5.00
[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Image File Execution Options\msdev.exe]
"DisableHeapLookaside"="1"
|
|
|
|
|
Yup it's win2k. That fixed it.
Thanks
Josh
josh@that-guy.net
|
|
|
|
|
Hi
I would like to disable the Ctrl+Alt+Del key combination on WinNT/2000. Am able to disable all other system key but this!
Please show me the way.
TIA
Anil.
|
|
|
|
|
You can't block Ctrl+Alt+Del on NT/2K. Check the KB article Q125614 for more info. BTW - why do you need this feature?
Tomasz Sowinski -- http://www.shooltz.com
|
|
|
|
|
I need this feature for a kiosk style application. I understand the security implications and also that a new keyboard filter or a replacement GINA dll are the only ways to do this. But just looking for info as there is a paucity of time. Of course, if these are the only ways to do it, then so be it.
Thanks
Anil
|
|
|
|
|
Actually, you CAN disable those keys, but it involves terminating winlogon. I didn't notice any drastic effects, the most notable thing was the lack of ctrl+alt+delete (at that point I could use RegisterHotKey and catch it in my own apps) and the inability to shut down the system. (You can't shut down by ANY MEANS. Blue screen or reset buttons are your only chioces.) Here's what you do:
Terminate smss.exe. (You'll need SeDebugPrivelege)
Terminate winlogon.exe.
There you go. I'm told that all smss.exe does after the system is running is wait on various things and bluescreen when they terminate. So by killing it before winlogon you get rid of winlogon without the bluescreen normally associated with it. If you wanted to get really advanced you could try to figure out which thread in smss.exe was the culprit and just kill the one thread.
Use at your own risk! I myself don't fully understand all of winlogon.exe's functions.
Stupid name? Yeah. I know.
|
|
|
|
|
Create a new user on the computer with limited security.
from your administrative account, remove read access from taskmgr.exe.
This will prevent the task manager from coming up.
There are ways of securing the rest of the 2K C+A+D screen like lock computer with the policy editor.
I also have a kiosk program running on XP now, and a locked down user account is the best option.
XP is nice because I can use terminal service into the admin account, and let the system auto-logon to the limited account.
Telnet and a collection of WSH scripts is also very VERY useful for when you dont want to inturupt the main screen, which terminal services does.
|
|
|
|
|
Does anyone know what Windows function/event I can use to trigger an action on a mouse rollover?
|
|
|
|
|
Rollover on what?
If the target is (derived from) a window, try the WM_MOUSEMOVE message. If the mouse is not captured, the WM_MOUSEMOVE message is received by the CWnd object beneath the mouse cursor.
|
|
|
|
|
Hi everyone,
Does anyone have *any* idea how the ApplicationGoo registry key parameter works on Win2K? In fact, does anyone have a list of what parameters are available under the keys within Image File Execution Options?
For those that know even less than me about this...
Under the key:
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\WindowsNT\CurrentVersion\Image File Execution Options
there are a Image File Name keys such as; qpw.EXE, cqw32.exe, install.exe, wpwin8.EXE, setup.exe (some of the vendor specific ones are obviously recognisable). Under each name key there are parameters such as DisableHeapLookAside, GlobalFlag (these seem to be booleam values as REG_SZ) and then there's the mysterious REG_BINARY ApplicationGoo...
I'm guessing its something to do with Win2K compatibility i.e. when Win2K starts executing an image file it checks this and gives appropriate warnings, but I could be utterly wrong.
I've hunted the Internet & MSDN for a quite a while but can't find anything of substance. Any ideas out there? Am I just being particularly dense? Someone give me some pointers?
Anyway, kudos to the first on the buzzer with the right answer.
TIA,
Trevor
|
|
|
|
|
ApplicationGoo looks like version resource stored in registry, at least on my machine. Have no idea what they're using it for. Maybe for comparing currently executing image with the one executed before?
Tomasz Sowinski -- http://www.shooltz.com
|
|
|
|
|
We have a client/Server application which, on the server side, runs on a dedicated, stand alone, server. We have a logon procedure by which we force users to authenticate themselves to us when they logon with a username and password. We would like to avoid this for those users who access our service if they are already properly logged onto their local network.
For example, what we want is to have some means of identifying a user trying to log onto our system as a properly authenticated member of domain XYZ of their network server. Is there any means of acquiring network information from the client side, which can be passed to our server which would allow us validate a user automatically? Clients might be running under Win 95/98.
Also, is there any way to uniquely identify a given network domain? That is, can we distinquish domain XYZ of one network, from domain XYZ of another network using the standard Windows networking API?
|
|
|
|
|
If the client use Named Pipes to talk to the server, you can do it pretty easy.
Named Pipes supports NT Security
If you want to know more about Named Pipes and NT Security, I recommend the book: "Programming Server Side Applications for Windows 2000", from MS Press...
- Anders
Money talks, but all mine ever says is "Goodbye!"
|
|
|
|
|
Thank you. I will look into that. However, one quick question, doesn't the use of a named pipe imply that you have a service running on the server in question in order to establish the converstation between client and server? In our situation, our service is running on a server external to the primary network that a client might be on. We simply want to be able to basically "piggy back" off of their existing network security so the user does not have to log on twice, once to get onto their network, and once to get into our service. No solution that requires us to have something running on thier server is going to work. We need to have information from the client relative to its status on the network so that we know that when the client calls us, it can tell us if it is, in fact, properly logged on to the its network.
|
|
|
|
|
Ohhh, now I know what you want, and you are right, you cannot use Named Pipes for that.
I guess I would get my client app. to chech that the user is authentificated against a DC, and send a special packet to the server, when connecting, saying "I'm ok
- Anders
Money talks, but all mine ever says is "Goodbye!"
|
|
|
|
|
Thanks. I think that is what we have decided to do. Unfortunantly, to do that authentication from 95/98 does not appear to be straight forward, and might involve the use of the old 16 bit network api.
|
|
|
|
|