blabberer wrote:if you scrap virtual kd you will be sitting there for hours waiting for windbg to return back from a single step with registers
I'm saying this with humour, so you wont lose it, but you are in fact admitting that vmware is a slow piece of crap.
Only Microsoft could have devised a debugger like this but we are stuck with it since Numega decided to dissolve. So, let's make the best of it, right?
blabberer wrote:pipe debugging turns of register dumping by default and will still be too slow for a saints patience.
I was reading up on pipes just to refresh rusted parts of my brain. They are implemented as FIFO buffers in the software rendition of serial communication. In normal serial hardware, FIFO buffers are used to store serial bits before sending. Interrupts are used by peripherals like serial ports to tell the CPU when service is required. The CPU controls the data flow using an RTS line, which it turns negative to tell the comm equipment to take a break.
VMs implement pipes as software buffers. So windbg is communicating with a buffer set up by the VM. Or. if the VM wants to set it's serial0 (com1) port to use the com1 port on the hosts hardware, it can do that. Unfortunately, I think windbg requires a null modem connection which I don't think a VM can supply, hence the use of pipes.
Stop me anytime you disagree.
blabberer wrote:instead of cribbing about wife however obtuse she is try to pretend to live with her atleast your dreams will be sweet.
many wives specialize in cribbing so why should I not crib when said wife does not perform in an ideal manner? After all, everything SHOULD be perfect, should it not?
blabberer wrote:i wrote ignore spurious breaks (they are not spurious perse but spurious for your context)
trash what you are doing sit back relaxed with a coffee unlearn all the shit you have learnt till now and start from scratch.
Ah...so you've done this before?
blabberer wrote:IopCreatefile is a processor it can process only one request at any given time
yes...I am aware of the time-slice processing of computers. I am aware that IopCreateFile has other things to do than wait for me. However, in SI, I could figure other ways around that whereas in this new learning curve I am somewhat handcuffed, especially when the mouse freezes.
blabberer wrote:you need to be in the context / context / context / context of what you doing for you to transact.
I have an intimate relationship with contexts from having plied my trade with SI. SI dealt with it differently. If something else had priority, SI would just break, allowing you access to the command line. Sometimes you could just hit go till it broke in the code you wanted. Other times there were too many intervening breaks.
The windbg/vm arrangement causes a conflict with the mouse and no one on the Net seems to have an answer for that.
blabberer wrote:if you leave the break on IopCreatfile as it is and merrily step away you will break thousands of time when you press p (single step)
familiar with that from SI. I was religious about disabling the breakpoint that got me there, especially if it was a general breakpoint that could break on system activity. Many a time I have sworn at myself after hitting go, realizing I had forgotten to disable a break, getting stuck in the middle of foreign code. Sometimes i could step out of it but just as often that triggered the app.
You are preaching to the converted. I am from the Kayaker school of contexts.