|
Microsoft's error messages have always been a pain. I suppose the best solution is to grit my teeth and bear it.
EDIT: (Nice to have a place to vent spleen!)
|
|
|
|
|
Jim_Snyder wrote: Microsoft's error messages have always been a pain.
Not exactly always - .NET handles exceptions quite nicely.
The COM error system, on the other hand, was a very bad joke made in the poorest of taste. A horrible, horrible, horrible thing! Quite possibly, the single worst error handling mechanism ever inflicted on mankind.
And it's still out their lurking in little (and sometimes not so little) corners of Windows - every time I see that dread string "HRESULT", my heart sinks because I see a trainload of grief approaching.
98.4% of statistics are made up on the spot.
|
|
|
|
|
That's all your (CPians) fault. I'm sure some of those QA questions were from Microsoft. If you (CPians) had helped them more with their Need Code questions, maybe Microsoft would improve.
|
|
|
|
|
|
Are you kidding? We're not doing that again - last time we got Windows 8!
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
|
Hey man, I'm using Windows 8 (yes, it's possible) right now!
|
|
|
|
|
I'm so terribly sorry. What did you to get that punishment?
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
Well, Stackoverflow did Vista...
|
|
|
|
|
That would explain a lot...
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
|
Gary Wheeler wrote: the Device Driver group's number one slot on my list of developers-first-against-the-wall-when-the-revolution-comes
What did we ever do to deserve that!
|
|
|
|
|
Have you ever written an installer for a device driver? The WDK docs describe three different, mutually exclusive ways to do it. The correct way is the fourth one. The whole thing is an abomination before the deity-of-your-choice.
Software Zen: delete this;
|
|
|
|
|
Well, they are pretty simple, and installation is supported by the win32 api, or there is divx that wraps that, or just an inf. They are simple these days, NT4, now that WAS a difficult inf to write.
Compared to using the kernel API it is all a walk in the park! 
modified 22-Nov-17 12:50pm.
|
|
|
|
|
Our history with Windows device driver development was sufficiently traumatic that all of our current custom hardware is built as single-board embedded computers that talk to our Windows-based controller via Ethernet TCP/IP. No Windows drivers, no awkward I/O constructs, vastly fewer undocumented rules that must not be broken.
Software Zen: delete this;
|
|
|
|
|
Or you can hire an expert to do it. Windows kernel work is viciously complex. IMO it takes many many years of practice before ones code is fit for use in the real world, there are so many ways of screwing it up.
However, after 20 years of it, I have probably found most of them, and now it is easy.
Of course Microsoft introduced the WDF model, a wrapper that takes all the PnP and Power handling on itself, and simplifies the IO stack. I have only briefly used it myself, but it is effective, not sure though if you can do all the down and dirty stuff you sometimes have to to fit a customers requirements.
|
|
|
|
|
I think you need anger management and good reading/typing skills. Why blaming Microsoft for your typo?
In Word you can only store 2 bytes. That is why I use Writer.
|
|
|
|
|
digimanus wrote: I think you need anger management
Possibly.digimanus wrote: I think you need ... good reading/typing skills
My reading and typing skills are fine.digimanus wrote: Why blaming Microsoft
I fault Microsoft for the necessity of typing command lines that are hundreds of character long.
I also fault their oh-so-detailed error messages that are effectively meaningless because of the barrage of worthless information that fails to address simple use cases. Don't give me random numeric error codes, and don't display HRESULT's. Stupid morons don't even use their own mechanisms for turning such things into meaningful text.
This is a toolset that's handed out to thousands of OEM's. If Microsoft is so all-fired worried about people doing this correctly, after 30 years there has got to be a better way than a mish-mash of command line garbage.
Software Zen: delete this;
|
|
|
|
|
and people say *nix is bad with all it's command lines to get things done
- at least with *nix you get a decent shell (of your own choice) and easier to edit scripts via an editor (of your choice).
Installing Signature...
Do not switch off your computer.
|
|
|
|
|
|
Quote: and random HRESULT's
 Hard to believe, funny anyway.
|
|
|
|
|
Gary Wheeler wrote: The Deployment bastards are working hard at taking the Device Driver group's number one slot on my list of developers-first-against-the-wall-when-the-revolution-comes.
As a developer, this both amuses and terrifies me.
I'm not part of either group, I swear... 
|
|
|
|
|
The problem is that Microsoft didn't hire enough DEC programmers to work on NT. 
|
|
|
|
|
Ah. I sense a story from the long ago that I've not heard before. I knew that Microsoft had poached a number of DEC folks to work on NT, but I didn't know it was that big of a thing.
Software Zen: delete this;
|
|
|
|
|
I wonder what OS/2 might have been like if it had been Microsoft and DEC rather than IBM.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
