Prev: Mount and dismount physicaldriver
Next: Can one force a USB disconnect then reconnect?
From: Pete Gontier on 4 May 2008 19:35
On 5/3/08 6:28 AM, in article
dedf9e87-a439-4534-9972-a2f1070c83c0(a)l64g2000hse.googlegroups.com,
"chris.aseltine(a)gmail.com" <chris.aseltine(a)gmail.com> wrote:

> ... you might look at this link to learn how to force-install a driver on a
> given device node, without touching anybody else:
>
> http://support.microsoft.com/kb/889763

Thanks. That seems to work well enough for me to send in another build to my
tester in hopes of ending this problem. Four notes for posterity:

-- The DDK contradicts the Knowledge Base article about the last
parameter. I trusted the KB article.

-- The sample code's declaration of the function pointer should
use APIENTRY.

-- InstallSelectedDriver has a file copying step unnecessary for
my usage, where UpdateDriverforPlugAndPlayDevices does not.
While the WHQL warning is posted, the progress window for this
copying step is sitting behind the warning, not painting, and
looking like a bug. Nice.

-- If you pass false for the backup parameter (probably not wise,
but I experimented with it to see if I could inhibit the copying
step) and the user says STOP at the WHQL warning, what appears
to be the last phase of the Found New Hardware Wizard appears
and reports the same error which will eventually be returned
by GetLastError after InstallSelectedDriver returns false.



Pete Gontier
http://www.m-audio.com/
pete (at) m-audio (dot) com

From: chris.aseltine on 4 May 2008 23:26
On May 4, 6:35 pm, Pete Gontier
<kokoroza...(a)discussions.microsoft.com> wrote:

> -- The DDK contradicts the Knowledge Base article about the last
> parameter. I trusted the KB article.

By the way, the article is also wrong about the "LPCWSTR Reserved"
argument. It says it should be null, but you really need to pass it a
pointer to a wchar_t buffer. Otherwise I think it crashes, but only
on Win2k, or only on XP, or something like that.
From: Pete Gontier on 5 May 2008 12:23
On 5/4/08 8:26 PM, in article
44ca0226-1a8b-408b-a846-28b12f3db37b(a)b64g2000hsa.googlegroups.com,
"chris.aseltine(a)gmail.com" <chris.aseltine(a)gmail.com> wrote:

> By the way, the (KB) article is also wrong about the "LPCWSTR Reserved"
> argument. It says it should be null, but you really need to pass it a
> pointer to a wchar_t buffer. Otherwise I think it crashes, but only
> on Win2k, or only on XP, or something like that.

Yes, I saw that somewhere on the web. The DDK documentation also says this
parameter is reserved and must be NULL, so it would seem both docs are at
least somewhat incorrect. However, I didn't have any problems on XP SP2, and
we don't support anything less than that, so I don't have to guess how big
to make a dummy buffer for an undocumented value. Thanks again...


Pete Gontier
http://www.m-audio.com/
pete (at) m-audio (dot) com

First  |  Prev  | 
Pages: 1 2
Prev: Mount and dismount physicaldriver
Next: Can one force a USB disconnect then reconnect?