ekohl(a)svn.reactos.com wrote:
>Add architecture-specific headers and move I386-specific definitions into them.
>
>
>
Thanks a lot Eric.
I did some research on another FIXME that is actually related to your
PnP work, which you might want to take a look at. You will notice that
DEVICE_NODE has 2 non-NT members. I saw that you had some previous
patches concerning this structure that removed some of the ROS-Only
fields, and I guess you might want to get rid of the two ones remaning.
The first is BusTypeGuid. The correct way it should be handled is
through a private array of GUIDs, which is indexed according
DeviceNode.ChildBusTypeIndex.
When querying, you just return the GUID in
IopXxxxx[DeviceNode.ChildBusTypeIndex], and when creating a bus type
guid, you just write the index there. I hope that's clear enough.
The second is "Address". That field is actually already present in the
DDK-documented DEVICE_CAPABILITIES structure. Therefore, when queried
for it, simply send an IRP_MJ_QUERY_CAPABILITIES (or whatever the name
is) to DeviceNode->PhysicalDeviceObject, and return the
DeviceCapabilities.Address.
That should be less then 20 lines of code to change, but I don't want to
break any PnP features since they aren't my speciality..
Best regards,
Alex Ionescu
Shouldn't that be 'of reg key'? ;)
On 7/4/05, gdalsnes(a)svn.reactos.com <gdalsnes(a)svn.reactos.com> wrote:
> fix premature close og reg key
>
>
> Updated files:
> trunk/reactos/lib/setupapi/devinst.c
>
> _______________________________________________
> Ros-svn mailing list
> Ros-svn(a)reactos.com
> http://reactos.com:8080/mailman/listinfo/ros-svn
>
--
"I had a handle on life, but then it broke"
ekohl(a)svn.reactos.com wrote:
>- Move definitions from ntpnp.h into NDK and DDK
>- Remove ntpnp.h
>
>
>Updated files:
>trunk/reactos/include/ndk/zwfuncs.h
>trunk/reactos/include/ndk/zwtypes.h
>trunk/reactos/lib/wdmguid/wdmguid.c
>trunk/reactos/ntoskrnl/include/ntoskrnl.h
>trunk/reactos/ntoskrnl/io/plugplay.c
>trunk/reactos/services/umpnpmgr/umpnpmgr.c
>trunk/reactos/w32api/include/ddk/wdmguid.h
>
>Deleted files:
>trunk/reactos/include/ddk/ntpnp.h
>
Hi,
Heh, you beat me to it! I asked Filip if it was OK to do this and was
going to do it myself, thanks for thinking foward. I just have two comments:
- I think the GUIDs are undocumented, while wdmguid.h is a DDK
header. Therefore, adding them there breaks MS DDK compatibility, since
we'll get undefined errors if we build with it. Sorry to nitpick about
compatibility again.
- To keep size small and to stick with the current way the NDK has
been made, I would prefer if there were no coments in it that describe
or document structures. The comments should go on top of the actual APIs
in the source files. Sorry for nitpicking again ;-).
Best regards,
Alex Ionescu
mf(a)svn.reactos.com wrote:
> prepare for WINE UNICODE patch
>
>
> Updated files:
> trunk/reactos/subsys/system/winefile/winefile.c
> trunk/reactos/subsys/system/winefile/winefile.h
>
> _______________________________________________
> Ros-svn mailing list
> Ros-svn(a)reactos.com
> http://reactos.com:8080/mailman/listinfo/ros-svn
>
This change broke compiling:
subsys\system\winefile\winefile.c: In function `path_from_pidlA':
subsys\system\winefile\winefile.c:706: warning: implicit declaration of
function
`free_strret'
subsys\system\winefile\winefile.c: At top level:
subsys\system\winefile\winefile.c:716: error: redefinition of 'wcscpyn'
subsys\system\winefile\winefile.c:671: error: previous definition of
'wcscpyn' w
as here
subsys\system\winefile\winefile.c:1096:2: #endif without #if
subsys\system\winefile\winefile.c: In function `CheckForFileInfo':
subsys\system\winefile\winefile.c:1821: warning: passing arg 1 of
`GetFileVersio
nInfoSizeW' discards qualifiers from pointer target type
subsys\system\winefile\winefile.c:1826: warning: passing arg 1 of
`GetFileVersio
nInfoW' discards qualifiers from pointer target type
subsys\system\winefile\winefile.c: At top level:
subsys\system\winefile\winefile.c:671: warning: 'wcscpyn' defined but
not used
subsys\system\winefile\winefile.c:698: warning: 'path_from_pidlA'
defined but no
t used
mingw32-make[1]: *** [obj-i386\subsys\system\winefile\winefile.o] Error 1
Even if hostlix guys wanted to help ReactOS team, they couldn't. From
ReactOS Development Policy:
8b) Developers who have had access to proprietary source code that
would have
originally fallen under trade secret protection (including, for
example,
the leaked Microsoft Windows source code) are a special case. The
copyright issues are unchanged, as are the patent issues, but any
developer who has the code is clearly either under an NDA or else does
not have a license to possess the code. In the latter case, there
is a
clear violation of the relevant trade secrets by the developer. As
such,
this kind of code is unsuitable even for research and
understanding. It
is the policy of the Project to not take submissions from
developers who
have or use any such proprietary code in their development efforts for
ReactOS.
All thouse hostlix guys have seen leaked windows source code and can not
contribute to ReactOS.
No changes made to hostlix code can be merged back to ReactOS.
I wrote all my ideas about the new ReactOS page (that came to my mind)
together:
http://reactosde.re.funpic.de/mediadetail.php?sec=Articles&cat=ideas&mdnr=1…
What is your opinion about that?
There are two (or more) possible solutions for the static main pages:
* wiki system that generate the static pages (HTML/XHTML) (maybe language
support problem ?)
* selfwritten simple CMS system (very flexible, easy to use, only a few
php-scripts so fewer bugs, better integration (design, layout), language
support, generate HTML/XHTML, ...)
(* CMS system like the current one)
What system do you prefer? (wiki, CMS, ...)
Compatibility and Package Manager section:
* wiki based (like in the current mediawiki only a long list)
* php based dynamic list (with filter features and tree structur; mysql
database): example: (first alpha page)
http://frik85.fr.funpic.de/packmgr/?page=tree&sec=view&sec2=2
* winehq's application database (or something else; redesign it?)
What system do you prefer? (wiki, dynamic list (tree structur), ...)
I will have time in july/august/sept. to implement the Compatibility-,
Package Manager section and the simple CMS system that can generate the
static pages, if it is okay for you.
But I will wait until I get a "start signal". If someone else want do it
alone or want help me, it will be no problem (for me; then I will have more
time to code for ReactOS).
I wrote this because the current ReactOS page is awful (design and CMS
system) and we need a better and more beautiful page to gain more visitors,
get a bigger ReactOS community (and more devs).
I hope the UI-Team will finalize the layout of the homepage in the next two
months.
Thanks for reading this, and maybe you find time to write your opinion about
that.
Best regards,
Klemens Friedl
(PS: Tomorrow, I will have no time to answer any emails, sorry.)
--
Weitersagen: GMX DSL-Flatrates mit Tempo-Garantie!
Ab 4,99 Euro/Monat: http://www.gmx.net/de/go/dsl
ion(a)svn.reactos.com wrote:
>PCH for csrss and win32csr, and remove rosrtl usage
>
>Modified: trunk/reactos/subsys/csrss/api/wapi.c
>
>
> *Modified: trunk/reactos/subsys/csrss/api/wapi.c*
>
>--- trunk/reactos/subsys/csrss/api/wapi.c 2005-07-01 02:47:15 UTC (rev 16360)
>+++ trunk/reactos/subsys/csrss/api/wapi.c 2005-07-01 03:03:06 UTC (rev 16361)
>@@ -10,13 +10,8 @@
>
>
> /* INCLUDES ******************************************************************/
>
>
>
>-#include <windows.h>
>-#define NTOS_MODE_USER
>-#include <ndk/ntndk.h>
>-#include <rosrtl/thread.h>
>
>
>+#include "csrss.h"
>
>
>
>
>
>-#include "api.h"
>-
>
>
> #define NDEBUG
> #include <debug.h>
>
>@@ -170,7 +165,7 @@
>
>
>
> /* Close the port and exit the thread */
> NtClose(ServerPort);
>
>
>- RtlRosExitUserThread(STATUS_SUCCESS);
>
>
>+ NtTerminateThread(NtCurrentThread(), STATUS_SUCCESS);
>
>
> }
>
>
>
Hi,
this change is wrong. If you remove RtlRosExitUserThread, you have to
insert the code from RtlRosExitUserThread. With the current code, the
user mode stack from the thread isn't freed. Csrss does crash after a
short time because csrss isn't able to create a stack for a new thread.
This occurs while compiling ros on ros.
- Hartmut
On Wed, Jun 29, 2005 at 01:17:11PM +0200, Rafal Kupiec
wrote:
> K McI wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> >
> > Rafal Kupiec wrote:
> >
> >>I think, now we're not violating the GPL.
> >>for eg:
>
>>http://svn.berlios.de/wsvn/hostilix/trunk/Sources/apps/notepad/dialog.c?op=…
> >
> >
> > I hope Revision 42 isn't the latest because the
> > URI is invalid (You lost a "w" somewhere from 1 to
> > 42)!
> > (...)
>
> What?
>
The site. The site that the above link points to.
Look at it.
* Copyright(C) 1999-2005r. BetaComp Team
<http://www.betacomp.info/>
becomes
* This file is a part of BetaComp Team Project
<http://ww.betacomp.info/>
Notice the missing W in "ww.betacomp.info".
-- Travis
____________________________________________________
Yahoo! Sports
Rekindle the Rivalries. Sign up for Fantasy Football
http://football.fantasysports.yahoo.com