Re: [ros-dev] Re: [ros-svn] [ea] 14244: SM: init system reading the registry

[Steven Edwards]

Hi Emanuele,

--- ea [e-mail] wrote:
> I tested twice fresh builds before committing and the 3rd time right now 
> with local repository at revision 14248. For me it compiles, installs 
> and boots correctly (qemu 0.6.1). Can you tell me more about the lock? I 
> can't rollback right now because I am going out to work in a few 
> seconds. Feel free to restore subsys/csrss and subsys/smss to 14243 if 
> you feel this is a blocking bug for you.

Sorry, its not your problem. I rm -fr'd my reactos tree and did a clean CVS checkout and things
are cool now. I guess the current build system left some crap around that caused problems. Nice
work btw on the smss stuff. How much trouble do you think it would be possible to make a native
mode console application in smss like the Windows recovery console? I have been thinking about
doing this for the boot CD for a while but would like to make it more powerful than the Windows
recovery console.

Thanks
Steven

__________________________________ 
Do you Yahoo!? 
Yahoo! Small Business - Try our new resources site!
http://smallbusiness.yahoo.com/resources/

[Saravanan Jayakumar]

Hi,
 I tried compiling the 0.26 RC1 from sf.net.Most part of the
compilation went thru fine and then came the below error.

halmp: [CC]      mpsirql.c
{standard input}: Assembler messages:
{standard input}:2474: Error: `%fs:(%bl)' is not a valid 32 bit base/index expre
ssion
mingw32-make[1]: *** [mpsirql.o] Error 1
mingw32-make: *** [halx86/mp] Error 2

Any suggestions.
Thanks
 Jay

On Mon, 21 Mar 2005 22:00:13 -0800 (PST), Steven Edwards
[e-mail] wrote:
> Hi Emanuele,
>
> --- ea [e-mail] wrote:
> > I tested twice fresh builds before committing and the 3rd time right now
> > with local repository at revision 14248. For me it compiles, installs
> > and boots correctly (qemu 0.6.1). Can you tell me more about the lock? I
> > can't rollback right now because I am going out to work in a few
> > seconds. Feel free to restore subsys/csrss and subsys/smss to 14243 if
> > you feel this is a blocking bug for you.
>
> Sorry, its not your problem. I rm -fr'd my reactos tree and did a clean CVS checkout and things
> are cool now. I guess the current build system left some crap around that caused problems. Nice
> work btw on the smss stuff. How much trouble do you think it would be possible to make a native
> mode console application in smss like the Windows recovery console? I have been thinking about
> doing this for the boot CD for a while but would like to make it more powerful than the Windows
> recovery console.
>
> Thanks
> Steven
>
> Do you Yahoo!?
> Yahoo! Small Business - Try our new resources site!
> http://smallbusiness.yahoo.com/resources/
> _______________________________________________
> Ros-dev mailing list
> [e-mail]
> http://reactos.com:8080/mailman/listinfo/ros-dev

[James Tabor]

Saravanan Jayakumar wrote:
> Hi,
>  I tried compiling the 0.26 RC1 from sf.net.Most part of the
> compilation went thru fine and then came the below error.
>
> halmp: [CC]      mpsirql.c
> {standard input}: Assembler messages:
> {standard input}:2474: Error: `%fs:(%bl)' is not a valid 32 bit base/index expre
> ssion
> mingw32-make[1]: *** [mpsirql.o] Error 1
> mingw32-make: *** [halx86/mp] Error 2
>
> Any suggestions.
> Thanks
>  Jay

Hi!
Try binary build version 2.15.94 20050118.

That might help,
James

[Magnus Olsen]

Hi
you need binutils 2.15.94 or higher, it will solv your problem.

----- Original Message ----- 
From: "Saravanan Jayakumar" [e-mail]
To: "ReactOS Development List" [e-mail]
Sent: Tuesday, March 22, 2005 7:26 AM
Subject: Re: [ros-dev] Re: [ros-svn] [ea] 14244: SM: init system reading
theregistry
> Hi,
>  I tried compiling the 0.26 RC1 from sf.net.Most part of the
> compilation went thru fine and then came the below error.
>
> halmp: [CC]      mpsirql.c
> {standard input}: Assembler messages:
> {standard input}:2474: Error: `%fs:(%bl)' is not a valid 32 bit base/index

expre
> ssion
> mingw32-make[1]: *** [mpsirql.o] Error 1
> mingw32-make: *** [halx86/mp] Error 2
>
> Any suggestions.
> Thanks
>  Jay
>
> On Mon, 21 Mar 2005 22:00:13 -0800 (PST), Steven Edwards
> [e-mail] wrote:
> > Hi Emanuele,
> >
> > --- ea [e-mail] wrote:
> > > I tested twice fresh builds before committing and the 3rd time right

now
> > > with local repository at revision 14248. For me it compiles, installs
> > > and boots correctly (qemu 0.6.1). Can you tell me more about the lock?

I
> > > can't rollback right now because I am going out to work in a few
> > > seconds. Feel free to restore subsys/csrss and subsys/smss to 14243 if
> > > you feel this is a blocking bug for you.
> >
> > Sorry, its not your problem. I rm -fr'd my reactos tree and did a clean

CVS checkout and things
> > are cool now. I guess the current build system left some crap around

that caused problems. Nice
> > work btw on the smss stuff. How much trouble do you think it would be

possible to make a native
> > mode console application in smss like the Windows recovery console? I

have been thinking about
> > doing this for the boot CD for a while but would like to make it more

powerful than the Windows
> > recovery console.
> >
> > Thanks
> > Steven
> >
> > Do you Yahoo!?
> > Yahoo! Small Business - Try our new resources site!
> > http://smallbusiness.yahoo.com/resources/
> > _______________________________________________
> > Ros-dev mailing list
> > [e-mail]
> > http://reactos.com:8080/mailman/listinfo/ros-dev
>
> Ros-dev mailing list
> [e-mail]
> http://reactos.com:8080/mailman/listinfo/ros-dev

[Richard Campbell]

By the way, the required compiler versions should be updated...binutils 
2.15.94 and gcc 3.4.2.

Magnus Olsen wrote:
> Hi
> you need binutils 2.15.94 or higher, it will solv your problem.
>
> ----- Original Message ----- 
> From: "Saravanan Jayakumar" [e-mail]
> To: "ReactOS Development List" [e-mail]
> Sent: Tuesday, March 22, 2005 7:26 AM
> Subject: Re: [ros-dev] Re: [ros-svn] [ea] 14244: SM: init system reading
> theregistry
> > Hi,
> > I tried compiling the 0.26 RC1 from sf.net.Most part of the
> > compilation went thru fine and then came the below error.
> >
> > halmp: [CC]      mpsirql.c
> > {standard input}: Assembler messages:
> > {standard input}:2474: Error: `%fs:(%bl)' is not a valid 32 bit base/index
>
> expre
> > ssion
> > mingw32-make[1]: *** [mpsirql.o] Error 1
> > mingw32-make: *** [halx86/mp] Error 2
> >
> > Any suggestions.
> > Thanks
> > Jay
> >
> > On Mon, 21 Mar 2005 22:00:13 -0800 (PST), Steven Edwards
> > [e-mail] wrote:
> > > Hi Emanuele,
> > >
> > > --- ea [e-mail] wrote:
> > > > I tested twice fresh builds before committing and the 3rd time right
>
> now
> > > > with local repository at revision 14248. For me it compiles, installs
> > > > and boots correctly (qemu 0.6.1). Can you tell me more about the lock?
>
> I
> > > > can't rollback right now because I am going out to work in a few
> > > > seconds. Feel free to restore subsys/csrss and subsys/smss to 14243 if
> > > > you feel this is a blocking bug for you.
> > >
> > > Sorry, its not your problem. I rm -fr'd my reactos tree and did a clean
>
> CVS checkout and things
> > > are cool now. I guess the current build system left some crap around
>
> that caused problems. Nice
> > > work btw on the smss stuff. How much trouble do you think it would be
>
> possible to make a native
> > > mode console application in smss like the Windows recovery console? I
>
> have been thinking about
> > > doing this for the boot CD for a while but would like to make it more
>
> powerful than the Windows
> > > recovery console.
> > >
> > > Thanks
> > > Steven
> > >
> > > Do you Yahoo!?
> > > Yahoo! Small Business - Try our new resources site!
> > > http://smallbusiness.yahoo.com/resources/
> > > _______________________________________________
> > > Ros-dev mailing list
> > > [e-mail]
> > > http://reactos.com:8080/mailman/listinfo/ros-dev
> >
> > Ros-dev mailing list
> > [e-mail]
> > http://reactos.com:8080/mailman/listinfo/ros-dev
>
> Ros-dev mailing list
> [e-mail]
> http://reactos.com:8080/mailman/listinfo/ros-dev

[ea]

Hi Steven,
> Sorry, its not your problem. I rm -fr'd my reactos tree and did a clean CVS checkout and things
> are cool now. I guess the current build system left some crap around that caused problems. Nice
> work btw on the smss stuff. How much trouble do you think it would be possible to make a native
> mode console application in smss like the Windows recovery console? I have been thinking about
> doing this for the boot CD for a while but would like to make it more powerful than the Windows
> recovery console.

SM needs some work mostly in the security and DBGSS area. I am exploring 
how to move Kmode in CSR, just like winlogon, because it does not belong 
there: this differs slightly from NT, but it's because multiuser support 
won't be like in NT/2k/XP/2k3. Major work and cleanup is needed in CSR: 
I'd like it to be a general purpose server DLL loader, so implementing a 
subsystem would be just the matter of wrinting a "special" library.

The major problem for a native shell (ntosh.exe?) is I/O. As no 
subsystem is present it has to open the keyboard and video devices. It 
would be better implementing first a static library with basic stdio 
functions that relies only on NTDLL. I've seen some code floating in the 
comp newsgroups. That could be a base. I think the ERD Commander by 
Russinovich is a native application (he also has some sample code on his 
site).

Emanuele

[Oliver Schneider]

> caused problems. Nice
> work btw on the smss stuff. How much trouble do you think it would be
> possible to make a native mode console application in smss like the
> Windows recovery console? I have been thinking about doing this for the
> boot CD for a while but would like to make it more powerful than the
> Windows recovery console.

Yes this is possible (early ERD Commander versions have proven this as well
as O&O BlueCon), but the recovery console of Windows is implemented in
kernel mode, I believe.
SMSS is usermode already.

I am working on something like this in a very early stage for Windows
NT/2000/XP to resemble something like ERD Commander (e.g. 2000) and extend
the functionality dynamically by native programs. I am currently stuck on
keyboard input - and for textmode output I am still not sure whether a
separate and specialized driver would be the better solution over a
"proxy-into-kernel" to the Inbv* and similar APIs.

Oliver
-- 
---------------------------------------------------
May the source be with you, stranger ;)

ICQ: #281645
URL: http://assarbad.net