reactos-kernel

[ros-kernel] Re: Re: Ros GUI

[Richard C. <rca...@ho...>]

Well, I'm not talking an exact clone, I'm talking certain elements NEED to 
be there. (start menu, all the same folders in the start menu, my computer 
on desktop, recycle bin on desktop, my network places on desktop, etc.)  If 
you radically change the user interface ReactOS is no longer a windows 
clone, but instead some sort of freak hybrid OS.  As popularity has shown, 
(just look at windows and also at kde/gnome) The 'windows way' may not be 
the best way, but it is the most accepted way.

_________________________________________________________________
MSN Photos is the easiest way to share and print your photos: 
http://photos.msn.com/support/worldwide.aspx

Re: [ros-kernel] Re: ROS gui

[Jarmo M. <ja...@mu...>]

Too many things are related to too many things. What do I mean?
Today I started IE as usual, but it didn't show up. It started to
do something very hard. The HD was loaded. I tried other apps too,
but everything which were related to shell didn't work. I went to
command prompt and wrote dir few times. As I thought IE was
deleting files, from cache I think (probably about 200MBs). After
about 10 minutes every pending operation was performed and finally
IE showed up.

It looks like IE is using SHFileOperation or something like that
and that function and other parts of shell are serialised. How
lame.

I have Windows 2000.

JMu

----- Original Message -----
From: "KJK::Hyperion" <no...@li...>
To: <rea...@li...>
Cc: <rea...@li...>
Sent: Saturday, July 20, 2002 4:43 PM
Subject: Re: [ros-kernel] Re: ROS gui


> At 05.31 20/07/2002, Royce Mitchell III wrote:
> >( For example, every time I delete a large quantity of files -
and I do
> >mean every time - the computer locks up for 5 - 45 seconds. As
a developer
> >I delete a large quantity of files rather frequently, and I
find this very
> >annoying. )
>
> I think it's a Windows 98 problem - still, I'll add this to the
wishlist
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> reactos-kernel mailing list
> rea...@li...
> https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>

Re: [ros-kernel] Re: ROS gui

[KJK::Hyperion <no...@li...>]

At 05.53 20/07/2002, Royce Mitchell III wrote:
>Here's an impartial feature/non-feature list for an explorer clone:

I'm keeping a list. If anyone wants to contribute theirs, go on

>* keep in-situ start menu editing

and right-click for items. Maybe also give a title bar to submenus, to let 
them be "teared" away and remain on-screen

>* remove "web view"

yeah! this was a big mistake

>* everything can be customized, and should be easy to do so ( no editing 
>text files - that's a Linux thing )

right! I'm glad I'm not the only one to think this. ReactOS is a Windows 
and should look and feel like one

>* xp-like user management ( ms finally got that right, if xp wasn't such a 
>flop )

what do you mean?

>* extension hiding should *not* be default for power users.
>* file hiding should *not* be default for power users.

well said. Implementing this will also require multiple skeleton/template 
profiles - something that Windows doesn't have

>* There should *never* be any reason for the OS to open a file unless the 
>user has performed some action on the file, or on a program that uses the 
>file. ( This will eliminate some of the nastier security holes explorer 
>has been plagued by ).

this sounds a bit generic...

>* file associations could work MUCH better. It's always a pain setting up 
>my .c, .h, .cpp, .hpp, etc associations.

any idea about how to do it? I have several, but they would break existing 
programs

Re: [ros-kernel] Re: ROS gui

[KJK::Hyperion <no...@li...>]

At 07.16 20/07/2002, Jarmo Muukka wrote:
>Has anyone pressed Num Pad * when focus is in Network Neighborhood (or 
>Computers Near Me)? This operation will expand A LOT and it cannot be 
>stopped! You have to kill the process. Pressing Esc would be nice way to 
>stop it and a Stop button.

right. I noticed this too. Wishlisted

>One more annoying thing that comes to mind. When you paste files (Copy and 
>Paste) to one folder and switch to another folder, Explorer activates the 
>previous folder when it displays Copy progress dialog. So, no call to 
>SetForegroundWindow, please.

same here. Wishlisted

Re: [ros-kernel] Service.exe

[Eric K. <ek...@rz...>]

"Jarmo Muukka" <ja...@mu...> wrote:


> Does it go something like this? My SCP calls OpenSCManager in
> advapi32.dll. OpenSCManager calls "some procedure" on SCM on
> remote computer through RPC. Remote computer opens its database
> and return "some info" to the caller.
> 
> If I have service open, my SCP calls StartService. StartService in
> advapi32.dll calls "some procedure" on remote computer through
> RPC. Remote computer's SCM starts the process (if not running) and
> creates/resumes a thread of the service.

Exactly!

> Now, where are the pipes needed for? I don't see need for pipes
> between SCM server and SP. They are in the same machine.... Now I
> got it, I think. StartServiceCtrlDispatcher is executing in
> address space of the service process. Services.exe runs in another
> address space. When services.exe receives a RPC, it communicates
> with SP (i.e. StartServiceCtrlDispatcher) with pipes. Does it? So
> you can implement it as you like.

Exactly!

Re: [ros-kernel] Re: ROS gui

[Eric K. <ek...@rz...>]

"Royce Mitchell III" <ro...@ev...> wrote:


> We have a fundamental difference in philosophy :)
>
> Here's why I said what I said:
>
> I consider myself a power user. Almost everyone who knows me calls me to
fix
> their computer problems. I'm the local "computer geek". I've had a lot of
> experience reinstalling Windows, and the most annoying thing about
> reinstalling Windows is having to reconfigure Windows and the applications
> with the settings I want. When applications use INI files, I don't have to
> reconfigure them! They work with the new OS! Applications that use the reg
> have to be reinstalled most of the time, although some applications (
> PageMaker 6.5 & TurboCad 6 Professional for example ) will just recreate
> their registry entries; although I still have to reconfigure them for my
> preferences.

I guess a registry backup applications that ignores system settings could
fix this.


> On animations, there are ways to show that a program is working that
doesn't
> require resource-intensive animations. Progress dialogs are a good
example.
> Another way is to have count-up/down timers, and status text windows
> explaining what's happening at that moment.

A progress bar that sticks at the 35% mark for five minutes will make the
user believe the application froze. That's what I thought when I installed
OpenOffice two weeks ago.


> Most settings are a matter of a bool or int. How much memory does that
> really take up?

How much memory do waste by not using hand-optimized assemly language but
C++, C# or a 4GL language? It always depends on which prize you are willing
to pay.


Eric

Re: [ros-kernel] Re: ROS gui

[KJK::Hyperion <no...@li...>]

At 05.31 20/07/2002, Royce Mitchell III wrote:
>( For example, every time I delete a large quantity of files - and I do 
>mean every time - the computer locks up for 5 - 45 seconds. As a developer 
>I delete a large quantity of files rather frequently, and I find this very 
>annoying. )

I think it's a Windows 98 problem - still, I'll add this to the wishlist

[ros-kernel] GCC 3.1 Summary

[Jurgen V. G. <jur...@st...>]

For those interested,

I was able to compile reactos under GCC 3.1 after a nasty hack. The
problem is that ntosknrl and win32k both reference _imp__pctype but it
seems to be undefined. After adding <unsigned short** _imp___pctype;> to
both files everything compiles fine. Any clues on what this could be?


Jurgen

PS: compilation != running: bochs exits when running reactos after
freeldr booted. Is there a way to get a bochs log or something?

RE: [ros-kernel] Service.exe

[Jurgen V. G. <jur...@st...>]

> If I have service open, my SCP calls StartService. 
> StartService in advapi32.dll calls "some procedure" on remote 
> computer through RPC. Remote computer's SCM starts the 
> process (if not running) and creates/resumes a thread of the service.
> 
> Now, where are the pipes needed for?

Correct me if I am mistaken, but services.exe communicates with services
over a pipe. It send start, stop, pause, ... Messages over a pipe to the
service.

Jurgen

Re: [ros-kernel] Re: ROS gui

[Jarmo M. <ja...@mu...>]

In the old days when you installed the OS, you just copied the
apps to the disk. Settings file included. Now many apps require
installation.

My own app uses registry because it was designed for multiple
users. With INI files the settings should be separate for every
user. Also INI files will be more difficult than registry if it
uses hierarchy as my next version will use. So there is pros &
cons for registry and INI files.

My app installs itself when it notices that registry settings are
not there. And of course, ALL personal registry settings are
removed when unistalled. And when the last user of the app
uninstalls it, everything is removed from the registry and all
files and directories are deleted, if user wants.

I have an idea of "application private registry". It is located in
the application's folder. Registry functions would work as before.
It is just bound to system registry so it seems to be normal. OS
could keep list of bound registries and if not found, the bound
registry is removed from the list. So you can just delete the
application folder and your settings are deleted from OS as well.
Unfortunately this idea cannot be implemented.

JMu

----- Original Message -----
From: "Royce Mitchell III" <ro...@ev...>
To: <rea...@li...>
Sent: Saturday, July 20, 2002 5:13 PM
Subject: Re: [ros-kernel] Re: ROS gui


> We have a fundamental difference in philosophy :)
>
> Here's why I said what I said:
>
> I consider myself a power user. Almost everyone who knows me
calls me to fix
> their computer problems. I'm the local "computer geek". I've had
a lot of
> experience reinstalling Windows, and the most annoying thing
about
> reinstalling Windows is having to reconfigure Windows and the
applications
> with the settings I want. When applications use INI files, I
don't have to
> reconfigure them! They work with the new OS! Applications that
use the reg
> have to be reinstalled most of the time, although some
applications (
> PageMaker 6.5 & TurboCad 6 Professional for example ) will just
recreate
> their registry entries; although I still have to reconfigure
them for my
> preferences.
>
> As a developer, I *never* use the registry unless it's necessary
to hide
> settings, or my customer requests it.
>
> On animations, there are ways to show that a program is working
that doesn't
> require resource-intensive animations. Progress dialogs are a
good example.
> Another way is to have count-up/down timers, and status text
windows
> explaining what's happening at that moment.
>
> Most settings are a matter of a bool or int. How much memory
does that
> really take up?
>
> Royce3
>
> ----- Original Message -----
> From: "Eric Kohl" <ek...@rz...>
> To: <rea...@li...>
> Sent: Saturday, July 20, 2002 5:52 AM
> Subject: Re: [ros-kernel] Re: ROS gui
>
>
> >
> > "Royce Mitchell III" <ro...@ev...> wrote:
> >
> >
> > > * there's no reason for explorer to access 20 registry
entries every
> time
> > > you double-click something. It should be blind to settings
changes made
> > > directly to it's config files/registry. All settings should
be available
> > > from within the program so that it's not necessary to edit
the config
> > > files/reg directly. At the very least it should re-read it's
settings
> > unless
> > > it detects they have changed. ( Probably would have to be an
ini file to
> > > monitor for file changes ).
> >
> > Looking up registry keys and values is not an expensive task
in terms of
> > time consumption, but keeping *all* settings in memory is
expensive in
> terms
> > of memory consumption. This also obsoletes the concept of
monitoring
> changes
> > to the registry, at least in most cases.
> >
> > Btw, classic .ini-files are *dead*. Don't ever use them to
store current
> > application settings! ReactOS should be an '.ini-free' zone.
> >
> > What really needs to be fixed is the registry management. I
really like
> the
> > idea of being able to build updated and customized install
media.
> >
> > > * this isn't just for explorer... but there should be very
little (
> > ideally
> > > none ) images/animations included. They just slow the O/S
down. If they
> > are
> > > necessary, then they should not be loaded until the first
time they are
> > > needed.
> >
> > IMO, animations are useful to show that lengthy jobs are still
running.
> > Usually, animations are resources of an application. They will
be loaded
> > when the application is started but they will be swapped to
the pagefile
> > when they are not in use.
> >
> > Regards,
> > Eric
> >
> >
> >
> > -------------------------------------------------------
> > This sf.net email is sponsored by:ThinkGeek
> > Welcome to geek heaven.
> > http://thinkgeek.com/sf
> > _______________________________________________
> > reactos-kernel mailing list
> > rea...@li...
> > https://lists.sourceforge.net/lists/listinfo/reactos-kernel
> >
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> reactos-kernel mailing list
> rea...@li...
> https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>

Re: [ros-kernel] Service.exe

[Jarmo M. <ja...@mu...>]

Does it go something like this? My SCP calls OpenSCManager in
advapi32.dll. OpenSCManager calls "some procedure" on SCM on
remote computer through RPC. Remote computer opens its database
and return "some info" to the caller.

If I have service open, my SCP calls StartService. StartService in
advapi32.dll calls "some procedure" on remote computer through
RPC. Remote computer's SCM starts the process (if not running) and
creates/resumes a thread of the service.

Now, where are the pipes needed for? I don't see need for pipes
between SCM server and SP. They are in the same machine.... Now I
got it, I think. StartServiceCtrlDispatcher is executing in
address space of the service process. Services.exe runs in another
address space. When services.exe receives a RPC, it communicates
with SP (i.e. StartServiceCtrlDispatcher) with pipes. Does it? So
you can implement it as you like.

BTW, RPC is unknown territory to me.

JMu

----- Original Message -----
From: "Eric Kohl" <ek...@rz...>
To: <rea...@li...>
Sent: Saturday, July 20, 2002 4:52 PM
Subject: Re: [ros-kernel] Service.exe


>
> "Jarmo Muukka" <ja...@mu...> wrote:
>
>
> > The protocol will matter. Have you ever thought that Win NT
and
> > ReactOS services would have to communicate with each other?
Say
> > you have a ReactOS machine and you want to start service on
> > Windows 2000 machine. ReactOS client part have to connect to
W2K
> > machine and W2K server part have to understand what client is
> > asking and ros have to understand W2K replies.
>
> The communication between a SCP and the SCM uses RPC, not pipes.
The contol
> pipes are only used for communication between the SCM and the
SPs. That's
> why the pipe protocol doesn't matter. Samba already implements
the RPC
> protocol.
>
> When you use a local SCP to start a service on a remote machine,
you tell
> the SCM on the remote machine to start the service.
>
> Eric
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> reactos-kernel mailing list
> rea...@li...
> https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>

Re: [ros-kernel] Re: ROS gui

[Royce M. I. <ro...@ev...>]

----- Original Message -----
> Looking up registry keys and values is not an expensive task in terms of
> time consumption, but keeping *all* settings in memory is expensive in
terms
> of memory consumption. This also obsoletes the concept of monitoring
changes
> to the registry, at least in most cases.

Also, I forget to mention that most settings will take up less space than
the code to call the registry and reload it.

Re: [ros-kernel] Re: ROS gui

[Royce M. I. <ro...@ev...>]

We have a fundamental difference in philosophy :)

Here's why I said what I said:

I consider myself a power user. Almost everyone who knows me calls me to fix
their computer problems. I'm the local "computer geek". I've had a lot of
experience reinstalling Windows, and the most annoying thing about
reinstalling Windows is having to reconfigure Windows and the applications
with the settings I want. When applications use INI files, I don't have to
reconfigure them! They work with the new OS! Applications that use the reg
have to be reinstalled most of the time, although some applications (
PageMaker 6.5 & TurboCad 6 Professional for example ) will just recreate
their registry entries; although I still have to reconfigure them for my
preferences.

As a developer, I *never* use the registry unless it's necessary to hide
settings, or my customer requests it.

On animations, there are ways to show that a program is working that doesn't
require resource-intensive animations. Progress dialogs are a good example.
Another way is to have count-up/down timers, and status text windows
explaining what's happening at that moment.

Most settings are a matter of a bool or int. How much memory does that
really take up?

Royce3

----- Original Message -----
From: "Eric Kohl" <ek...@rz...>
To: <rea...@li...>
Sent: Saturday, July 20, 2002 5:52 AM
Subject: Re: [ros-kernel] Re: ROS gui


>
> "Royce Mitchell III" <ro...@ev...> wrote:
>
>
> > * there's no reason for explorer to access 20 registry entries every
time
> > you double-click something. It should be blind to settings changes made
> > directly to it's config files/registry. All settings should be available
> > from within the program so that it's not necessary to edit the config
> > files/reg directly. At the very least it should re-read it's settings
> unless
> > it detects they have changed. ( Probably would have to be an ini file to
> > monitor for file changes ).
>
> Looking up registry keys and values is not an expensive task in terms of
> time consumption, but keeping *all* settings in memory is expensive in
terms
> of memory consumption. This also obsoletes the concept of monitoring
changes
> to the registry, at least in most cases.
>
> Btw, classic .ini-files are *dead*. Don't ever use them to store current
> application settings! ReactOS should be an '.ini-free' zone.
>
> What really needs to be fixed is the registry management. I really like
the
> idea of being able to build updated and customized install media.
>
> > * this isn't just for explorer... but there should be very little (
> ideally
> > none ) images/animations included. They just slow the O/S down. If they
> are
> > necessary, then they should not be loaded until the first time they are
> > needed.
>
> IMO, animations are useful to show that lengthy jobs are still running.
> Usually, animations are resources of an application. They will be loaded
> when the application is started but they will be swapped to the pagefile
> when they are not in use.
>
> Regards,
> Eric
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> reactos-kernel mailing list
> rea...@li...
> https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>

Re: [ros-kernel] Service.exe

[Eric K. <ek...@rz...>]

"Jarmo Muukka" <ja...@mu...> wrote:


> The protocol will matter. Have you ever thought that Win NT and
> ReactOS services would have to communicate with each other? Say
> you have a ReactOS machine and you want to start service on
> Windows 2000 machine. ReactOS client part have to connect to W2K
> machine and W2K server part have to understand what client is
> asking and ros have to understand W2K replies.

The communication between a SCP and the SCM uses RPC, not pipes. The contol
pipes are only used for communication between the SCM and the SPs. That's
why the pipe protocol doesn't matter. Samba already implements the RPC
protocol.

When you use a local SCP to start a service on a remote machine, you tell
the SCM on the remote machine to start the service.

Eric

Re: [ros-kernel] Service.exe

[Jarmo M. <ja...@mu...>]

Hello,

The protocol will matter. Have you ever thought that Win NT and
ReactOS services would have to communicate with each other? Say
you have a ReactOS machine and you want to start service on
Windows 2000 machine. ReactOS client part have to connect to W2K
machine and W2K server part have to understand what client is
asking and ros have to understand W2K replies.

About working on with same thing as a group of programmers. This
would be easy if you first design interfaces. Pipe has two ends
and it clearly has interface on both ends. If they are first
designed, both ends may write their code separately and modules
that use the pipe ends could be programmed as well.

JMu

----- Original Message -----
From: "Eric Kohl" <ek...@rz...>
To: <rea...@li...>
Sent: Saturday, July 20, 2002 4:11 PM
Subject: Re: [ros-kernel] Service.exe


>
> "Jarmo Muukka" <ja...@mu...> wrote:
>
> Hi!
>
> > One of the most difficult part, to me, is the pipe. What is
the
> > protocol?
> The protocol doesn't matter as long as our implementation
behaves the same
> way as the original implementation. ;-)
>
>
> > So, one could implement SCP, one could write both ends of the
> > pipe, one could write SCMC and one could write SCMS. This
could be
> > easily splitted for four developers.
> That might work if you have a well defined communication
protocol, otherwise
> it doesn't work.
>
>
> > At point I thought that I could implement them, but I don't
have
> > time. I checked ReactOS code and noticed that it wasn't
> > implemented "at all". The code that were implemented started
all
> > services in the process at the same time. That's wrong.
> At present only service processes that constain a single service
can be
> started. Services that share a process are not supported yet.
>
>
> Regards,
> Eric
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> reactos-kernel mailing list
> rea...@li...
> https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>

Re: [ros-kernel] Service.exe

[Eric K. <ek...@rz...>]

"Jarmo Muukka" <ja...@mu...> wrote:

Hi!

> One of the most difficult part, to me, is the pipe. What is the
> protocol?
The protocol doesn't matter as long as our implementation behaves the same
way as the original implementation. ;-)


> So, one could implement SCP, one could write both ends of the
> pipe, one could write SCMC and one could write SCMS. This could be
> easily splitted for four developers.
That might work if you have a well defined communication protocol, otherwise
it doesn't work.


> At point I thought that I could implement them, but I don't have
> time. I checked ReactOS code and noticed that it wasn't
> implemented "at all". The code that were implemented started all
> services in the process at the same time. That's wrong.
At present only service processes that constain a single service can be
started. Services that share a process are not supported yet.


Regards,
Eric

Re: [ros-kernel] Service.exe

[Eric K. <ek...@rz...>]

"Jurgen Van Gael" <jur...@st...> wrote:

Hello Jurgen!

> As I am picking up reactos again I was just working on extending the
> services.exe a bit more as a starting point. I was looking into
> communication etc but as you are working on it maybe that isn't such a
> good idea after all. Let me know if I should look into it further or if
> I should look into WINE stuff or something!

It is fine with me if you want to work on the service manager and services!

I'll commit my changes to services.exe, eventlog.exe and advapi32.dll now.
Please remember that this is work in progress. I will have a close look at
the npfs driver and implement the tranceive request which is needed by
TransactNamedPipe().

I use the eventlog service to implement the communication between the
service manager and the service. To enable the eventlog service you will
need to change the 'start' value of the eventlog service in the system.hiv
file from 4 (disabled) to 2 (auto start).

Regards,
Eric

Re: [ros-kernel] Re: ROS gui

[Eric K. <ek...@rz...>]

"Royce Mitchell III" <ro...@ev...> wrote:


> * there's no reason for explorer to access 20 registry entries every time
> you double-click something. It should be blind to settings changes made
> directly to it's config files/registry. All settings should be available
> from within the program so that it's not necessary to edit the config
> files/reg directly. At the very least it should re-read it's settings
unless
> it detects they have changed. ( Probably would have to be an ini file to
> monitor for file changes ).

Looking up registry keys and values is not an expensive task in terms of
time consumption, but keeping *all* settings in memory is expensive in terms
of memory consumption. This also obsoletes the concept of monitoring changes
to the registry, at least in most cases.

Btw, classic .ini-files are *dead*. Don't ever use them to store current
application settings! ReactOS should be an '.ini-free' zone.

What really needs to be fixed is the registry management. I really like the
idea of being able to build updated and customized install media.

> * this isn't just for explorer... but there should be very little (
ideally
> none ) images/animations included. They just slow the O/S down. If they
are
> necessary, then they should not be loaded until the first time they are
> needed.

IMO, animations are useful to show that lengthy jobs are still running.
Usually, animations are resources of an application. They will be loaded
when the application is started but they will be swapped to the pagefile
when they are not in use.

Regards,
Eric

Re: [ros-kernel] Re: ROS gui

[Robert K. <ro...@ko...>]

Do you also know these people?
Word belongs to Windows. Explorer is the OS/windows. A display runs 
sometimes in DOS-mode. The internet is an application on the desktop and 
consists of just HTML sites. It belongs to either AOL or M$ who 
maintains all the content (including warez-sites?) !

Richard Campbell schrieb:

> I think you guys need to stop thinking like geeks and start thinking 
> like AOLers... (i.e. explorer DOES need to be cloned because it is 
> part of windows nt.)
>
> PS, I've missed out on alot of ROS development (although very few 
> people have missed my flames I'm sure :)) I hope to be up to speed
>
>
>> From: rea...@li...
>> Reply-To: rea...@li...
>> To: rea...@li...
>> Subject: reactos-kernel digest, Vol 1 #239 - 13 msgs
>> Date: Thu, 18 Jul 2002 12:10:27 -0700
>>
>> Send reactos-kernel mailing list submissions to
>> rea...@li...
>>
>> To subscribe or unsubscribe via the World Wide Web, visit
>> https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>> or, via email, send a message with subject or body 'help' to
>> rea...@li...
>>
>> You can reach the person managing the list at
>> rea...@li...
>>
>> When replying, please edit your Subject line so it is more specific
>> than "Re: Contents of reactos-kernel digest..."
>>
>>
>> Today's Topics:
>>
>> 1. =?iso-8859-1?Q?YAW:_yet_another_windows_(old_project)?= 
>> (=?utf-8?Q?ea...@io...?=)
>> 2. Re: YAW: yet another windows (old project) (Royce Mitchell III)
>> 3. Re: YAW: yet another windows (old project) (Thomas J. Hruska)
>> 4. Re: finally got good stack trace... (David Welch)
>> 5. Re: The Explorer replacement we need. (KJK::Hyperion)
>> 6. FW: Proposal - alt.os.reactos Discussion forum for the OpenSource 
>> WindowsNT/2K clone ReactOS (Steven Edwards)
>> 7. newsgroup (Steven Edwards)
>> 8. Re: newsgroup (David Welch)
>> 9. Re: newsgroup (Steven Edwards)
>> 10. Re: finally got good stack trace... (Eric Kohl)
>> 11. Re: finally got good stack trace... (Royce Mitchell III)
>> 12. Re: The Explorer replacement we need. (Nick Date)
>> 13. Re: The Explorer replacement we need. 
>> (=?iso-8859-2?Q?H=E1morszky_Bal=E1zs?=)
>>
>> --__--__--
>>
>> Message: 1
>> Date: Wed, 17 Jul 2002 14:37:44 +0200
>> From: "=?utf-8?Q?ea...@io...?=" <ea...@io...>
>> To: rea...@li...
>> Subject: [ros-kernel] 
>> =?iso-8859-1?Q?YAW:_yet_another_windows_(old_project)?=
>> Reply-To: rea...@li...
>>
>> SXMgaXQgYSBkZWFkIHByb2plY3Q/DQoNCmh0dHA6Ly93d3cuaWNlLnJ1L3lhdy8=
>>
>>
>>
>>
>> --__--__--
>>
>> Message: 2
>> Date: Wed, 17 Jul 2002 15:17:04 -0600 (Central Standard Time)
>> From: Royce Mitchell III <ro...@ev...>
>> To: rea...@li...
>> Subject: Re: [ros-kernel] YAW: yet another windows (old project)
>> Reply-To: rea...@li...
>>
>> That's the impression I got when I checked them out a few
>> weeks ago.
>>
>> Royce3
>>
>> >
>> >Subject: [ros-kernel] YAW: yet another windows (old project)
>> > From: "=?utf-8?Q?ea...@io...?=" <ea...@io...>
>> > Date: Wed, 17 Jul 2002 14:37:44 +0200
>> > To: rea...@li...
>> >
>> >SXMgaXQgYSBkZWFkIHByb2plY3Q/DQoNCmh0dHA6Ly93d3cuaWNlLnJ1L3lhdy8=
>> >
>> >
>> >
>> >
>> >-------------------------------------------------------
>> >This sf.net email is sponsored by:ThinkGeek
>> >Welcome to geek heaven.
>> >http://thinkgeek.com/sf
>> >_______________________________________________
>> >reactos-kernel mailing list
>> >rea...@li...
>> >https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>>
>>
>>
>> --__--__--
>>
>> Message: 3
>> Date: Wed, 17 Jul 2002 16:08:06 -0400
>> To: rea...@li...
>> From: "Thomas J. Hruska" <shi...@sh...>
>> Subject: Re: [ros-kernel] YAW: yet another windows (old project)
>> Reply-To: rea...@li...
>>
>> At 02:37 PM 7/17/2002 +0200, =?utf-8?Q?ea...@io...?= writeth:
>> >SXMgaXQgYSBkZWFkIHByb2plY3Q/DQoNCmh0dHA6Ly93d3cuaWNlLnJ1L3lhdy8=
>>
>> For those who can't read/speak/understand the Base64 language:
>> ------------Start of Base64 Decode---------------
>> Is it a dead project?
>>
>> http://www.ice.ru/yaw/
>> -------------End of Base64 Decode----------------
>>
>> Hope this helps!
>>
>>
>> Thomas J. Hruska -- shi...@sh...
>> Shining Light Productions -- "Meeting the needs of fellow programmers"
>> http://www.shininglightpro.com/
>>
>>
>> --__--__--
>>
>> Message: 4
>> Date: Wed, 17 Jul 2002 22:07:52 +0000
>> From: David Welch <we...@cw...>
>> To: rea...@li...
>> Subject: Re: [ros-kernel] finally got good stack trace...
>> Reply-To: rea...@li...
>>
>> On Tue, Jul 16, 2002 at 10:26:26PM -0500, Royce Mitchell III wrote:
>> > Any ideas?
>> >
>> Yes, ntoskrnl is an exception, just add 0xc0000000 to the printed
>> address when using addr2line. I'm sorry figuring out where the kernel 
>> crashed
>> can be such a bear.
>>
>> > It is still reporting div/0 by the way, and I have latest CVS...
>> >
>> Seems to be working here.
>>
>>
>>
>> --__--__--
>>
>> Message: 5
>> Date: Wed, 17 Jul 2002 23:04:47 +0200
>> To: rea...@li...
>> From: "KJK::Hyperion" <no...@li...>
>> Subject: Re: [ros-kernel] The Explorer replacement we need.
>> Reply-To: rea...@li...
>>
>> At 20.25 17/07/2002, you wrote:
>> >Also wine has some bugs with its CxxFrameHandler
>>
>> what's this?
>>
>> >Dont get me wrong I am not oposed to a project to rewrite explorer.exe.
>>
>> and I'm not supporting such a project :-) Explorer isn't exactly the 
>> best
>> desktop around, it's Shell32 that matters
>>
>> >1. It doesnt link to msvcp60.dll
>>
>> C++ runtime? whoa, it's big. It implements the STL, I see. We could use
>> STLPort for this, it's the STL the OpenOffice guys are using, and it 
>> looks
>> good (http://www.stlport.org/). The problem is that the DLL uses
>> Microsoft's C++ name decoration. I started documenting their decoration
>> scheme some time ago, but I can't find the BNF grammar I hacked together
>> anymore. If you need help on adding this to g++, I can try 
>> reconstructing
>> what I had found out
>>
>> >4. After only 1 day of work I can now compile 50% of geoShell under
>> >mingw/gcc 3.1. I was working on litestep I would be banging my head 
>> in to
>> >the wall right now.
>>
>> ;-)
>>
>> > > but was is usable? would you use it? I downloaded it
>> > > and gave it a try, and
>> > > my answer is ...
>> >I was using it....the only real problem I had was the lack of 
>> Desktop Icons.
>>
>> me too. Anyway, what do you think of my idea of the desktop as an
>> always-on-bottom bar with a Desktop widget in it?
>>
>> > > ... nope, geoshell is not ready for "prime time". It
>> > > didn't even *work* for
>> > > me. Some menus didn't open, some plugins behaved
>> > > odd, some menus flickered.
>> > > A disappointment
>> >The only thig left that really bothers me is so bugs in minamizing the
>> >geobars.
>>
>> I noticed a long series of bug and quirks, instead. The Favorites menu
>> doesn't open (not that I really miss it, since I use Opera full-time).
>> There's a duplicate Programs menu because I have a localized Windows 
>> and it
>> was expecting the folder to be literally called "Programs". The "lock
>> screen" doesn't lock the screen at all, it starts the screen saver. Some
>> control panel applets had the wrong icon or caption. Some icons appeared
>> shrunk. If you want a full bug report, I can reinstall geOShell
>>
>> >I agree adding stuff to the bars could be better... I think the size of
>> >the default geObar doesnt help to much.
>>
>> that's why Office (and most programs with similar toolbars) has a 
>> separate
>> dialog
>>
>> [...]
>> >Come on dude how hard is it to rename the resources?
>>
>> I think you misunderstood me. I'm just pointing out usability problems.
>> Mine is a designer's point of view, not a developer's. But if it's 
>> easy to
>> solve, even better :-)
>>
>> [...]
>> >I dont see why we couldnt implement this if we adopt R4. they are 
>> working
>> >on R6 but are willing to help us if we do future development on R4 
>> for ReactOS.
>>
>> again: it's not (strictly) a technical matter. I just want an opinion 
>> on my
>> remarks: do they make sense to you? do you agree? anything to add?
>>
>> BTW, I didn't ask directly to the geOShell team because I was afraid 
>> they
>> were a bit too "1337" to understand usability arguments (no offense 
>> meant,
>> I just seem to have strange prejudices against IRC'ers)
>>
>>
>>
>> --__--__--
>>
>> Message: 6
>> Date: Wed, 17 Jul 2002 23:50:23 -0700 (PDT)
>> From: Steven Edwards <ste...@ya...>
>> To: rea...@li...
>> Subject: [ros-kernel] FW: Proposal - alt.os.reactos Discussion forum 
>> for the OpenSource WindowsNT/2K clone ReactOS
>> Reply-To: rea...@li...
>>
>> Proposal - alt.os.reactos Discussion forum for the
>> OpenSource WindowsNT/2K clone ReactOS
>>
>> CHARTER: alt.os.reactos discusses design and
>> implementation of the ReactOS operating systems and
>> its
>> relation to the Microsoft Windows Operating Systems.
>> Subjects will include but not be limited to
>> device drivers, win32api support, the posix subsystem
>> and license issues X11/BSD and GNU GPL/LGPL.
>>
>> The newsgroup will not be moderated. Binary,
>> off-topic, spam and advertising posts are not
>> permitted.
>> Discussion of hacks, cracks, warez, serial numbers or
>> the illegal exchange of copyrighted software is
>> banned.
>>
>> JUSTIFICATION: A scan of messages in Google Groups in
>> the 3 months from 5/17/98 to 7/17/98 showed
>> *exactly* 1,010 relevant matches discussing ReactOS.
>> ReactOS is the key topic of no less then 3
>> mailing lists and 2 web message boards and as such
>> central repository such as a news group would be
>> convenient.
>>
>> Thanks
>> Steven Edwards
>>
>> -- 
>> "Every revolution was once a thought in one man's
>> mind"
>> - Ralph Waldo Emerson
>>
>>
>> __________________________________________________
>> Do You Yahoo!?
>> Yahoo! Autos - Get free new car price quotes
>> http://autos.yahoo.com
>>
>>
>> --__--__--
>>
>> Message: 7
>> Date: Wed, 17 Jul 2002 23:52:11 -0700 (PDT)
>> From: Steven Edwards <ste...@ya...>
>> To: rea...@li...
>> Subject: [ros-kernel] newsgroup
>> Reply-To: rea...@li...
>>
>> I forgot to add that you can help by responding in
>> alt.config to help show support.
>>
>> Thanks
>> Steven
>>
>> __________________________________________________
>> Do You Yahoo!?
>> Yahoo! Autos - Get free new car price quotes
>> http://autos.yahoo.com
>>
>>
>> --__--__--
>>
>> Message: 8
>> Date: Thu, 18 Jul 2002 09:35:53 +0000
>> From: David Welch <we...@cw...>
>> To: rea...@li...
>> Subject: Re: [ros-kernel] newsgroup
>> Reply-To: rea...@li...
>>
>> On Wed, Jul 17, 2002 at 11:52:11PM -0700, Steven Edwards wrote:
>> > I forgot to add that you can help by responding in
>> > alt.config to help show support.
>> >
>> Do we really need a newsgroup? I don't see a lot of discussion of
>> ReactOS on usenet and we could always use alt.os.development.
>>
>>
>> --__--__--
>>
>> Message: 9
>> Date: Thu, 18 Jul 2002 01:53:29 -0700 (PDT)
>> From: Steven Edwards <ste...@ya...>
>> Subject: Re: [ros-kernel] newsgroup
>> To: rea...@li...
>> Reply-To: rea...@li...
>>
>> I guess/hope not now. I forgot about the sig. juliet
>> uses and that showed up a 1000 times and got flamed
>> for it. Cancel the request as I have killed the thread
>> and probly pissed everyone off anyway. =P What a bunch
>> of nice people there are on usenet.
>>
>> Thanks
>> Steven
>>
>> > Do we really need a newsgroup? I don't see a lot of
>> > discussion of
>> > ReactOS on usenet and we could always use
>> > alt.os.development.
>>
>>
>> __________________________________________________
>> Do You Yahoo!?
>> Yahoo! Autos - Get free new car price quotes
>> http://autos.yahoo.com
>>
>>
>> --__--__--
>>
>> Message: 10
>> From: "Eric Kohl" <ek...@rz...>
>> To: <rea...@li...>
>> Subject: Re: [ros-kernel] finally got good stack trace...
>> Date: Thu, 18 Jul 2002 13:04:03 +0200
>> Reply-To: rea...@li...
>>
>> "Royce Mitchell III" <ro...@ev...> wrote:
>>
>> > I put a bunch of DbgPrint's all over that function last night, and 
>> didn't
>> > see any of them on-screen...
>>
>> Please try a clean rebuild of the current CVS tree. David fixed some 
>> severe
>> bugs and I got my 9 year old Seagate ST3144A (130 MB) running which 
>> reported
>> 576 byte sectors. ;-)
>>
>> Eric
>>
>>
>>
>> --__--__--
>>
>> Message: 11
>> Date: Thu, 18 Jul 2002 08:44:08 -0600 (Central Standard Time)
>> From: Royce Mitchell III <ro...@ev...>
>> To: rea...@li...
>> Subject: Re: [ros-kernel] finally got good stack trace...
>> Reply-To: rea...@li...
>>
>> I can't wait to try!
>>
>> /me rubbing hands with glee
>>
>> :)
>>
>> Royce3
>>
>> >
>> >Subject: Re: [ros-kernel] finally got good stack trace...
>> > From: "Eric Kohl" <ek...@rz...>
>> > Date: Thu, 18 Jul 2002 13:04:03 +0200
>> > To: <rea...@li...>
>> >
>> >"Royce Mitchell III" <ro...@ev...> wrote:
>> >
>> >> I put a bunch of DbgPrint's all over that function last night, and 
>> didn't
>> >> see any of them on-screen...
>> >
>> >Please try a clean rebuild of the current CVS tree. David fixed some 
>> severe
>> >bugs and I got my 9 year old Seagate ST3144A (130 MB) running which 
>> reported
>> >576 byte sectors. ;-)
>> >
>> >Eric
>> >
>> >
>> >
>> >-------------------------------------------------------
>> >This sf.net email is sponsored by:ThinkGeek
>> >Welcome to geek heaven.
>> >http://thinkgeek.com/sf
>> >_______________________________________________
>> >reactos-kernel mailing list
>> >rea...@li...
>> >https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>>
>>
>>
>> --__--__--
>>
>> Message: 12
>> From: "Nick Date" <nic...@ya...>
>> To: <rea...@li...>
>> Subject: Re: [ros-kernel] The Explorer replacement we need.
>> Date: Thu, 18 Jul 2002 17:28:43 +0100
>> Reply-To: rea...@li...
>>
>> Hiya.
>>
>> Sorry to step in after being away for so long (my life's up in the 
>> air at
>> the moment to say the least).
>>
>> ----- Original Message -----
>> From: "KJK::Hyperion" <no...@li...>
>> To: <rea...@li...>
>> Sent: Wednesday, July 17, 2002 10:04 PM
>> Subject: Re: [ros-kernel] The Explorer replacement we need.
>>
>>
>> > At 20.25 17/07/2002, you wrote:
>>
>> > >Dont get me wrong I am not oposed to a project to rewrite 
>> explorer.exe.
>>
>> > and I'm not supporting such a project :-) Explorer isn't exactly 
>> the best
>> > desktop around, it's Shell32 that matters
>>
>> [snip]
>>
>> You're right of course, but I think it would be nice to have a shell 
>> with a
>> windoze-a-like feel to it. End users hate change of any sort and a
>> familiar(ish) interface would probably stop them having a heart 
>> attack, or
>> mental breakdown or whatever. ;-)
>>
>> Hi to everyone and sorry I'm not active in the message lists at the 
>> moment.
>> I've got a new job starting on monday (warehouse work at a plumbing
>> merchants) so things should start to pick up and I'll have more free 
>> time to
>> keep up with the project.
>>
>> Best regards.
>>
>> Nick.
>>
>> -- 
>> Nick Date
>> Bath, England, UK
>>
>>
>> __________________________________________________
>> Do You Yahoo!?
>> Everything you'll ever need on one web page
>> from News and Sport to Email and Music Charts
>> http://uk.my.yahoo.comm
>>
>>
>> --__--__--
>>
>> Message: 13
>> From: =?iso-8859-2?Q?H=E1morszky_Bal=E1zs?= <ba...@cr...>
>> To: rea...@li...
>> Subject: Re: [ros-kernel] The Explorer replacement we need.
>> Date: Thu, 18 Jul 2002 20:11:53 +0200
>> Reply-To: rea...@li...
>>
>>
>> Take two look:
>> http://www.geoshellx.com/userguide.asp?doc=desktop.inc
>> http://www.geoshellx.com/plugins/r48/geoDeskInstaller_1_0.exe
>>
>>
>>
>>
>>
>> --__--__--
>>
>> _______________________________________________
>> reactos-kernel mailing list
>> rea...@li...
>> https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>>
>>
>> End of reactos-kernel Digest
>
>
>
>
>
> _________________________________________________________________
> Join the world's largest e-mail service with MSN Hotmail. 
> http://www.hotmail.com
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> reactos-kernel mailing list
> rea...@li...
> https://lists.sourceforge.net/lists/listinfo/reactos-kernel

Re: [ros-kernel] Re: ROS gui

[Robert K. <ro...@ko...>]

>
>
>Okay, I was thinking that he was talking about explorer the browser not the 
>shell interface. The shell should be replaceable with something that the user 
>wants. It would seem to me that this is a flaw in design, unless you allow 
>things like themes which can change all the button and other widget styles 
>along with the window apperance and things of this nature.
>
What you think of ist a customizable GUI. The actual windwos and buttons 
are drawn by a lib.
A Shellprogram gives an more less universal interface to the OS/its 
files and should not provide Skins ot th. like that.

>
>James Marjie
>
>
>-------------------------------------------------------
>This sf.net email is sponsored by:ThinkGeek
>Welcome to geek heaven.
>http://thinkgeek.com/sf
>_______________________________________________
>reactos-kernel mailing list
>rea...@li...
>https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>  
>

Re: [ros-kernel] Service.exe

[Jarmo M. <ja...@mu...>]

Hello,

I think that you both can implement pieces of the services at the
same time. I have some knowledge of services. I see services as
following classes (I'm OO designer/programmer):

- Service Control Program (SCP)
- Service Control Manager Client (SCMC)
- Service Control Manager Server (SCMS)
- Service Process (SP)

Service Control Program is the Services applet. This is clearly
the easiest module to implement.

Service Control Manager Client is the module which SCP uses and
it's in advapi32.dll. This module implements client part. SCMC
connects to SCMS by named pipe.

Service Control Manager Server is the module where SCMC connects
to and it's in services.exe. This module implement server part.
SCMS starts the Service Process and creates a thread for service
when the service is started.

Service Process uses advapi32.dll too, but uses different APIs
that SCP.

One of the most difficult part, to me, is the pipe. What is the
protocol?

So, one could implement SCP, one could write both ends of the
pipe, one could write SCMC and one could write SCMS. This could be
easily splitted for four developers.

At point I thought that I could implement them, but I don't have
time. I checked ReactOS code and noticed that it wasn't
implemented "at all". The code that were implemented started all
services in the process at the same time. That's wrong.

JMu

----- Original Message -----
From: "Jurgen Van Gael" <jur...@st...>
To: <rea...@li...>
Sent: Saturday, July 20, 2002 1:35 AM
Subject: RE: [ros-kernel] Service.exe


> Well,
>
>
> As I am picking up reactos again I was just working on extending
the
> services.exe a bit more as a starting point. I was looking into
> communication etc but as you are working on it maybe that isn't
such a
> good idea after all. Let me know if I should look into it
further or if
> I should look into WINE stuff or something!
>
> Jurgen
>
> -----Original Message-----
> From: rea...@li...
> [mailto:rea...@li...] On Behalf Of
Eric
> Kohl
> Sent: Friday, July 19, 2002 23:14
> To: ReactOS Kernel
> Subject: Re: [ros-kernel] Service.exe
>
>
>
> "Jurgen Van Gael" <jur...@st...> wrote:
>
>
> > Well I was extending the functionality of services.exe. It's
"My"
> > starting point to get more into the internals of ReactOS. Do
you mind
> > me working on this as this is currently within my reach? (I'd
be more
> > than happy if you could email me your work on the
communication and I
> > might even extend your work ...)
>
> What part of services are you working on exactly?
>
> I'm currently working on the creation of service contol pipes
and
> sending a start command to the first service. This also includes
several
> modifications to lib/advapi32/service/sctrl.c.
>
> Eric
>
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
_______________________________________________
> reactos-kernel mailing list rea...@li...
> https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> reactos-kernel mailing list
> rea...@li...
> https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>

[ros-kernel] FURPS

[Jarmo M. <ja...@mu...>]

Hello,

I thought that it would be cool idea to have a web page where
users would have a possibility to see and add requirements. These
could be possible to see/add per module. It should be controlled
someway. There would be many requirements that would be the same.
Also, there would be requirements that is not possible to
implement because other requirements are the opposite. If they all
are reasonabe requirements, this way you could see that some kind
of option is required. One person per module could accept the
requirements.

We are using FURPS (functionality, usability, reliability,
performance, and supportability) at work. It's good way of seeing
and documenting what could be done. We collect all requirements
and write them down. We collect all that we can think of, but we
will accept only must requirements for version 1.0 which we need
as soon as possible. After the version 1.0 is up and running, we
add less important, but still important features to system. I know
that some of my ideas are too expensive to implement, but I added
them to the list as well.

My way of doing it is to first add a general requirement, like
"Company has employees". Then I start splitting this req to
smaller pieces:
- Add new employee
- Update employee
- Delete employee
- Delete is not possible, if employee has hour bookings
- Reject duplicate employee by full name
- List employees
- List employees of department
- Mark employee as deleted
- Mark deleted employee is not listed generally (this is a bit too
abstract req)
- List mark deleted employees
- Employee has following data: xxxxx, yyyyy, zzzzz, etc.
- etc.

Most of my reqs are FRs (functional requirements) and some are UR
(usability reqs). I collect data too and put them as SR
(supportability reqs).

I hope that you got the point. When you read this you might not
like something, or think that it's too complicated, or there is
something missing, or req is not needed, or it can be implemented
later, etc.

When you are very detailed in your reqs, you get a lot of
information
a) for what could/should be done
b) for test cases
c) for user interface design
d) for database design
e) for programming

One detailed req is easy to design, easy to program and easy to
test. I am very proud of my work what I have done in two weeks. I
have received positive feedback from my colleagues and from
customer. I feel that adding detailed requirements before we write
a line of code will pay it back. By the way, the reqs are so
general that there are no UI specific things. Some reqs may be
easy with Centura/SQLWindows and difficult with web browser.

When this is done for ReactOS Explorer, the list would be very
long. So, you see how much code you have to write. It also helps
designing source code files. And putting things to classes. And
organising development.

What next? Then the most complicated reqs are analysed with UML. I
think that we make use case diagrams, statechart diagrams and
activity diagrams. We don't make class diagrams. I design database
with ERwin. I will do this after summer vacation.

Was this interesting or not?

JMu

Re: [ros-kernel] Re: ROS gui

[James M. <jid...@sa...>]

Okay, I was thinking that he was talking about explorer the browser not t=
he=20
shell interface. The shell should be replaceable with something that the =
user=20
wants. It would seem to me that this is a flaw in design, unless you allo=
w=20
things like themes which can change all the button and other widget style=
s=20
along with the window apperance and things of this nature.

James Marjie

Re: [ros-kernel] Re: ROS gui

[Jarmo M. <ja...@mu...>]

Has anyone pressed Num Pad * when focus is in Network Neighborhood
(or Computers Near Me)? This operation will expand A LOT and it
cannot be stopped! You have to kill the process. Pressing Esc
would be nice way to stop it and a Stop button.

One more annoying thing that comes to mind. When you paste files
(Copy and Paste) to one folder and switch to another folder,
Explorer activates the previous folder when it displays Copy
progress dialog. So, no call to SetForegroundWindow, please.

JMu

----- Original Message -----
From: "Royce Mitchell III" <ro...@ev...>
To: <rea...@li...>
Sent: Saturday, July 20, 2002 6:53 AM
Subject: Re: [ros-kernel] Re: ROS gui


> > Why should we clone explorer? If we use a free one or wite a
free one, if
> > the person doesn't like it could they just not download the
newest
> > explorer.
>
> I think an Explorer clone should adhere closely to the
"interface" of the
> original, so newbies will be immediately familiar with it. Just
because MS's
> is insecure, doesn't mean ours has to be.
>
> Here's an impartial feature/non-feature list for an explorer
clone:
>
> * keep quick-launch
> * keep in-situ start menu editing
> * remove start menu reordering ( should always alphabetize )
> * remove "web view"
> * keep desktop
> * everything can be customized, and should be easy to do so ( no
editing
> text files - that's a Linux thing )
> * it should be possible to edit settings in a text file - if you
want to ( a
> single, simple ini file - there are several reasons for this one
if you want
> me to expound )
> * keep control panel
> * xp-like start menu ( it really is a much better layout when
you get rid of
> the performance-sapping fancy graphics )
> * xp-like user management ( ms finally got that right, if xp
wasn't such a
> flop )
> * extension hiding should *not* be default for power users.
> * file hiding should *not* be default for power users.
> * no web-integration. web browser must be a separate app.
> * There should *never* be any reason for the OS to open a file
unless the
> user has performed some action on the file, or on a program that
uses the
> file. ( This will eliminate some of the nastier security holes
explorer has
> been plagued by ).
> * file associations could work MUCH better. It's always a pain
setting up my
> .c, .h, .cpp, .hpp, etc associations.
> * there's no reason for explorer to access 20 registry entries
every time
> you double-click something. It should be blind to settings
changes made
> directly to it's config files/registry. All settings should be
available
> from within the program so that it's not necessary to edit the
config
> files/reg directly. At the very least it should re-read it's
settings unless
> it detects they have changed. ( Probably would have to be an ini
file to
> monitor for file changes ).
> * 9x/NT4-like network neighborhood, unless someone can explain
to me how the
> 2000/XP net hoods are better ( I have yet to find a good
reason ). Note I'm
> talking about browsing the network here, not configuring it.
> * this isn't just for explorer... but there should be very
little ( ideally
> none ) images/animations included. They just slow the O/S down.
If they are
> necessary, then they should not be loaded until the first time
they are
> needed.
>
> That's all I can think of right now.
>
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> reactos-kernel mailing list
> rea...@li...
> https://lists.sourceforge.net/lists/listinfo/reactos-kernel
>

Re: [ros-kernel] Re: ROS gui

[Royce M. I. <ro...@ev...>]

> Why should we clone explorer? If we use a free one or wite a free one, if
> the person doesn't like it could they just not download the newest
> explorer.

I think an Explorer clone should adhere closely to the "interface" of the
original, so newbies will be immediately familiar with it. Just because MS's
is insecure, doesn't mean ours has to be.

Here's an impartial feature/non-feature list for an explorer clone:

* keep quick-launch
* keep in-situ start menu editing
* remove start menu reordering ( should always alphabetize )
* remove "web view"
* keep desktop
* everything can be customized, and should be easy to do so ( no editing
text files - that's a Linux thing )
* it should be possible to edit settings in a text file - if you want to ( a
single, simple ini file - there are several reasons for this one if you want
me to expound )
* keep control panel
* xp-like start menu ( it really is a much better layout when you get rid of
the performance-sapping fancy graphics )
* xp-like user management ( ms finally got that right, if xp wasn't such a
flop )
* extension hiding should *not* be default for power users.
* file hiding should *not* be default for power users.
* no web-integration. web browser must be a separate app.
* There should *never* be any reason for the OS to open a file unless the
user has performed some action on the file, or on a program that uses the
file. ( This will eliminate some of the nastier security holes explorer has
been plagued by ).
* file associations could work MUCH better. It's always a pain setting up my
.c, .h, .cpp, .hpp, etc associations.
* there's no reason for explorer to access 20 registry entries every time
you double-click something. It should be blind to settings changes made
directly to it's config files/registry. All settings should be available
from within the program so that it's not necessary to edit the config
files/reg directly. At the very least it should re-read it's settings unless
it detects they have changed. ( Probably would have to be an ini file to
monitor for file changes ).
* 9x/NT4-like network neighborhood, unless someone can explain to me how the
2000/XP net hoods are better ( I have yet to find a good reason ). Note I'm
talking about browsing the network here, not configuring it.
* this isn't just for explorer... but there should be very little ( ideally
none ) images/animations included. They just slow the O/S down. If they are
necessary, then they should not be loaded until the first time they are
needed.

That's all I can think of right now.