Installing VS .NET 2003 SP1 on Vista

[DavidW]

The application installed okay, but attempted installation of SP1 produces an
error 1606 ("Could not access network location VS Developers (cannot create
group)"). It's hard to believe that any network access error should be fatal to
installing the SP, but it's enough to kill it.

KB article 886549 describes a fix for the problem by modifying the registry, but
it's clearly intended for an OS pre-Vista. A couple of the reg entries listed in
the article were missing in my registry (Local Settings and Common Favorites). I
tried adding these with guesses for the correct values (and made sure the
locations exist), but it still fails. Does anyone have a fix? Thanks.

 

[David Wilkinson]

The application installed okay, but attempted installation of SP1 produces an
error 1606 ("Could not access network location VS Developers (cannot create
group)"). It's hard to believe that any network access error should be fatal to
installing the SP, but it's enough to kill it.

KB article 886549 describes a fix for the problem by modifying the registry, but
it's clearly intended for an OS pre-Vista. A couple of the reg entries listed in
the article were missing in my registry (Local Settings and Common Favorites). I
tried adding these with guesses for the correct values (and made sure the
locations exist), but it still fails. Does anyone have a fix? Thanks.

David:

AFAIK VS2003 is not supported on Vista. You can't upgrade to VS2008?

 

[DavidW]

David:

AFAIK VS2003 is not supported on Vista. You can't upgrade to VS2008?

Just because it is "not supported" doesn't mean it won't run. It is evident from
a Google search and from the KB article that error 1606 crops up in a multitude
of circumstances on previous OSes. I really doubt that incompatibility with
Vista is the cause of the problem, particularly when VS 2003 installed okay in
its entirety and only the SP is giving a problem.

Again on "not supported", this is another reason almost the entire world detests
Microsoft. VS is an expensive product and people don't have endless money to
throw at upgrades, or the time to install another huge application and learn
another new interface, because MS deigns from its lofty, monopolistic position
that it's no longer going to support it. Furthermore, it is highly unlikely that
a large project built with VS 2003 can simply be rebuilt with VS 2008 and it
will all work fine. There are a lot of products out there that developers just
want to maintain with a minimum of fuss. People don't have infinite time and
money to deal with every obstacle that Microsoft deliberately places in their
path.

 

[Brian Muth]

Again on "not supported", this is another reason almost the entire world detests
Microsoft. VS is an expensive product and people don't have endless money to
throw at upgrades...

Well, there is always Visual Studio 2008 Express. You can't beat that price.

or the time to install another huge application and learn
another new interface, because MS deigns from its lofty, monopolistic position
that it's no longer going to support it.

The interface of VS 2008 is very similar, and won't require a learning curve of any significance.

Furthermore, it is highly unlikely that
a large project built with VS 2003 can simply be rebuilt with VS 2008 and it
will all work fine.

I personally had no issues upgrading any of my old VS 2003 or VS 2005 projects to VS 2008. If there are problems, it's likely
because VS 2008 is more standards conformant that VS 2003, and may in fact detect problems with your coding. Isn't that a good
thing?

There are a lot of products out there that developers just
want to maintain with a minimum of fuss. People don't have infinite time and
money to deal with every obstacle that Microsoft deliberately places in their
path.

Ok, we get the message that you don't want to waste time. That's why David Wilkinson suggested that it may be better to upgrade.
Besides you get a much better compiler.

By the way, I found your statements highly inflammatory. If you want to post such comments you should be prepared to back it up with
specifics. What obstacles are you talking about?

Regards,

Brian

 

[DavidW]

Well, there is always Visual Studio 2008 Express. You can't beat that price.


The interface of VS 2008 is very similar, and won't require a learning curve
of any significance.


I personally had no issues upgrading any of my old VS 2003 or VS 2005 projects
to VS 2008. If there are problems, it's likely because VS 2008 is more
standards conformant that VS 2003, and may in fact detect problems with your
coding. Isn't that a good thing?

Fine, but it's not my usual experience in upgrading development systems, and I
can't know it will be trouble-free without getting it, installing it, building
the project and then re-testing the whole thing rigorously. Are you are trying
to tell me that doing all that is easier and quicker than just using what I've
got?

Ok, we get the message that you don't want to waste time. That's why David
Wilkinson suggested that it may be better to upgrade. Besides you get a much
better compiler.

By the way, I found your statements highly inflammatory. If you want to post
such comments you should be prepared to back it up with specifics. What
obstacles are you talking about?

I'm talking about the particular obstacle under discussion. New projects would
have been built using VS 2003 up to, and probably significantly beyond, the
release of VS 2005. So we are talking barely two years since then and already MS
says it won't support that product on their new shiny OS. That's not good
enough.

 

[Brian Muth]

Fine, but it's not my usual experience in upgrading development systems, and I can't know it will be trouble-free without getting
it, installing it, building the project and then re-testing the whole thing rigorously. Are you are trying to tell me that doing
all that is easier and quicker than just using what I've got?

Given the choice of either trying to get VS 2003 to run under Vista or upgrading to VS 2008, I'd choose the latter in a heartbeat.

Remember, you have other choices. You don't _have_ to run it under Vista. Why not develop it under the current OS? Or, you could run
Virtual PC (which is free) under Vista, and keep your VS 2003 running in its current environment.

I'm talking about the particular obstacle under discussion. New projects would have been built using VS 2003 up to, and probably
significantly beyond, the release of VS 2005. So we are talking barely two years since then and already MS says it won't support
that product on their new shiny OS. That's not good enough.

VS 2003 came out long before Vista hit the market. VS 2005 was a much better product, as it replaced the unreliable Managed C++
extensions with the C++/CLI technology, and of course supported .NET Framework 2.0 instead of the more immature 1.1. So there are
other good technical reasons to upgrade. Like it or not, the reality is that the sun is starting to set on this product, whether or
not it's capable of running on Vista.

Regards,

Brian

 

[Ben Voigt [C++ MVP]]

I personally had no issues upgrading any of my old VS 2003 or VS 2005

Fine, but it's not my usual experience in upgrading development systems,
and I can't know it will be trouble-free without getting it, installing
it, building the project and then re-testing the whole thing rigorously.
Are you are trying to tell me that doing all that is easier and quicker
than just using what I've got?

I totally fail to see how that argues in favor of VS2003 on Vista.

If the retesting is too expensive, you have no business using Vista. If
you're going to the trouble of retesting with a new OS, you should upgrade
your compiler at the same time to get the most return from your testing
effort.

 

[DavidW]

Given the choice of either trying to get VS 2003 to run under Vista or
upgrading to VS 2008, I'd choose the latter in a heartbeat.

Remember, you have other choices. You don't _have_ to run it under Vista. Why
not develop it under the current OS?

I might have to do that, and then transfer the built application to Vista every
time I want to run it. Hardly ideal.

Or, you could run Virtual PC (which is free) under Vista, and keep your VS
2003 running in its current environment.

Yes, and hope that's all hassle-free too. Certainly neither of these is
preferable to simply running VS 2003 on Vista, which I suspect it would do if
not for this silly error that probably has nothing to do with Vista
incompatibilty. If I'm wrong then it's a bad sign for Vista compatibility with
the world's pre-Vista software if MS's own products won't run on it.

VS 2003 came out long before Vista hit the market. VS 2005 was a much better
product, as it replaced the unreliable Managed C++ extensions with the C++/CLI
technology, and of course supported .NET Framework 2.0 instead of the more
immature 1.1.

All irrelevant, since my project doesn't use any of that, and has been built and
works. It's just in a maintenance phase now and VS 2003 is fine for that.

So there are other good technical reasons to upgrade. Like it or not, the
reality is that the sun is starting to set on this product, whether or not
it's capable of running on Vista.

Hell, we still have a number of large VC6 projects (incidentally, VC6 _does_ run
under Vista), primarily because of the hassle we had of upgrading to whatever VS
came after that. There's software out in the world that's still running after
thirty years. You don't seem to live in the real world or have any appreciation
of what small teams with limited resources face. You don't _have_ to always have
the latest and greatest to produce software that is useful to people. You are
just making excuses for Microsoft's lack of consideration for its customers. You
must be an MS paid employee. No other type of person on the planet would be
defending it so gallantly.

 

[Brian Muth]

You are

just making excuses for Microsoft's lack of consideration for its customers. You
must be an MS paid employee. No other type of person on the planet would be
defending it so gallantly.

To be clear, I am not an MS employee.

Brian

 

[DavidW]

To be clear, I am not an MS employee.

Okay. My apologies for the accusation.

 

[DavidW]

I personally had no issues upgrading any of my old VS 2003 or VS 2005 projects
to VS 2008.

Well, I've just tried to build it with VS 2008 and, as expected, there are a
million errors I wasn't getting before.

If there are problems, it's likely because VS 2008 is more standards
conformant that VS 2003, and may in fact detect problems with your coding.
Isn't that a good thing?

Nope. Not right now. It's funny that every time you get a new version of
something everything you did on previous versions becomes unacceptable and needs
to be fixed, even though it was fine at the time and the code worked. There's no
telling how long it's going to take just to get VS 2008 to produce an
executable.

 

[Brian Muth]

Well, I've just tried to build it with VS 2008 and, as expected, there are a
million errors I wasn't getting before.


Nope. Not right now. It's funny that every time you get a new version of
something everything you did on previous versions becomes unacceptable and needs
to be fixed, even though it was fine at the time and the code worked. There's no
telling how long it's going to take just to get VS 2008 to produce an
executable.

Would you care to list the first few? I'm sure there are people here willing to help.

Brian

 

[DavidW]

Would you care to list the first few? I'm sure there are people here willing
to help.

Thanks, but I don't need help with the errors at this stage. I know C++ pretty
well and should be able to fix them. For your interest, the most common ones are
C3867, 2065, 4430 and 4867. These are in a library of primitives, so it didn't
attempt to compile any of the thirteen projects that depend on it.

 

[Mark Salsbery [MVP]]

Thanks, but I don't need help with the errors at this stage. I know C++
pretty well and should be able to fix them. For your interest, the most
common ones are C3867, 2065, 4430 and 4867. These are in a library of
primitives, so it didn't attempt to compile any of the thirteen projects
that depend on it.

All those errors are the result of C++ standard compiler conformance,
starting with VS 2005.

Aside from having to use the std namespace where needed, the rest can be
quieted with #pragma options, although making the code compliant is probably
more useful in the long run.

Mark
[/QUOTE]

 

[DavidW]

All those errors are the result of C++ standard compiler conformance, starting
with VS 2005.

Aside from having to use the std namespace where needed, the rest can be
quieted with #pragma options, although making the code compliant is probably
more useful in the long run.

Yes, it would be, but I might go for the #pragmas to get it going. Some of this
code goes back to VC++ 1.0 for Windows 3.1, including this nasty but very useful
macro that sets up for the sending of events to an object of any class,
regardless of the hierarchy it belongs to.

class EventHandler
{
};

typedef void (EventHandler::*CLASS_FUNC)(
UINT eventId, WPARAM wparam, LPARAM lparam);

typedef void (WINAPI *GLOBAL_FUNC)(
UINT eventId, WPARAM wparam, LPARAM lparam);

#define DECLARE_AS_EVENT_HANDLER(className) \
typedef void (className::*FEVENT)(EVENT, WPARAM, LPARAM)

// FEVENT is below context-dependent
#define SET_CLASS_EVENT(manager, eventId, instance, handler, eventType) \
{ FEVENT _fevent = handler; \
CLASS_FUNC _handler; \
*(GLOBAL_FUNC*)&_handler = *(GLOBAL_FUNC*)&_fevent; \
(manager).SetEvent(eventId, (EventHandler*)instance, _handler, eventType); \
}

Windows 3.1 had no threads, so events were needed more than nowadays, and VC++
1.0 had no templates. To do this in a standard-compliant way would probably take
quite a bit of work. The above is post-modification to work on VS 2003. The
original was worse.