Damian Dixon wrote in message <1107_979228530@dilbert>...
>Jan, Walter, When do you plan to improve the template support in the
compiler?


I do hear and understand the problems you're having with templates and namespaces.

The template support in the compiler is one of the trickiest and most complex pieces of code in it (aside from multiple inheritance!). A quick hack isn't going to do the job and will likely make things worse - not only that, doing it right may require some significant work on the linker.

My main focus for the near term is to get the whole package into a usable form (there's some legal footwork to be done, too), so the template support will not likely change in the near future.

I'm also facing the reality that it's just myself & Pat working on this, so we'll have to be a bit ruthless in narrowing the scope of the project. We'll likely not be making any changes at all to the IDE nor the 16 bit support, and will focus on the command line tools and win32 support.

I appreciate all the work everyone here in the newsgroup is doing to produce workarounds for the template issues.

Wise.

Ed

Walter wrote:

> I've found it the most productive to develop under NT or 2000, and then port it to 98/95. The reason is that the former is a much more robust development environment (a lot fewer reboots needed).

Isn't it amazing!
I *never* reboot Unix (actually FreeBSD) on which I also do development by the
way. (Check the system status at http://www.digitaldaemon.com/)
But I agree, Windows NT or 2000 is a huge improvement over 95 or even 98. I
never changed to 95 or 98 since I was running NT before that. Never regret it
for one second...

Jan

Jan Knepper wrote in message <3A5E83CA.1EEFD54B@smartsoft.cc>...
>Walter wrote:
>
>> I've found it the most productive to develop under NT or 2000, and then
port
>> it to 98/95. The reason is that the former is a much more robust
development
>> environment (a lot fewer reboots needed).
>
>Isn't it amazing!
>I *never* reboot Unix (actually FreeBSD) on which I also do development by
the
>way. (Check the system status at http://www.digitaldaemon.com/)
>But I agree, Windows NT or 2000 is a huge improvement over 95 or even 98. I
>never changed to 95 or 98 since I was running NT before that. Never regret
it
>for one second...
>
>Jan


I've had to reboot my linux box a few times, but the people in the linux newsgroups called me a liar because linux *never* freezes up <g>.

As Clint Eastwood put it, "a man's got to know his limitations!" <grins>


Edward F. Sowell wrote in message <3A5E8175.5BA87BB9@home.com>...
>Wise.
>
>Ed
>

On Thu, 11 Jan 2001 14:26:48 -0800, "Edward F. Sowell" <sowelled@home.com> wrote:
> 
> 
> Damian Dixon wrote:
> 
> >
> >
> > Is the browser type being used it a large number of places?
> 
> No. Seems to be used only once!
> 
> > or just a few? If in a few places then
> > replacing the use of browser should not be too difficult. It may be fairly simple to move the functionality of browser
> > into the TVector template, with minimal impact. I've done some modifications that eleminate the use of the
> > browser template, though I have not tested them to see if the functionality is the same.
> >
> > Sorry I could not be of any more help. If you want the changes I made email me and I will send them to you.
> 
> Yes, please send. I will try to continue the work and ask for help if needed.
> 
Sent......

> Thanks.
> 
> Ed
> 

On Thu, 11 Jan 2001 18:49:57 -0800, "Walter" <walter@digitalmars.com> wrote:
> 
> Damian Dixon wrote in message <1107_979228530@dilbert>...
> >Jan, Walter, When do you plan to improve the template support in the
> compiler?
> 
> 
> I do hear and understand the problems you're having with templates and namespaces.
> 
> The template support in the compiler is one of the trickiest and most complex pieces of code in it (aside from multiple inheritance!). A quick hack isn't going to do the job and will likely make things worse - not only that, doing it right may require some significant work on the linker.
> 

I understand. I had looked at doing a pre-process jog on templates into a supported form for partial specialisation etc.... but decided for the moment it was too large a job given my other commitments :(

> My main focus for the near term is to get the whole package into a usable form (there's some legal footwork to be done, too), so the template support will not likely change in the near future.
> 

Thanks understandable.

> I'm also facing the reality that it's just myself & Pat working on this, so we'll have to be a bit ruthless in narrowing the scope of the project. We'll likely not be making any changes at all to the IDE nor the 16 bit support, and will focus on the command line tools and win32 support.
> 

win32 is probably the thing to concentrate on. I can live with IDE and 16 bit support as they are.

> I appreciate all the work everyone here in the newsgroup is doing to produce workarounds for the template issues.
> 
>

Is there anything the rest of us could do to help?

On Thu, 11 Jan 2001 13:11:43 -0500, Jan Knepper <jan@smartsoft.cc> wrote:
> Damian Dixon wrote:
> 
> > With all these compilers we have found that template support varies greatly to such an
> > extent that something that worked with VC++ v6 failed on nearly all the Unix platforms we need to
> > support, so we have had to re-write a large chuck of code (template meta programming). So we now
> > keep our templates fairly simple.
> 
> Cool huh?!
> 
Not so cool, we found that many of our developers and very very few customers could understand what was going on. The only benefit we had was performance, but we also got code bloat!

So for sanity and for porting we removed the code.

Regards,
Damian

Damian Dixon wrote:

> On Thu, 11 Jan 2001 13:11:43 -0500, Jan Knepper <jan@smartsoft.cc> wrote:
> > Damian Dixon wrote:
> >
> > > With all these compilers we have found that template support varies greatly to such an
> > > extent that something that worked with VC++ v6 failed on nearly all the Unix platforms we need to
> > > support, so we have had to re-write a large chuck of code (template meta programming). So we now
> > > keep our templates fairly simple.
> >
> > Cool huh?!
> >
> Not so cool, we found that many of our developers and very very few customers could understand what was going on.

Know exactly what you mean...
Although, my customers NEVER look into my code.

> The only benefit we had was performance, but we also got code bloat!

Hmmm.

> So for sanity and for porting we removed the code.

For this reason I never use any STL template directly, but always derive a protected class and implement an interface. This way, if I ever go out of luck on a platform or compiler I will have to implement the class myself without the template, but it will work.

Jan

Walter wrote:

> I've had to reboot my linux box a few times, but the people in the linux newsgroups called me a liar because linux *never* freezes up <g>.

<grin>
The truth is, as far as I know, that Linux is more on the bleeding edge of
technology than BSD. Drivers for new hardware are mostly implemented in Linux
first and than (quite) a bit later BSD follows.

I would not say that I would not know how to crash BSD, but it has been proven
that BSD is more stable, secure and has a more crash proof file system. BSD
seems to have the fastest TCP stack which is great with all the current DoS
attact that happen on the Internet these days.
I could show you some log files. You would be amazed how many port scans and
other things happen. One of the cutest is IP packages comming in via the
external interface that claim to come from 192.168.0.0/16, 172.16.0.0/12 or
10.0.0.0/8... These addresses are for private networks defined so by RFC 1918,
which means they legally NEVER can come from the Internet and thus never should
come in via the external interface of the system.

Jan