Jan, Walter

Sorry if this appears twice.

Just a followup on the contents of the download page with regards to STL

I have extracted the ASCII from the page and put line numbers on the relevant section for easy of reference.


1   o Hewlett Packard (312,667 bytes)
2
3   o Silicon Graphic	International (3.11) (As far as I got it done until 08/23/1998) (238,828 bytes)
4
5   o Silicon	Graphic International (3.20) (As far as I got it done until 09/08/2000) (241,618 bytes) (NEW!)
6
7   o STLport (4.0) (Thanks to Damian Dixon) (NEW)
8
9   Warning! SGI STL and STPport will not (yet) compile with compiler version 8.0 and later as this compiler implements __out as keyword!
10
11
12 The following files are in .tar.gz format. The are a checkout of the CVS repository. Every day at 00:00a checkout is being performed. These files are provided for the users that do not
have CVS access.
13
14  o Hewlett Packard
15
16  o Silicon Graphics International (3.20)
17
18  o STLport (4.0)


The statement on line 9 is correct for lines 5 and 7. It is also probably correct for lines 1, 3 and 5 but I have not tried these.

Lines 1 to 9 should be moved to some sort of Old Un-supported section.

The following should be added in place of lines 1 to 10. Lines 3 and 5 need to be confirmed or removed.


1  The following STL snapshots have been confirmed to work with Digital Mars C++ version 8.1B2
2
3    o Hewlett Packard
4
5    o Silicon Graphics International (3.20)
6
7    o STLport (4.0) (limied to STL part of STLport only)
8
9
10


I believe we should keep the CVS checkout for those that do not have access to CVS. So lines 11 to 18 should be kept.

I hope this clears things up. I should have been more explicit in my previous postings.

I will be looking at porting 'STLport 4.1b5' as soon as I have some spare time. I need to start again to ensure that all the changes ncessary to support DMC++ get feed back into STLport officially, so I also must ensure the number of changes are kept very low.

If you need to discuss this in more detail or require a slightly faster update rate please email me directly at 'damian.dixon@tenetdefence.com'

Regards,
Damian

On Wed, 28 Mar 2001 15:05:36 GMT, Damian Dixon <damian.dixon@tenetdefence.com> wrote:
> On Tue, 27 Mar 2001 03:04:49 -0500, Jan Knepper <jan@smartsoft.cc> wrote:
> > STLport as well for HP's, SGI's and MFC and ATL have been tagged with a
> > "RELEASE" tag.
> > The nightly checkout's (now exports) use this "RELEASE" tag.
> > So the .tar.gz files available on the web site are not the latest snapshot
> > anymore as these might not be STABLE.
> > 
> > Hope this solves the problems
> 
> Should do.
> 
> However you may wish to provide people with access to the bleeding edge :>
> 
> Damian
> > 
> > Jan
> > 
> > 
> > 
> > Damian Dixon wrote:
> > 
> > > Jan,
> > >
> > > Ummmm.....
> > >
> > > The STL part of STLport works with DMC++ v8.12. The other bits are still un-supported.
> > >
> > > This means nightly snapshot should be upto date.
> > >
> > > The comments on STLport4.0 not working with DMC++v8.12 is not quite true. That particular shapshot will not work with the latest compiler.
> > >
> > > We need to take a snapshot and mark that as working with DMC++ v8.12, and mark the others as obsolete or remove them.
> > >
> > > I don't think we should expect people to use the snapshot out of CVS, as there is a remote possability that it may get broken :>
> > >
> > > Did you find the ANSI documents?
> > >
> > > Regards,
> > > Damian
> > >
> > > Jan Knepper <jan@smartsoft.cc> wrote in message news:3ABF7434.224D69E@smartsoft.cc...
> > > > What is wrong with it?
> > > >
> > > > Jan
> > > >
> > > >
> > > >
> > > > Walter wrote:
> > > >
> > > > > Hi Damian - It looks like the download page for STLport is not right.
> > > Can
> > > > > you let me know what it should be, and I'll fix it?
> > > > >
> > > > >     -Walter
> > > > >
> > > > > Damian Dixon wrote in message <1103_985016466@dilbert>...
> > > > > >The STLport cvs repository has been updated to work
> > > > > > with DMC++ version 8.1.
> > > > > >
> > > > > >The changes relate to '__i' being used by the compiler
> > > > > >to denote the imaginary number 1.
> > > > > >
> > > > > >I have run the regression tests and can see no problems with the generated output (please note that regression tests are limited in scope).
> > > > > >
> > > > > >If you do not have access to CVS you will need to wait
> > > > > >until the tar files are updated on the server (please
> > > > > >refer to the notes on the web site for update times).
> > > > > >
> > > > > >- Damian
> > > > > >
> > > > > >
> > > >
> > 
> 
>