Digital Mars - c++ - DM8.45 and boost

↑ ↓ ← → "yxh" <yangxh98 hotmail.com> writes:

Perhaps nobody interested in boost here!

"John Fletcher" <J.P.Fletcher aston.ac.uk> 
??????:43314771.692E79FA aston.ac.uk...
 What is the status of DM8.45 and the boost library?

 I have checked up on the boost site and there is some information about
 DM8.44 in relation to boost 1.33.0

 Thanks

 John

Sep 22 2005

↑ ↓ ← → "Bob Paddock" <graceindustries gmail.com> writes:

On Thu, 22 Sep 2005 08:32:06 -0400, yxh <yangxh98 hotmail.com> wrote:

 Perhaps nobody interested in boost here!


I am.

 What is the status of DM8.45 and the boost library?

 I have checked up on the boost site and there is some information about
 DM8.44 in relation to boost 1.33.0




What is the URI for that info, I'd like to take a look?

I have not tried to build Boost with DM since V8.40 and 1.30.
It was a nightmare and I gave up, and switched to BCC.

I'd switch back to DM if someone can tell me that the current
Boost builds out of the box, with DM8.45.

Sep 22 2005

↑ ↓ ← → John Fletcher <J.P.Fletcher aston.ac.uk> writes:

Bob Paddock wrote:

 On Thu, 22 Sep 2005 08:32:06 -0400, yxh <yangxh98 hotmail.com> wrote:

 Perhaps nobody interested in boost here!


 I am.


There was work done before.

 What is the status of DM8.45 and the boost library?

 I have checked up on the boost site and there is some information about
 DM8.44 in relation to boost 1.33.0




 What is the URI for that info, I'd like to take a look?


http://engineering.meta-comm.com/boost-regression/1_33_0/user/

(It is a link from http://www.boost.org/ near the top righthand corner)

Not "out of the box" yet.

 I have not tried to build Boost with DM since V8.40 and 1.30.
 It was a nightmare and I gave up, and switched to BCC.

 I'd switch back to DM if someone can tell me that the current
 Boost builds out of the box, with DM8.45.


John

Sep 22 2005

↑ ↓ ← → "Bob Paddock" <graceindustries gmail.com> writes:

On Thu, 22 Sep 2005 11:27:52 -0400, John Fletcher  
<J.P.Fletcher aston.ac.uk> wrote:

 What is the URI for that info, I'd like to take a look?


 http://engineering.meta-comm.com/boost-regression/1_33_0/user/


http://engineering.meta-comm.com/boost-regression/1_33_0/user/summary.html

Shows only four passed out of about fifty test for DM8.44, pretty dismal.

Sep 22 2005

↑ ↓ ← → Daniel James <daniel calamity.org.uk> writes:

Bob Paddock wrote:
 
 http://engineering.meta-comm.com/boost-regression/1_33_0/user/summary.html
 
 Shows only four passed out of about fifty test for DM8.44, pretty dismal.


That page is only an overview, and makes DMC look much worse because no 
one has annotated the failures on digital mars. For example see the 
detail for type_traits:

http://engineering.meta-comm.com/boost-regression/1_33_0/user/type_traits.html

Dmc passes more tests than vc-7, but on the summary vc-7 is green, dmc 
is yellow (I'm just comparing those two because they're adjacent).

Also, many of the failures are down to problems with Boost.Test or 
problems with the build system. Not the actual library itself. As far as 
dmc is concerned that page is meaningless.

It's difficult to get dmc working with the build system though, 
especially if you don't have the CD.

Daniel

Sep 22 2005

↑ ↓ ← → "Walter Bright" <newshound digitalmars.com> writes:

"Daniel James" <daniel calamity.org.uk> wrote in message
news:dgvdgi$250p$1 digitaldaemon.com...
 It's difficult to get dmc working with the build system though,
 especially if you don't have the CD.


Why? DMC is exceptionally easy to install/use. In fact, it doesn't even need
to be "installed", as it doesn't muck about with the registry or environment
variables.

Sep 22 2005

↑ ↓ ← → Daniel James <daniel calamity.org.uk> writes:

Walter Bright wrote:
 "Daniel James" <daniel calamity.org.uk> wrote in message
 news:dgvdgi$250p$1 digitaldaemon.com...
 
It's difficult to get dmc working with the build system though,
especially if you don't have the CD.


 
 Why? DMC is exceptionally easy to install/use. In fact, it doesn't even need
 to be "installed", as it doesn't muck about with the registry or environment
 variables.


Oh, I wouldn't disagree with that at all - most of what I've done so far 
has been done via the compile line or using make and it's worked fine. 
But Boost.Build is a bit odd. It's designed for using multiple compilers 
and mulitple configurations simultaneously. So it tries to control 
everything through the command line - such as include paths, libraries 
linked, etc. The default setup tries to use versions of the libraries 
not supplied by the download (eg. by default it uses the dynamic version 
of the libraries).

There's probably a combination of Boost.Build options that will get it 
to work with the download, but I haven't really tried to find them yet. 
It doesn't help that Boost.Build works very badly with Windows 9x (I'm 
mainly a linux user, so I never upgraded). Another part of the problem 
is that I'm not sure what is causing the link errors in the test 
results. Such as: http://tinyurl.com/c8tdo. Hopefully over the next week 
or two I'll get a better idea.

Boost.Build v1 has been due to be replaced with v2 for a long time - I 
had been planning to look into this when this happens, as v2 is a lot 
easier to use. But the transition never seems to happen.

Daniel

Sep 23 2005

↑ ↓ ← → John Fletcher <J.P.Fletcher aston.ac.uk> writes:

Bob Paddock wrote:

 On Thu, 22 Sep 2005 11:27:52 -0400, John Fletcher
 <J.P.Fletcher aston.ac.uk> wrote:

 What is the URI for that info, I'd like to take a look?


 http://engineering.meta-comm.com/boost-regression/1_33_0/user/


 http://engineering.meta-comm.com/boost-regression/1_33_0/user/summary.html

 Shows only four passed out of about fifty test for DM8.44, pretty dismal.


I was hoping that 8.45 would be better.  Perhaps there are a few reasons why a
lot fail.

John

Sep 23 2005

↑ ↓ ← → "Walter Bright" <newshound digitalmars.com> writes:

"Bob Paddock" <graceindustries gmail.com> wrote in message
news:op.sxiarkc5ocfjro grace002.graceinc.com...
 I have not tried to build Boost with DM since V8.40 and 1.30.
 It was a nightmare and I gave up, and switched to BCC.


It's likely because many have made the effort to provide workarounds in
boost for BCC, not that BCC is more compatible.

Sep 22 2005