[boost] Re: Comeau status/advice?

From: David Abrahams (dave_at_[hidden])
Date: 2003-09-23 06:05:17

"Misha Bergal" <mbergal_at_[hidden]> writes:

> "Mat Marcus" <mat-boost_at_[hidden]> wrote in message
> news:32551045.1064251288@[172.25.40.82]...
>
>> I read through the archives for awhile and did some tests with bjam
>> trying to build boost with the latest Comeau compiler (433b8) with VC
>> 7.1 back end. I tried "-sTOOLS=como" and "-sTOOLS=como-win32". The
>> former didn't work at all (bad option -OO) and the latter produced
>> quite a few errors in regex, signals, data, etc. What I'd really like
>> to do is build for como43x/VC7.1 under cygwin, but I'd settle for it
>> working under the NT shell. Before I spend much more time on this I
>> wanted to ask if anyone has already succeeded at it. In particular,
>> the metacomm regression results seem quite a bit better than what I am
>> getting. Any advice?
>
> Mat,
>
> We are using 4.3.2beta and our own local toolsets. See
> http://tinyurl.com/obj0 (patch to como-win32-tools.jam)  and
> //tinyurl.com/obos extension toolset
> You will have to apply como-win32-tools.jam.diff  patch to file
> como-win32-tools.jam
> in your tools/build directory and just copy meta-como-4.3.2b-vc7-tools.jam
> to tools/build.
>
> Invoking bjam -sTOOLS=meta-como-4.3.2b-vc7 should give you what you want.
>
> 4.3.3b8 has a bug in como.exe which prevents it from successfully building
> most of
> the boost. I am working with Boost.Build people and Greg Comeau on that,
> and I think it is going to be resolved Very Soon Now.

Do you know what the status is with being able to handle __declspec()
in a more standard-conforming mode than the one with all the
microsoft bugs in it?  I asked Greg, but got no answer.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com