Re: [Boost-testing] metacomm Digital Mars results

From: Arjan Knepper (arjan.knepper_at_[hidden])
Date: 2006-04-12 05:38:28

Vladimir Prus wrote:
> Arjan wrote:
>
>   
>> Vladimir Prus wrote:
>>     
>>> Hi,
>>> the results for the Digital Mars compiler from metacomm are pretty bad at
>>> the moment -- just 5 libraries show up green. I doubt that this sitation
>>> can change unless serious effort is spend both on Boost side and on
>>> compiler side, and it does not seem like this effort is happening.
>>>
>>> So: maybe we should stop testing that compiler?
>>>
>>> - Volodya
>>>       
>> No please don't. I (and other people in the DMC camp) just started to
>> look at some of the problems with DMC.
>> Please have a little more patience.
>>     
>
> Ok, if you're starting working on this, guess we can wait. However, please
> start with Boost.Build V2 right away, and try testing the existing V2
> toolset as documented at 
>
>    https://zigzag.cs.msu.su/boost.build/wiki/TestingNewToolset
>   
Thanks for the pointer.
> including full regression tests for Boost.Build. Last time it was tried we
> run into this:
>
>    http://lists.boost.org/boost-testing/2006/03/2702.php
>
> which sounds like some issues with creating DLLs with DMC. If you can look
> into this that would be great.
>   
Will do, but at first I will stick to the static builds and linking. 
When that is in a reasonable state I will start with the dynamic build 
and linking.

BTW AFAIK the free version also allows building dll's and IIRC def files 
are not really required it is possible the specify the setting(s) on the 
command line.

Arjan