Re: [boost] [build] Default compiler options for a toolset

Subject: Re: [boost] [build] Default compiler options for a toolset
From: Edward Diener (eldiener_at_[hidden])
Date: 2013-11-21 20:40:08

On 11/21/2013 3:43 PM, Lars Viklund wrote:
> On Thu, Nov 21, 2013 at 03:28:26PM -0500, Edward Diener wrote:
>> On 11/21/2013 11:07 AM, Lars Viklund wrote:
>>> On Wed, Nov 20, 2013 at 05:33:33PM -0500, Edward Diener wrote:
>>>> I disagree. When/if clang changes the particular jam file should change
>>>> appropriately.
>>>
>>> What about people that must, want to, or need to use a Boost version
>>> released with the proposed /EH muting in effect together with a future
>>> clang-cl compiler that actually implements /EH.
>>
>> Why in the world would I add a compiler option to a compiler which does not
>> support it just in case that compiler might support that compiler option in
>> the future ?
>
>  From the sounds of it, this seems to be more about actively removing
> things that are already in place.
>
> And as for "might", I'd bet three pinecones on that it is actually
> "will", unless some patent or other fundamental stupidity _prevents_
> proper EH/RTTI in future Clang.
>
> To be honest, you're weighing some minor warning spam from the compiler
> against forever making a release unusable with future versions of that
> compiler.
>
> Sadly I don't get to make the choice, but I sure will say my opinion on
> things that look intrinsically odd.
>
> (I'd like my bikeshed purple, thank you.)

Please go to the e-mail called "Clang windows toolset" on the Boost 
Build mailing list and download the attached files there by 'tr1gun' and 
do whatever you wish with them. I am just using those jam files locally 
in order to test clang-cl on Windows using Boost Build for myself. You 
can do the same also if you desire.

Given that the Boost Build mavens had no interest, and may still have no 
interest, in a toolset for clang-cl, I am just thankful to be using the 
product of another person's efforts to achieve my goal of testing 
clang-cl with Boost and fixing problems in Boost that will make it 
easier for Boost end-users to use clang-cl if Boost Build ever 
officially supports it. I do not know who 'tr1gun' is but am thankful 
for his efforts. What I do locally in order to correctly test clang-cl 
with Boost Build is my own business, while whatever fixes I make to 
Boost for clang-cl in various Boost libraries is Boost's business. But I 
have made no effort to promote the jam files created by 'trgun' in any 
official way as a part of Boost or Boost Build.