Re: Re: [boost] Comeau setup

From: gclbb-jamboost_at_[hidden]
Date: 2002-07-07 11:35:18

In article <20020707000102-r01010800-8bf4f5ac-0860-0108_at_12.100.89.43>,
Rene Rivera <jamboost_at_[hidden]> wrote:
>[2002-07-06] gclbb-jamboost_at_[hidden] wrote:
>>In article <20020706132802-r01010800-a585b50a-0860-0108_at_12.100.89.43>,
>>Rene Rivera <jamboost_at_[hidden]> wrote:
>>>[2002-07-06] Greg Comeau wrote:
>>Now that this is in the right NG, I'll just restate that the
>>compiler status info at boost site has reported of Comeau C++
>>for LINUX for quite some time now (I believe by Jens).
>
>Now that I've looked at the status stuff :-) There is an easy answer to the
>dilema... The regression status pages don't use Boost.Build to compile the
>tests. It uses direct commands, as stated in the .../status/compiler.cfg
>file. I think some of the regression testing is going to change soon but I
>don't know the details there.

So, you're saying that *NO* compile is done for the regression
status pages using bjam? It looks like some something reads
that .cfg file???? Guess I'll be snooping around that directory.

>>>>* I'm not sure what http://www.boost.org/tools/build/como-tools.html
>>>> represents...
>> "The como toolset responds to the following variables which
>> can be set in the environment"
>Somewhat of a misleading statement :-) The clarification...
>"como toolset" == "Boost.Build support for Comeau C++"
>
>The confusing part is that those variables can be set either directly or in
>the environment. And even more confusing is that we try and use environment
>variables as pre-configured normally by a user of a given toolset.

As I understand the purposes of the different part, this will no
doubt sink in.

>>>> Also, is this bjam'ing procedure different for different platforms,
>>>> or does bjam recognize what platform it is on? (IOWs, invoking
>>>> como can be slightly different for different platforms, and so I'm
>>>> unclear how that's recognized programmatically, or described in your
>>>> documentation.)
>>>...
>
>* I looked at the como toolset a bit more, and it's definately as specific
>in support as the description says. It only works in Win32/VC.

What does 'if $(NT)' check for?

>* We can tell which platform we are running in, and therefore adjust the
>compiler invocations accordingly. But again that's not done in the current
>como toolset support.

Right, and I see for Intel that that wasn't done either, and instead
a seperate file was used. ???

>* I'd be willing to add support for more of the Comeau backends, but I'd
>need help in doing so. Specifically I'd need:
>
> 1. Documentation for the compiler options based on which backend. I
>haven't looked at the documentation you have on the website but hopefully
>the info is in there.
> 2. Someone(s) who is willing to test the new toolset. This someone(s)
>would need to know how to access the Boost CVS files. And compile on the
>various platforms we want to support.

I suspect that the current info is wrong and/or superfluous,
so I'd rather get that changed first. And at that point I'll
probably understand it enough to be able to make suggestions
for the new backends.

As such, let me start with a question: I'm unclear as to
the difference between COMO_PATH, COMO_BIN_PATH, COMO_BIN_DIRECTORY,
COMO_BASE_SETUP, and COMO_PATH_SETUP and how they [need to] interact
and/or how they are [necessary to be] different from just
PATH and COMO_PATH?
And also what program uses these "new" variables, and why?
-- 
Greg Comeau 4.3.0 NEWS: New Windows Backends + 'export' IN July!
Comeau C/C++ ONLINE ==> http://www.comeaucomputing.com/tryitout
World Class Compilers: Breathtaking C++, Amazing C99, Fabulous C90.
Comeau C/C++ with Dinkumware's Libraries... Have you tried it?