Re: [boost] Where to put support for missing C++ C headers?

From: David Abrahams (abrahams_at_[hidden])
Date: 2001-04-13 17:22:56

Everyone,

I do not want to leave Ralf hanging, but I'm not sure what to tell him. He
is making an important contribution to Boost.Python, and the generation of
missing <cxxx> headers is a useful contribution to boost overall, consistent
with our mission. Does anyone have an idea about this?

Perhaps we should start with a web page on library compatibility, with links
to the new <limits> file for GCC and to Ralfs scripts/generated files...

Thanks,
Dave

----- Original Message -----
From: <rwgk_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Thursday, April 12, 2001 8:34 PM
Subject: [boost] Where to put support for missing C++ C headers?

> I have a small Python script that generates the missing
> C++ C headers for a few platforms (Compaq Alpha cxx,
> both Linux & Tru64, Silicon Graphics IRIX CC). I am
> routinely using the headers generated by the script to
> compile the Boost.Python library. The makefiles in
> boost/python/libs/build rely on the headers being present.
>
> Is there a good place somewhere in the boost tree to
> put the Python script? An ad-hoc approach would be to
> put the script into boost/python/libs/build, but there
> it would be somewhat hidden. A more prominent location
> would be useful since the headers could also be used
> with other boost libraries.
>
> Alternatively, would there be a good place to put the
> results of running the script, i.e. the actual headers,
> instead of the script itself? What about boost/C++_C_headers
> or boost/fixes/C++_C_headers? -- I do not really like
> this idea myself, though.
>
> I would appreciate some guidence.
>
> Ralf