|
Boost : |
From: Beman Dawes (bdawes_at_[hidden])
Date: 2002-08-20 10:29:00
At 10:01 AM 8/20/2002, Gennadiy Rozental wrote:
>"Beman Dawes" <bdawes_at_[hidden]> wrote in message
>news:4.3.2.7.2.20020820085314.023cc928_at_mailhost.esva.net...
>> At 04:41 AM 8/20/2002, Gennadiy Rozental wrote:
>>
>> >> Well, that isn't quite right. "inline" isn't 100% right either,
but
>it
>>
>> >is
>> >> seems better than "online" to me.
>> >I think that inline would be misleading (at least for the
Test/Program
>> >Execution Monitor headers ) since in fact library is implemented
>offline.
>> >Maybe we could separate it into
>> >
>> >boost/test/online/test_execution_monitor.hpp
>> >boost/test/online/prg_execution_monitor.hpp
>> >and
>> >boost/test/minimal.hpp or boost/test/lightweight.hpp ?
>>
>> I think changing the directory name to "basic", "lite", "lightweight",
or
>> similar would be better:
>>
>> boost/test/basic/...
>
>But there is nothing basic or lite in test_execution_monitor.hpp and
>prg_execution_monitor.hpp. It's simply inlined through direct source
>inclusion of all sources.
Sorry, I lost the thread. I thought we were talking about the basic, V1,
functionality versions.
Anyhow, "online" doesn't sound right to me. "inline" isn't really right
either, because everything in those headers isn't inline, IIRC.
Since the concept to be captured is "direct source inclusion", maybe you
should consider a sub-directory name like "src" (or "source"), and consider
changing the names to .cpp. That was what V1 did to indicate these were
not normal header files, but source files being directly included.
--Beman
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk