incubator-stdcxx-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Martin Sebor <se...@roguewave.com>
Subject Re: r. 380238 broken
Date Fri, 24 Feb 2006 00:52:11 GMT
Liviu Nicoara wrote:
> I meant using a custom separator (instead of space) in the list of
> functions you test in libc_decl.sh.

But then we wouldn't be able to take advantage of shell tokenization,
would we? But I don't think quoting is going to do it, either. What
a pain! I think I just fixed it with the latest commit.

Martin

> 
> Liviu
> 
> Martin Sebor wrote:
> 
>>>Liviu Nicoara wrote:
>>>Perhaps a custom separator would be more beneficial in the long run?
>>
>>I'm not sure what you mean?
>>
>>
>>>>>It seems I rushed to a wrong conclusion - I haven't noticed the
>>>>>definitions for _RWSTD_INSTANTIATE_type's macros. Scratch the diff since
>>>>>it leads to unsats because of the lack of type instantiations in the
>>>>>library. Martin, any off-the-shelf ideas as to why I get this:
>>>>
>>>>
>>>>It's caused by this change:
>>>>http://svn.apache.org/viewcvs.cgi?rev=379694&view=rev
>>>>
>>>>The space between const and the type is causing problems (the bogus
>>>>macro _RWSTD_NO_CHAR ends up getting defined to nothing which makes
>>>>the the preprocessor unhappy when it sees it in conditionals like
>>>>
>>>>   #if !_RWSTD_NO_CHAR
>>>>
>>>>I need to fix it by some clever application of shell quoting (ugh).
>>>>
>>>>Incidentally, this also means that the same breakage will occur
>>>>when _RWSTD_NO_WCHAR_T is #defined. That will need to be fixed in
>>>>_defs.h, maybe along these lines:
>>>>
>>>>   #if !(_RWSTD_NO_WCHAR_T + 0)
>>>>
>>>>Martin
>>>>
>>
>>


Mime
View raw message