Re: [BBv2] fix for response file regeneration

From: Reece Dunn (msclrhd_at_[hidden])
Date: 2005-10-30 01:53:09

David Abrahams wrote:
> Michael van der Westhuizen <r1mikey_at_[hidden]> writes:
>>On 10/29/05, Reece Dunn <msclrhd_at_[hidden]> wrote:
>>[snip]
>>
>>>I have attached the patch for an initial port of native bjam response
>>>file support based on Matt's code. This allows for:
>>
>>[snip]
>>
>>>The patch currently only supports NT (because of getTempPath()), so I
>>>would like help porting to other OSes. Linux would probably be:
>>>
>>> const char * getTempPath( void )
>>> {
>>> return "/tmp/";
>>> }
>>
>>It's normally accepted (in POSIX systems) to allow a user to override
>>the temporary directory using the TMPDIR environment variable. The
>>following code will provide that behaviour, and, from your example
>>above, should just plug in where you need the system specific code.
> 
> It's not much use until we can find a POSIX compiler that supports
> response files, ne?

Currently, the response file logic I have supplied is in variable.c and 
this makes use of getTempPath() (that I have made global to share 
between execnt.c and variable.c). Are you suggesting that (at the 
moment) they should only be available on NT/CYGWIN (or alternatively, if 
OPT_RESPONSE is defined). If so, isn't that going to make some of the 
language conditional?

Q: Does intel-linux have response file support?
Q: Does CodeWarrior for the Mac have response file support?

- Reece