[Rd] grep in version 1.8 (PR#4231)

[Paul Gilbert]

Friedrich.Leisch at ci.tuwien.ac.at wrote:

>>>>>>On Mon, 22 Sep 2003 06:07:00 -0400,
>>>>>>Duncan Murdoch (DM) wrote:
>>>>>>            
>>>>>>
>
>  > On Sun, 21 Sep 2003 16:58:10 -0500, you wrote:
>  >> On Sunday 21 September 2003 16:50, Duncan Murdoch wrote:
>  >>> On Sun, 21 Sep 2003 11:23:57 +0200 (MET DST), you wrote:
>  >>> >This is not a bug! It works when compiling from clean sources.
>  >>> >
>  >>> >I guess you have unpacked new sources over old sources? Please clean the
>  >>> >directory before unpacking a new version, and try to compile again. A
>  >>> >couple of us ran in the same problem during the last weeks.
>  >>> 
>  >>> I don't know if this is what happened here, but this kind of thing
>  >>> also happens when using rsync, not just when unpacking tarballs:
>  >>> rsync won't delete obsolete files.  Anonymous cvs wouldn't have this
>  >>> problem.
>  >> 
>  >> rsync should delete files if you give the --delete flag
>
>  > Yes, but then you end up doing a clean build, because it will also
>  > delete the *.o files, etc.  I don't think there's a concept of
>  > deleting files that were there yesterday but aren't there today.
>
>Yes, there is:
>
>1) Use the --delete flag when updating the sources
>
(From another strand of this thread)  I don't think the -P option to the 
cvs update command does this, it just prunes empty subdirectories, and 
-d does roughly the opposite. If cvs has a way to delete files that are 
not in the archive I have never found it, and would be curious to know 
what it is so that I can be careful to avoid it.

>2) Use a build directory that is different from your source tree, that
>   way rsync will never see the *.o et al files.
>
This is probably the answer, but I don't understand how it solves what I 
think was the original question. Could someone please expand on
 
  i/  How make will know to ignore certain .o files which should not be 
there, but leave other ones that are up-to-date and not rebuild them.

  ii/  How to build R in a different  directory from the sources.

Thanks,
Paul Gilbert