This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Succeeded with Alpha+19980803, some feedback



  In message <2.2.32.19980817174913.00e851a0@po9.mit.edu>you write:
  > At 11:06 AM 8/17/98 -0600, you wrote:
  > >What references to /usr/local exist in your gcc Makefiles?
  > 
  > I've hunted around, and the only interesting use I could find
  > (ie: perform a search via find+grep and ignore documentation
  > files and source files that mention /usr/local) was in
  > $objdir/gcc/Makefile, line# 262:
  > 
  >    local_prefix = /usr/local
Yup.

  > which gets used in $objdir/gcc/Makefile, line# 278:
  > 
  >    includedir= $(local_prefix)/include
Got it.

  > 
  > which is then used in the attempt to create the directory
  > in line# 3402.  There is no /usr/local/include directory
  > on the Alpha I'm working on, so the test beginning
  > 
  >   "-if [ -d $(includedir) ] ..."
  > 
  > would fail, causing the Makefile to try and create a
  > directory that it won't be using. 
It won't put anything in there for a native configuration.  That
directory may get populated in some cross compilation environments
by other tools (I'm not 100% sure).

  > Probably its just the attempt to create the directory that
  > is bogus (and it fails without aborting the Makefile anyways),
  > and there is no other particular misbehaviour.
Actually I don't think trying to create the directory is necessarily
bogus.  And I think we both agree that nothing bad happens.

Thanks for investigating this.

jeff


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]