[1831] in Kerberos_V5_Development
Re: make vs. gmake and the Athena source tree
daemon@ATHENA.MIT.EDU (Sam Hartman)
Mon Sep 30 15:34:52 1996
To: Greg Hudson <ghudson@MIT.EDU>
Cc: Marc Horowitz <marc@MIT.EDU>, krbdev@MIT.EDU
From: Sam Hartman <hartmans@MIT.EDU>
Date: 30 Sep 1996 15:33:57 -0400
In-Reply-To: Greg Hudson's message of Mon, 30 Sep 1996 14:46:39 EDT
Just in case someone decides to use something along the lines
of Greg's suggestion, they should fix the following problems first:
* You don't always need to build shared objects on platforms without a
concept of PIC code.
* We need the flexibility to only build shared libraries on AIX.
* On some platforms, you need to specify the version part of the name
as a flag to LD; that probably isn't too difficult.
* That doesn't work with platforms where you want to build .obj
objects instead of .o objects.
Besides that, it is workable for TCL. I'm not convinced that
when you add the cruft to deal with sources in multiple directories,
etc, etc, that it will be particularly workable. However, I will
admit that Greg's objections to gmake are fairly compelling.
--Sam
