Nope, CPPFLAGS is for the c preprocessor (all your -D defines, -I directories, etc.). CFLAGS is for the actual C compiler. Granted, most of the time both of those steps are done in one pass, but this change makes it so someone can override CPPFLAGS, say, to just add another -D define without messing with the CFLAGS for optimization settings or anything. I almost went through the same steps for LDFLAGS (for the linker), but most people don't mess with those much. In my case, I was just be anal and compiling everything in sight with -O3 -mcpu=athlon =march=athlon to see what broke :-) (so far, nothing has, it all works, even GCC 3.0.4 itself, amazing :-) ----- Original Message ----- From: "Mark D. Studebaker" <mds at paradyne.com> To: "Kevin P. Fleming" <kevin at labsysgrp.com> Cc: <sensors at stimpy.netroedge.com> Sent: Saturday, April 27, 2002 9:32 PM Subject: Re: Patch for lm_sensors-2.6.3 Makefile to support user specified CFLAGS and CPPFLAGS > good idea - but isn't CPPFLAGS for c++? > Why do we need both CFLAGS and CPPFLAGS? > > "Kevin P. Fleming" wrote: > > > > The attached patch allows the user to add CFLAGS and CPPFLAGS entries (via > > the environment or the make command line) and have them actually get used > > compiling the lm_sensors objects... I used this because I wanted to try > > other optimization settings without massaging the Makefile every time. > > Enjoy! > > > > > > Name: lm_sensors-Makefile.patch > > lm_sensors-Makefile.patch Type: unspecified type (application/octet-stream) > > Encoding: quoted-printable > >
