Re: Another attempt at fixing sane-gcc-compiler

2010-12-02 Thread David Holmes
Thanks Mike. I'd been struggling to work out why this was failing (as I didn't want to pull in your back-out changeset. I was left puzzling why this is even checking for something gcc related on Solaris when I'm not even building with gcc ??? In any case this looks ok now. David Mike Duigou

Re: Another attempt at fixing sane-gcc-compiler

2010-12-02 Thread Mike Duigou
On Dec 2 2010, at 14:41 , Kelly O'Hair wrote: > I think the rules are broken. > > The ifdef and ifndef and ifeq and ifneq statements are Make-isms and should > not use TAB characters. > But the shell logic of the rule needs to have the line start with a TAB, so > the "@if [..." line needs to s

Re: Another attempt at fixing sane-gcc-compiler

2010-12-02 Thread Kelly O'Hair
I think the rules are broken. The ifdef and ifndef and ifeq and ifneq statements are Make-isms and should not use TAB characters. But the shell logic of the rule needs to have the line start with a TAB, so the "@if [..." line needs to start with a TAB. It is hard to see in a webrev, but I t

Another attempt at fixing sane-gcc-compiler

2010-12-02 Thread Mike Duigou
The previous patch (CR 6998016) which attempted to fix the incorrect ifdef nesting in the Sanity.gmk sane-gcc-compiler rule ran into a problem on 32 bit versions of solaris due to a previously unreported error in the use of REQUIRED_GCC_VERSION (which should have been REQUIRED_GCC_VER). The fixe