Nick thanks for looking into this. Nick Kew wrote: > > On 10 Mar 2009, at 03:08, Sriram Natarajan wrote: > >> - A friendly reminder... can you kindly take a look at the attached >> diff and let me know your thoughts. >>> >>> http://cr.opensolaris.org/~sn123202/php529.mar9/webrev/ > > What's happening in Makefile.sfw? In particular, do the lines > > +PRECONF_ENVLINE= \ > + PATH=/usr/gnu/bin:/usr/bin \ > + MAKE=$(GMAKE) > + > > together with > > + (cd $(PHP_DIR); env - $(PRECONF_ENVLINE) \ > + ./buildconf --force ) > > switch an entire build to using gcc? > no, the dtrace patch that we are including here introduces the 'enable-dtrace' configuratin flag. so, we are using 'buildconf' to regenerate the configure script . since this requires gnu tools and gmake to do so, i am setting up the path properly before generating the configure.
why we are doing this way - we want to keep the dtrace patch to the minimum so that is is easier for us to carry and maintain in a longer term - for the life of php 5.2.. we are still compiling with sun studio 12 > And what's php_streams.c.patch about? Looks like broken C syntax? > yes, i already took care of this. looks like , i missed syncing the latest version to the webrev url. thanks for noticing it. > But I don't think I follow the patch format used here: patch to a patch > in php_dtrace.patch and php_build_config.m4.patch? > not sure, what you mean by patch to a patch.. php_dtrace.patch -> carries necessary source code changes so that dtrace probe can be enabled within php engine php_build_config_m4.patch -> this is the necessary change to config.m4 files within php build environment that is necessary to build in opensolaris environment. earlier, we were not (re) generating the configure script. hence, we were patching the existing configure script to carry these changes. now, with addition of dtrace probes - we are regenerating the configure script - hence, it makes more sense to actually patch these config.m4 files which are easier to maintan - sriram
