| Message ID | 20100822161448.GA16116@bromo.med.uc.edu |
|---|---|
| State | New |
| Headers | show |
On Aug 22, 2010, at 9:14 AM, Jack Howarth wrote: > Index: gcc/config.gcc > =================================================================== > --- gcc/config.gcc (revision 163449) > +++ gcc/config.gcc (working copy) > @@ -237,6 +237,8 @@ > | pdp11-*-bsd \ > | sparc-hal-solaris2* \ > | thumb-*-* \ > + | *-*-darwin[0-5] \ > + | *-*-darwin[0-5].* \ Either . is a regex, in which case this matches darwin10, or it isn't and it doesn't match darwin5 and darwin5.0. > | *-*-linux*aout* \ > | *-*-linux*coff* \ > | *-*-linux*libc1* \ > @@ -1950,14 +1952,14 @@ > extra_options="${extra_options} rs6000/darwin.opt" > extra_parts="crt2.o" > case ${target} in > - *-darwin1[0-9]* | *-darwin[8-9]*) > + *-darwin[8-9]*) ? This seems to change darwin10?
On Sun, Aug 22, 2010 at 08:41:21PM -0700, Mike Stump wrote: > On Aug 22, 2010, at 9:14 AM, Jack Howarth wrote: > > Index: gcc/config.gcc > > =================================================================== > > --- gcc/config.gcc (revision 163449) > > +++ gcc/config.gcc (working copy) > > @@ -237,6 +237,8 @@ > > | pdp11-*-bsd \ > > | sparc-hal-solaris2* \ > > | thumb-*-* \ > > + | *-*-darwin[0-5] \ > > + | *-*-darwin[0-5].* \ > > Either . is a regex, in which case this matches darwin10, or it isn't and it doesn't match darwin5 and darwin5.0. Doesn't this cover darwin0, darwin1, darwin2, darwin3, darwin4 and darwin5 as well the same with .*? > > > | *-*-linux*aout* \ > > | *-*-linux*coff* \ > > | *-*-linux*libc1* \ > > @@ -1950,14 +1952,14 @@ > > extra_options="${extra_options} rs6000/darwin.opt" > > extra_parts="crt2.o" > > case ${target} in > > - *-darwin1[0-9]* | *-darwin[8-9]*) > > + *-darwin[8-9]*) > > ? This seems to change darwin10? No, this is in the powerpc-*-darwin* subsection where someone got widely optimistic and assumed the that there would be darwin releases past darwin9 for powerpc darwin. Jack
On Sun, Aug 22, 2010 at 08:41:21PM -0700, Mike Stump wrote: > On Aug 22, 2010, at 9:14 AM, Jack Howarth wrote: > > Index: gcc/config.gcc > > =================================================================== > > --- gcc/config.gcc (revision 163449) > > +++ gcc/config.gcc (working copy) > > @@ -237,6 +237,8 @@ > > | pdp11-*-bsd \ > > | sparc-hal-solaris2* \ > > | thumb-*-* \ > > + | *-*-darwin[0-5] \ > > + | *-*-darwin[0-5].* \ > > Either . is a regex, in which case this matches darwin10, or it isn't and it doesn't match darwin5 and darwin5.0. Mike, I still don't understand your concerns. If you look a bit further down in the list of unsupported targets in gcc/config.gcc, you will find... | *-*-solaris2.[0-7] \ | *-*-solaris2.[0-7].* \ which is what I used as the template for... | *-*-darwin[0-5] \ | *-*-darwin[0-5].* \ In the first case, solaris2.0 through solaris2.7 as well as solaris2.0.* through solaris2.7.* will be unsupported. Likewise, the section I added covers darwin0 through darwin5 as well as darwin0.* through darwin5.*. What is the confusion here? Certainly the current entries for solaris2 aren't broken. > > > | *-*-linux*aout* \ > > | *-*-linux*coff* \ > > | *-*-linux*libc1* \ > > @@ -1950,14 +1952,14 @@ > > extra_options="${extra_options} rs6000/darwin.opt" > > extra_parts="crt2.o" > > case ${target} in > > - *-darwin1[0-9]* | *-darwin[8-9]*) > > + *-darwin[8-9]*) > > ? This seems to change darwin10? As I mentioned before, there is and will never be a darwin release past darwin9.* for powerpc-apple-darwin*. So this removes the bogus target entry *-darwin1[0-9]* entry from within the powerpc section. In context... powerpc-*-darwin*) extra_options="${extra_options} rs6000/darwin.opt" extra_parts="crt2.o" case ${target} in *-darwin[8-9]*) <=== removed bogus entry here tmake_file="${tmake_file} rs6000/t-darwin8" tm_file="${tm_file} rs6000/darwin8.h" ;; *-darwin7*) tm_file="${tm_file} rs6000/darwin7.h" ;; *-darwin6*) ;; esac tmake_file="${tmake_file} t-slibgcc-darwin" lto_binary_reader=lto-macho extra_headers=altivec.h ;; Is http://gcc.gnu.org/ml/gcc-patches/2010-08/msg01703.html, okay with those clarifications? Jack
Index: gcc/config.gcc =================================================================== --- gcc/config.gcc (revision 163449) +++ gcc/config.gcc (working copy) @@ -237,6 +237,8 @@ | pdp11-*-bsd \ | sparc-hal-solaris2* \ | thumb-*-* \ + | *-*-darwin[0-5] \ + | *-*-darwin[0-5].* \ | *-*-linux*aout* \ | *-*-linux*coff* \ | *-*-linux*libc1* \ @@ -1950,14 +1952,14 @@ extra_options="${extra_options} rs6000/darwin.opt" extra_parts="crt2.o" case ${target} in - *-darwin1[0-9]* | *-darwin[8-9]*) + *-darwin[8-9]*) tmake_file="${tmake_file} rs6000/t-darwin8" tm_file="${tm_file} rs6000/darwin8.h" ;; *-darwin7*) tm_file="${tm_file} rs6000/darwin7.h" ;; - *-darwin[0-6]*) + *-darwin6*) ;; esac tmake_file="${tmake_file} t-slibgcc-darwin"
This revised patch limits the depreciated darwin releases to those before darwin6 as well as removes the non-existent target entries for darwin10+ on powerpc-*-darwin*. Bootstrap and regression tested on x86_64-apple-darwin10. Okay for gcc trunk? Jack 2010-08-22 Jack Howarth <howarth@bromo.med.uc.edu> * gcc/config.gcc: Add darwin[0-5] and darwin[0-5].* to unsupported target list. Remove non-existent *-darwin1[0-9]* and *-darwin[0-5] from powerpc-*-darwin* case.