[tree-ssa] Fix bootstrap failure

[tree-ssa] Fix bootstrap failure

[Diego Novillo]

2003-11-30  Daniel Berlin  <dberlin@dberlin.org>

        * c-config-lang.in: Move tree-alias-* from here
        * Makefile.in (GTFILES): To here.
	[ ... ]

causes this bootstrap failure:

  /home/cygnus/dnovillo/perf/sbox/tree-ssa-branch/local.i686/bld.tobiano/gcc/gengtype-lex.c -o gengtype-lex.o
gcc -c   -g  -DIN_GCC   -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long  -Wno-error  -DHAVE_CONFIG_H -DGENERATOR_FILE    -I. -I. -I/home/cygnus/dnovillo/perf/sbox/tree-ssa-branch/local.i686/src/gcc -I/home/cygnus/dnovillo/perf/sbox/tree-ssa-branch/local.i686/src/gcc/. -I/home/cygnus/dnovillo/perf/sbox/tree-ssa-branch/local.i686/src/gcc/../include -I/home/cygnus/dnovillo/perf/sbox/tree-ssa-branch/local.i686/src/gcc/../libbanshee/libcompat -I/home/cygnus/dnovillo/perf/sbox/tree-ssa-branch/local.i686/src/gcc/../libbanshee -I/home/cygnus/dnovillo/perf/sbox/tree-ssa-branch/local.i686/src/gcc/../libbanshee/points-to  \
  /home/cygnus/dnovillo/perf/sbox/tree-ssa-branch/local.i686/bld.tobiano/gcc/gengtype-yacc.c -o gengtype-yacc.o
gcc   -g  -DIN_GCC   -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long    -DHAVE_CONFIG_H -DGENERATOR_FILE  -o gengtype \
 gengtype.o gengtype-lex.o gengtype-yacc.o ../libiberty/libiberty.a
make[2]: *** No rule to make target `\/home/cygnus/dnovillo/perf/sbox/tree-ssa-branch/local.i686/src/gcc/tree-alias-common.c', needed by `s-gtype'.  Stop.
make[2]: Leaving directory `/notnfs/dnovillo/sbox/tree-ssa-branch/bld.tobiano/gcc'
make[1]: *** [stage1_build] Error 2
make[1]: Leaving directory `/notnfs/dnovillo/sbox/tree-ssa-branch/bld.tobiano/gcc'
make: *** [bootstrap] Error 2


Fixed with


	* Makefile.in (GTFILES): Fix typo in tree-alias-common.c
	entry.

Index: Makefile.in
===================================================================
RCS file: /cvs/gcc/gcc/gcc/Makefile.in,v
retrieving revision 1.903.2.145
diff -d -c -p -r1.903.2.145 Makefile.in
*** Makefile.in	30 Nov 2003 01:00:49 -0000	1.903.2.145
--- Makefile.in	30 Nov 2003 14:45:53 -0000
*************** GTFILES = $(srcdir)/input.h $(srcdir)/co
*** 2206,2212 ****
    $(srcdir)/tree-ssa.c $(srcdir)/tree-dfa.c $(srcdir)/tree-ssa-ccp.c \
    $(srcdir)/tree-ssa-pre.c $(srcdir)/tree-iterator.c $(out_file) \
    $(srcdir)/tree-alias-type.h $(srcdir)/tree-alias-common.h \
!   $(srcdir)/tree-alias-type.c \$(srcdir)/tree-alias-common.c \
    @all_gtfiles@
  
  GTFILES_FILES_LANGS = @all_gtfiles_files_langs@
--- 2206,2212 ----
    $(srcdir)/tree-ssa.c $(srcdir)/tree-dfa.c $(srcdir)/tree-ssa-ccp.c \
    $(srcdir)/tree-ssa-pre.c $(srcdir)/tree-iterator.c $(out_file) \
    $(srcdir)/tree-alias-type.h $(srcdir)/tree-alias-common.h \
!   $(srcdir)/tree-alias-type.c $(srcdir)/tree-alias-common.c \
    @all_gtfiles@
  
  GTFILES_FILES_LANGS = @all_gtfiles_files_langs@


Diego.

Re: [tree-ssa] Fix bootstrap failure

[Daniel Berlin]

>
> Fixed with
>
>
> 	* Makefile.in (GTFILES): Fix typo in tree-alias-common.c
> 	entry.
>

Update again, it was already fixed :P

Re: [tree-ssa] Fix bootstrap failure

[Toon Moene]

Daniel Berlin wrote:
>>
>> Fixed with
>>
>>     * Makefile.in (GTFILES): Fix typo in tree-alias-common.c
>>     entry.

> Update again, it was already fixed :P

Talking about "fixed" - as far as I can see your PTA only takes effect 
if HAVE_BANSHEE is defined during compilation.  Now, despite of the fact 
that libbanshee is built correctly (and used !) in my bootstrap 
(powerpc-unknown-linux-gnu), I still don't see a -DHAVE_BANSHEE on the 
compile lines ...

Is that OK ?  Is there some other way to have your code take effect ?

-- 
Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://gcc.gnu.org/fortran/ (under construction)

Re: [tree-ssa] Fix bootstrap failure

[Daniel Berlin]

On Nov 30, 2003, at 1:33 PM, Toon Moene wrote:

> Daniel Berlin wrote:
>>>
>>> Fixed with
>>>
>>>     * Makefile.in (GTFILES): Fix typo in tree-alias-common.c
>>>     entry.
>
>> Update again, it was already fixed :P
>
> Talking about "fixed" - as far as I can see your PTA only takes effect 
> if HAVE_BANSHEE is defined during compilation.  Now, despite of the 
> fact that libbanshee is built correctly (and used !) in my bootstrap 
> (powerpc-unknown-linux-gnu), I still don't see a -DHAVE_BANSHEE on the 
> compile lines ..

Because it gets thrown into gcc's config.h.

Toplevel configure automatically passes --with-libbanshee to gcc 
configure if libbanshee exists.

(of course, there is no --without-libbanshee flag right now for 
toplevel configure, you have to remove the libbanshee dir for it to not 
pass --with-libbanshee, but at least *that* works)

Re: [tree-ssa] Fix bootstrap failure

[Toon Moene]

Daniel Berlin wrote:

> On Nov 30, 2003, at 1:33 PM, Toon Moene wrote:

>> Talking about "fixed" - as far as I can see your PTA only takes effect 
>> if HAVE_BANSHEE is defined during compilation.

> Because it gets thrown into gcc's config.h.

Yep, got it - was looking in the wrong spot.  Sorry.

-- 
Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://gcc.gnu.org/fortran/ (under construction)

Re: [tree-ssa] Fix bootstrap failure

[Toon Moene]

Toon Moene wrote:

> Daniel Berlin wrote:
> 
>> On Nov 30, 2003, at 1:33 PM, Toon Moene wrote:

>>> Talking about "fixed" - as far as I can see your PTA only takes 
>>> effect if HAVE_BANSHEE is defined during compilation.

>> Because it gets thrown into gcc's config.h.

> Yep, got it - was looking in the wrong spot.  Sorry.

BTW, the real reason I'm interested in alias analysis is to implement 
Fortran's "if you don't tell us you're not aliasing" rules.

I'm looking at tree-alias-common.c (intra_function_call):

   /* We assume that an actual parameter can point to any global. */

where Fortran would change the "can" to "cannot" :-)

Is that the point to change ?

Thanks in advance,



-- 
Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://gcc.gnu.org/fortran/ (under construction)

Re: [tree-ssa] Fix bootstrap failure

[Daniel Berlin]

On Nov 30, 2003, at 2:13 PM, Toon Moene wrote:

> Toon Moene wrote:
>
>> Daniel Berlin wrote:
>>> On Nov 30, 2003, at 1:33 PM, Toon Moene wrote:
>
>>>> Talking about "fixed" - as far as I can see your PTA only takes 
>>>> effect if HAVE_BANSHEE is defined during compilation.
>
>>> Because it gets thrown into gcc's config.h.
>
>> Yep, got it - was looking in the wrong spot.  Sorry.
>
> BTW, the real reason I'm interested in alias analysis is to implement 
> Fortran's "if you don't tell us you're not aliasing" rules.
>
> I'm looking at tree-alias-common.c (intra_function_call):
>
>   /* We assume that an actual parameter can point to any global. */
>
> where Fortran would change the "can" to "cannot" :-)

Can parameters not point to each other, too?

If so, you can just make it stop calling intra_function_call for 
fortran. :P

Re: [tree-ssa] Fix bootstrap failure

[Toon Moene]

Daniel Berlin wrote:

> On Nov 30, 2003, at 2:13 PM, Toon Moene wrote:

>> BTW, the real reason I'm interested in alias analysis is to implement 
>> Fortran's "if you don't tell us you're not aliasing" rules.
>>
>> I'm looking at tree-alias-common.c (intra_function_call):
>>
>>   /* We assume that an actual parameter can point to any global. */
>>
>> where Fortran would change the "can" to "cannot" :-)

> Can parameters not point to each other, too?

Indeed.

> If so, you can just make it stop calling intra_function_call for 
> fortran. :P

Well, that would be great !

-- 
Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://gcc.gnu.org/fortran/ (under construction)