Bootstrap problem...

Bootstrap problem...

[Alan Grimes]

Hey; I'm trying to bootstrap GCC 3.0 under BeOS. 

It goes fairly well, (only a few hundred warning messages) right through
stage 3 build. But then it fails during the comparison between 2 and 3. 

The config line was: 

   --host=i686-pc-beos

The build compiler was: 

   GCC ver 2.9-beos-991026

The ./gcc/xgcc  (stage 3 compiler) is a valid executable which responds
to inquiries about its capabilities normally. (I don't know how to test
it on actual code..)

It is equal in size to the ./gcc/stage2/xgcc compiler except nearly
every file in the build faild comparison... 

As for performance: the stage 2 compiler was 7% smaller than the stage 1
compiler. and the stage 2 compiler. The stage 1 compiled the code in
about 17 minutes while the stage 2 compiler worked in only 13 minutes. 

The underlying hardware is an Athalon 800 with 256mb of ram.

-- 
Libertarians: Left wing conservatives. =)   www.lp.org
http://users.erols.com/alangrimes/  <my website.
Any usage of this e-mail account is subject to the terms and conditions
specified on my website.

Re: Bootstrap problem

[Salvatore Filippone]

Il giorno gio, 16/09/2010 alle 22.19 -0700, Ian Lance Taylor ha scritto:
> 
> That is code which is only ever executed when compiling for the ia64
> target.  I would guess that something changed and nobody tested it for
> ia64.  If you're not up for debugging it yourself, then please file a
> bug report according to the instructions at
> http://gcc.gnu.org/bugs.html
> No need to include a test case, as it is part of gcc itself.
> 
> Ian 

Whatever it was, it disappeared at r164336, where bootstrap completed;
it could be a fix was applied, or it could be that a change was split in
multiple commits and I got an intermediate step that was inconsistent. 

I have seen a great many instances of 

/tmp/ccvLT0FG.s:59: Warning: rest of line ignored; first ignored character is `i'

and similarly with 'd'; I have no idea if this is significant or not;
will now proceed to compile and test a couple of codes and see what
happens.

Does it make sense to open a PR now? I guess not... 
I'll rebuild every now and then, and will let you know if it surfaces
again. 

Regards
Salvatore 

Re: Bootstrap problem

[Ian Lance Taylor]

Salvatore Filippone <salvatore.filippone@uniroma2.it> writes:

> ../../../../gcc/libstdc++-v3/libsupc++/array_type_info.cc:33:1: internal
> compiler error: tree check: expected tree that contains 'decl with RTL'
> structure, have 'addr_expr' in output_constant, at varasm.c:4408
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See <http://gcc.gnu.org/bugs.html> for instructions.
> make[4]: *** [array_type_info.lo] Error 1

That is code which is only ever executed when compiling for the ia64
target.  I would guess that something changed and nobody tested it for
ia64.  If you're not up for debugging it yourself, then please file a
bug report according to the instructions at http://gcc.gnu.org/bugs.html
No need to include a test case, as it is part of gcc itself.

Ian

Bootstrap problem

[Salvatore Filippone]

Hi, 
on an Itanium system I seeing this error in stage 3, trunk at  r164313:

libtool: compile:  /home3/pasqua/GNUBUILD/obj/./gcc/xgcc -shared-libgcc
-B/home3/pasqua/GNUBUILD/obj/./gcc -nostdinc++
-L/home3/pasqua/GNUBUILD/obj/ia64-unknown-linux-gnu/libstdc++-v3/src
-L/home3/pasqua/GNUBUILD/obj/ia64-unknown-linux-gnu/libstdc
++-v3/src/.libs -B/home3/pasqua/gnu46/ia64-unknown-linux-gnu/bin/
-B/home3/pasqua/gnu46/ia64-unknown-linux-gnu/lib/
-isystem /home3/pasqua/gnu46/ia64-unknown-linux-gnu/include
-isystem /home3/pasqua/gnu46/ia64-unknown-linux-gnu/sys-include
-I/home3/pasqua/GNUBUILD/gcc/libstdc++-v3/../gcc
-I/home3/pasqua/GNUBUILD/obj/ia64-unknown-linux-gnu/libstdc
++-v3/include/ia64-unknown-linux-gnu
-I/home3/pasqua/GNUBUILD/obj/ia64-unknown-linux-gnu/libstdc++-v3/include
-I/home3/pasqua/GNUBUILD/gcc/libstdc++-v3/libsupc++
-fno-implicit-templates -Wall -Wextra -Wwrite-strings -Wcast-qual
-fdiagnostics-show-location=once -ffunction-sections -fdata-sections -g
-O2 -D_GNU_SOURCE -c ../../../../gcc/libstdc++-v3/libsupc
++/array_type_info.cc  -fPIC -DPIC -o array_type_info.o
../../../../gcc/libstdc++-v3/libsupc++/array_type_info.cc:33:1: internal
compiler error: tree check: expected tree that contains 'decl with RTL'
structure, have 'addr_expr' in output_constant, at varasm.c:4408
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.
make[4]: *** [array_type_info.lo] Error 1
make[4]: Leaving directory
`/home3/pasqua/GNUBUILD/obj/ia64-unknown-linux-gnu/libstdc++-v3/libsupc
++'
make[3]: *** [all-recursive] Error 1
make[3]: Leaving directory
`/home3/pasqua/GNUBUILD/obj/ia64-unknown-linux-gnu/libstdc++-v3'
make[2]: *** [all] Error 2
make[2]: Leaving directory
`/home3/pasqua/GNUBUILD/obj/ia64-unknown-linux-gnu/libstdc++-v3'
make[1]: *** [all-target-libstdc++-v3] Error 2
make[1]: Leaving directory `/home3/pasqua/GNUBUILD/obj'
make: *** [bootstrap] Error 2

I have searched on bugzilla, there are some similar items, but probably
not identical. 
Is this really new? Ideas anyone? 

Thanks 
Salvatore