Howdy, I am still waiting to get results from gcc60, but after reading whiteknight++'s guess about the nature of this bug, these warnings seem very suspect:
src/gc/gc_ms2.c: In function 'gc_ms2_free_pmc_header': src/gc/gc_ms2.c:724:49: warning: cast increases required alignment of target type [-Wcast-align] src/gc/gc_ms2.c:729:58: warning: cast increases required alignment of target type [-Wcast-align] src/gc/gc_ms2.c: In function 'gc_ms2_mark_pmc_header': src/gc/gc_ms2.c:752:38: warning: cast increases required alignment of target type [-Wcast-align] src/gc/gc_ms2.c: In function 'gc_ms2_free_string_header': src/gc/gc_ms2.c:834:49: warning: cast increases required alignment of target type [-Wcast-align] src/gc/gc_ms2.c:842:61: warning: cast increases required alignment of target type [-Wcast-align] src/gc/gc_ms2.c: In function 'gc_ms2_sweep_pmc_pool': src/gc/gc_ms2.c:1144:5: warning: cast increases required alignment of target type [-Wcast-align] src/gc/gc_ms2.c: In function 'gc_ms2_destroy_pmc_pool': src/gc/gc_ms2.c:1190:5: warning: cast increases required alignment of target type [-Wcast-align] src/gc/gc_ms2.c: In function 'gc_ms2_sweep_string_pool': src/gc/gc_ms2.c:1220:5: warning: cast increases required alignment of target type [-Wcast-align] src/gc/gc_ms2.c: In function 'gc_ms2_is_ptr_owned': src/gc/gc_ms2.c:1262:30: warning: cast increases required alignment of target type [-Wcast-align] Duke On Wed, Jun 13, 2012 at 10:53 AM, Jonathan "Duke" Leto <[email protected]> wrote: > Howdy, > > Just for reference, gcc60 and gcc64 in the GCC Compile Farm [0] are both ia64. > > I am attempting to build parrot on gcc60 now and will report back with > results. > > Duke > > [0] http://gcc.gnu.org/wiki/CompileFarm > > On Mon, Jun 11, 2012 at 5:25 AM, Andrew Whitworth <[email protected]> > wrote: >> On the surface this seems nonsensical. The String PMC does not >> implement the get_pointer VTABLE (as the error message points out) but >> no place inside our build process should be attempting to use it that >> way. >> >> This suggests to me something GC-related. Somewhere a PMC is probably >> not being marked and then the memory is being recycled to allocate a >> String. When we try to call get_pointer on the old reference we get to >> the String PMC and have problems. >> >> Considering that the GC is one of the few places that relies on >> hardware-specific code, and that code is not always well tested, this >> doesn't surprise me. >> >> Do we have any developers knowledgable about IA64 architecture or >> access to a machine for testing? >> >> Thanks, >> >> --Andrew Whitworth >> >> >> On Sun, Jun 10, 2012 at 7:00 AM, Alessandro Ghedini <[email protected]> >> wrote: >>> Hi all, >>> >>> Parrot 4.3.0 failed (twice) to build on Debian's ia64 buildd [0] with the >>> error: >>> >>>> ./parrot-nqp --target=pir --output=compilers/opsc/gen/Ops/Emitter.pir >>>> compilers/opsc/src/Ops/Emitter.pm >>>> get_pointer() not implemented in class 'String' >>>> current instr.: 'parrot;PAST;Compiler;post_children' pc 2489 >>>> (compilers/pct/src/PAST/Compiler.pir:508) >>>> called from Sub 'parrot;PAST;Compiler;as_post' pc 3172 >>>> (compilers/pct/src/PAST/Compiler.pir:773) >>>> called from Sub 'parrot;PAST;Compiler;as_post' pc 3275 >>>> (compilers/pct/src/PAST/Compiler.pir:821) >>>> called from Sub 'parrot;PAST;Compiler;post_children' pc 2595 >>>> (compilers/pct/src/PAST/Compiler.pir:553) >>>> called from Sub 'parrot;PAST;Compiler;as_post' pc 3172 >>>> (compilers/pct/src/PAST/Compiler.pir:773) >>>> called from Sub 'parrot;PAST;Compiler;post_children' pc 2595 >>>> (compilers/pct/src/PAST/Compiler.pir:553) >>>> called from Sub 'parrot;PAST;Compiler;pirop' pc 5280 >>>> (compilers/pct/src/PAST/Compiler.pir:1388) >>>> called from Sub 'parrot;PAST;Compiler;post_children' pc 2595 >>>> (compilers/pct/src/PAST/Compiler.pir:553) >>>> called from Sub 'parrot;PAST;Compiler;as_post' pc 4565 >>>> (compilers/pct/src/PAST/Compiler.pir:1176) >>>> called from Sub 'parrot;PAST;Compiler;post_children' pc 2595 >>>> (compilers/pct/src/PAST/Compiler.pir:553) >>>> called from Sub 'parrot;PAST;Compiler;as_post' pc 3172 >>>> (compilers/pct/src/PAST/Compiler.pir:773) >>>> called from Sub 'parrot;PAST;Compiler;as_post' pc 3275 >>>> (compilers/pct/src/PAST/Compiler.pir:821) >>>> called from Sub 'parrot;PAST;Compiler;post_children' pc 2595 >>>> (compilers/pct/src/PAST/Compiler.pir:553) >>>> called from Sub 'parrot;PAST;Compiler;as_post' pc 3172 >>>> (compilers/pct/src/PAST/Compiler.pir:773) >>>> called from Sub 'parrot;PAST;Compiler;post_children' pc 2595 >>>> (compilers/pct/src/PAST/Compiler.pir:553) >>>> called from Sub 'parrot;PAST;Compiler;pirop' pc 5280 >>>> (compilers/pct/src/PAST/Compiler.pir:1388) >>>> called from Sub 'parrot;PAST;Compiler;post_children' pc 2595 >>>> (compilers/pct/src/PAST/Compiler.pir:553) >>>> called from Sub 'parrot;PAST;Compiler;as_post' pc 4565 >>>> (compilers/pct/src/PAST/Compiler.pir:1176) >>>> called from Sub 'parrot;PAST;Compiler;to_post' pc 1931 >>>> (compilers/pct/src/PAST/Compiler.pir:215) >>>> called from Sub 'parrot;PCT;HLLCompiler;compile' pc 464 >>>> (compilers/pct/src/PCT/HLLCompiler.pir:331) >>>> called from Sub 'parrot;HLL;Compiler;eval' pc 24537 >>>> (ext/nqp-rx/src/stage0/HLL-s0.pir:8848) >>>> called from Sub 'parrot;PCT;HLLCompiler;evalfiles' pc 1501 >>>> (compilers/pct/src/PCT/HLLCompiler.pir:764) >>>> called from Sub 'parrot;PCT;HLLCompiler;command_line' pc 1726 >>>> (compilers/pct/src/PCT/HLLCompiler.pir:879) >>>> called from Sub 'parrot;NQP;Compiler;main' pc 92462 >>>> (ext/nqp-rx/src/stage0/NQP-s0.pir:28374) >>>> make[1]: *** [compilers/opsc/gen/Ops/Emitter.pir] Error 1 >>> >>> The problem here is that I cannot reproduce the failure on Debian's ia64 >>> porterbox. Is anyone able to reliably reproduce it or is aware of any change >>> between 4.0.0 and 4.3.0 that may be causing this? >>> >>> Cheers >>> >>> [0] >>> https://buildd.debian.org/status/fetch.php?pkg=parrot&arch=ia64&ver=4.3.0-1&stamp=1339321350 >>> >>> -- >>> perl -E '$_=q;$/= @{[@_]};and s;\S+;<inidehG ordnasselA>;eg;say~~reverse' >>> >>> _______________________________________________ >>> http://lists.parrot.org/mailman/listinfo/parrot-dev >>> >> >> _______________________________________________ >> Pkg-parrot-devel mailing list >> [email protected] >> http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-parrot-devel > > > > -- > Jonathan "Duke" Leto <[email protected]> > Leto Labs LLC http://labs.leto.net > 209.691.DUKE http://dukeleto.pl -- Jonathan "Duke" Leto <[email protected]> Leto Labs LLC http://labs.leto.net 209.691.DUKE http://dukeleto.pl _______________________________________________ http://lists.parrot.org/mailman/listinfo/parrot-dev
