-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I hate mozilla. I really hate mozilla. I violently dislike having to repeat I 
hate mozilla.

Oh well, I have a strange one here I'm trying to figure out with esr31-mozilla.

Environment is pkgsrc 32-bit on omnios bloody with gcc4.9.2 (and patches)...

With a recent workaround patch, I'm finally able to build xulrunner31 that has 
been stifled for a while due to 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63779

Great! Let's move on to firefox31 and thunderbird31.

The subject line is what I'm getting with the link of libxul.so in firefox31.
Strange, libxul.so just linked okay a minute ago when building xulrunner31!
(unfortunately xulrunner and firefox don't share libxul.so [yet])

These should be more or less identical since the patches and primary options 
are the same
(true, the configure options are a bit different) so the resulting libxul.so 
should for all intents and purposes be equivalent.

With xulrunner31's version I notice the following elfdump extracts:
> Section Header[179]:  sh_name: .data.rel.ro._ZTVN2JS16CustomAutoRooterE
>     sh_addr:      0               sh_flags:   [ SHF_WRITE SHF_ALLOC SHF_GROUP 
> ]
>     sh_size:      0xc             sh_type:    [ SHT_PROGBITS ]
>     sh_offset:    0x16490         sh_entsize: 0
>     sh_link:      0               sh_info:    0
>     sh_addralign: 0x8       
> 
> Section Header[180]:  sh_name: .rel.data.rel.ro._ZTVN2JS16CustomAutoRooterE
>     sh_addr:      0               sh_flags:   0
>     sh_size:      0x8             sh_type:    [ SHT_REL ]
>     sh_offset:    0x27a30         sh_entsize: 0x8 (1 entry)
>     sh_link:      187             sh_info:    179
>     sh_addralign: 0x4       
> 
...
>      [668]  0x00000000 0x0000000c  OBJT WEAK  D    0 .data.rel.ro._ 
> _ZTVN2JS16CustomAutoRooterE
>  
...
> Group Section:  .group
>      index    flags / section         signature symbol
>        [0]   [ COMDAT ]               _ZTVN2JS16CustomAutoRooterE
>        [1]   .data.rel.ro._ZTVN2JS16CustomAutoRooterE [179]
> 
...
> Relocation Section:  .rel.data.rel.ro._ZTVN2JS16CustomAutoRooterE
>     type                       offset             section        symbol
>   R_386_32                        0x8             .rel.data.rel. 
> __cxa_pure_virtual
> 

but in firefox31:
> Section Header[176]:  sh_name: .data.rel.ro._ZTVN2JS16CustomAutoRooterE
>     sh_addr:      0               sh_flags:   [ SHF_WRITE SHF_ALLOC SHF_GROUP 
> ]
>     sh_size:      0xc             sh_type:    [ SHT_PROGBITS ]
>     sh_offset:    0x196b8         sh_entsize: 0
>     sh_link:      0               sh_info:    0
>     sh_addralign: 0x8       
> 
> Section Header[177]:  sh_name: .rel.data.rel.ro._ZTVN2JS16CustomAutoRooterE
>     sh_addr:      0               sh_flags:   0
>     sh_size:      0x8             sh_type:    [ SHT_REL ]
>     sh_offset:    0x2ad70         sh_entsize: 0x8 (1 entry)
>     sh_link:      184             sh_info:    176
>     sh_addralign: 0x4       
> 
...
>      [667]  0x00000000 0x0000000c  OBJT WEAK  H    0 .data.rel.ro._ 
> _ZTVN2JS16CustomAutoRooterE
...
> Group Section:  .group
>      index    flags / section         signature symbol
>        [0]   [ COMDAT ]               _ZTVN2JS16CustomAutoRooterE
>        [1]   .data.rel.ro._ZTVN2JS16CustomAutoRooterE [176]
> 
...
> Relocation Section:  .rel.data.rel.ro._ZTVN2JS16CustomAutoRooterE
>     type                       offset             section        symbol
>   R_386_32                        0x8             .rel.data.rel. 
> __cxa_pure_virtual
> 



The only thing that seems different are the two symtab entries:
>      index    value      size      type bind oth ver shndx          name
..
>      [668]  0x00000000 0x0000000c  OBJT WEAK  D    0 .data.rel.ro._ 
> _ZTVN2JS16CustomAutoRooterE
>      [667]  0x00000000 0x0000000c  OBJT WEAK  H    0 .data.rel.ro._ 
> _ZTVN2JS16CustomAutoRooterE

Is the 'H' indicating hidden?

Since this is a class defined in jsapi.o (with more or less the same 
characteristics),
could that be having an negative effect on jsobj.o in this particular case?

I put the output from building with 'LD_OPTIONS=-Dsymbols,reloc,detail,output' 
http://www.netbsd.org/~richard/xulrunner31-g++-issue/ff31-ld-debug.out.gz
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQEcBAEBAgAGBQJUgKXJAAoJECAB22fHtp27DOUH/jxvZFiu2rOeCCHb+rNfTNZP
u+7FDIkZCXcEkWlw1T31zs85A8GjDFLxNwpXsM05S0lBPtFNCZeLm2y21NxtCGfc
SRE/xDclXtvUlfoqwLJiaETLl1RuXOl2hzknk6KGsm6Fzqu7ywFrMOC+CCUCuD1A
1D9d2bG7UnkvB3arbR5eTYZSi2NIwkv+De5TgOxqO6jJ5V5IMwUZxMG2F1JGKRrm
TUY7paR98Rlmg9mUbV9SIG+6ECw+jGmlvyNz7+kaBOaquZQXFKQYPrepa6yI5yVm
E1ekrIXpXMrWBnt1DkDOZNjxb9LG9D7h/BdKU4SAVcsmFKblTZl8OmiOlT/rRfk=
=GdUT
-----END PGP SIGNATURE-----



-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175430-2e6923be
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21175430&id_secret=21175430-6a77cda4
Powered by Listbox: http://www.listbox.com

Reply via email to