[Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 fails on undefined symbol in libgcj.so

[Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 fails on undefined symbol in libgcj.so

[dhazeghi@yahoo.com]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10802



------- Additional Comments From dhazeghi@yahoo.com  2003-05-26 07:03 -------
That's bizarre. The symbol in question is in libgcc.so.1 on Linux. Can you check whether that 
symbol is in that library on Solaris? Can you add -v to the failing command line to see whether the 
right version of libgcc is being linked in? Thanks,

Dara



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 fails on undefined symbol in libgcj.so

[michaud.bernard@ic.gc.ca]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10802



------- Additional Comments From michaud.bernard@ic.gc.ca  2003-05-26 12:01 -------
Subject: RE: [Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 f
	ails on undefined symbol in libgcj.so

-----Original Message-----
From: dhazeghi@yahoo.com [mailto:gcc-bugzilla@gcc.gnu.org]
Sent: Monday, May 26, 2003 03:03
To: Michaud, Bernard: FIN
Subject: [Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86
fails on undefined symbol in libgcj.so


PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10802



------- Additional Comments From dhazeghi@yahoo.com  2003-05-26 07:03
-------
That's bizarre. The symbol in question is in libgcc.so.1 on Linux. Can you
check whether that 
symbol is in that library on Solaris? Can you add -v to the failing command
line to see whether the 
right version of libgcc is being linked in? Thanks,

Dara



------- You are receiving this mail because: -------
You reported the bug, or are watching the reporter.


Hi Dara,

Thank you for replying.  I am attaching to this message the output from the
failing command line with "-v".  As well, I've attached the output from the
entire build as a compressed file.  As for the libgcc library, would that be
the one from the compiling compiler (which is gcc-3.2.3, or the compiler
being compiled (gcc-3.3)?

Regards,
Bernard Michaud


------- Additional Comments From michaud.bernard@ic.gc.ca  2003-05-26 12:01 -------
Created an attachment (id=4077)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=4077&action=view)

------- Additional Comments From michaud.bernard@ic.gc.ca  2003-05-26 12:01 -------
Created an attachment (id=4078)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=4078&action=view)




------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 fails on undefined symbol in libgcj.so

[dhazeghi@yahoo.com]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10802



------- Additional Comments From dhazeghi@yahoo.com  2003-05-26 17:28 -------
Hello,

I don't see anything obvious out of order. I meant the libgcc and libgcc_s from the 3.3 you're 
building. Could you check whether the symbol libgcj isn't finding is indeed in libgcc*? I guess what 
I'm wondering is whether perhaps it's picking up the wrong version of the library without the 
necessary symbol. Could you check whether the 3.2 libgcc* also has the symbol? Thanks,

Dara



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 fails on undefined symbol in libgcj.so

[pinskia@physics.uc.edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10802


pinskia@physics.uc.edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING





------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 fails on undefined symbol in libgcj.so

[michaud.bernard@ic.gc.ca]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10802



------- Additional Comments From michaud.bernard@ic.gc.ca  2003-05-26 18:04 -------
Subject: RE: [Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 f
	ails on undefined symbol in libgcj.so

The libgcc_s.so.1 from gcc-3.2.3 does not contain the
_Unwind_FindEnclosingFunction symbol, as follows:

$ elfdump -s /opt/gcc-3.2.3/lib/libgcc_s.so.1|fgrep "Unwind"
       [6]  0x00004550 0x0000001b  FUNC GLOB  D    2 .text
_Unwind_DeleteException
      [14]  0x00002ec0 0x0000000b  FUNC GLOB  D    2 .text
_Unwind_GetTextRelBase
      [17]  0x000045b0 0x0000000e  FUNC GLOB  D    2 .text
_Unwind_SetIP
      [18]  0x00002eb0 0x0000000b  FUNC GLOB  D    2 .text
_Unwind_GetDataRelBase
      [26]  0x00005770 0x0000017c  FUNC GLOB  D    2 .text
_Unwind_Find_FDE
      [30]  0x00004570 0x00000010  FUNC GLOB  D    2 .text
_Unwind_GetGR
      [52]  0x00002e90 0x0000000b  FUNC GLOB  D    2 .text
_Unwind_GetLanguageSpecificData
      [54]  0x00004580 0x00000013  FUNC GLOB  D    2 .text
_Unwind_SetGR
      [58]  0x00004100 0x00000171  FUNC GLOB  D    2 .text
_Unwind_RaiseException
      [62]  0x000043a0 0x000000c4  FUNC GLOB  D    2 .text
_Unwind_ForcedUnwind
      [67]  0x00002ea0 0x0000000b  FUNC GLOB  D    2 .text
_Unwind_GetRegionStart
      [75]  0x00004470 0x000000d7  FUNC GLOB  D    2 .text
_Unwind_Resume
      [78]  0x000045a0 0x0000000b  FUNC GLOB  D    2 .text
_Unwind_GetIP
     [123]  0x00004030 0x000000cd  FUNC LOCL  D    0 .text
_Unwind_RaiseException_Phase2
     [124]  0x00004280 0x00000120  FUNC LOCL  D    0 .text
_Unwind_ForcedUnwind_Phase2
     [161]  0x00004550 0x0000001b  FUNC GLOB  D    0 .text
_Unwind_DeleteException
     [169]  0x00002ec0 0x0000000b  FUNC GLOB  D    0 .text
_Unwind_GetTextRelBase
     [172]  0x000045b0 0x0000000e  FUNC GLOB  D    0 .text
_Unwind_SetIP
     [173]  0x00002eb0 0x0000000b  FUNC GLOB  D    0 .text
_Unwind_GetDataRelBase
     [181]  0x00005770 0x0000017c  FUNC GLOB  D    0 .text
_Unwind_Find_FDE
     [185]  0x00004570 0x00000010  FUNC GLOB  D    0 .text
_Unwind_GetGR
     [207]  0x00002e90 0x0000000b  FUNC GLOB  D    0 .text
_Unwind_GetLanguageSpecificData
     [209]  0x00004580 0x00000013  FUNC GLOB  D    0 .text
_Unwind_SetGR
     [213]  0x00004100 0x00000171  FUNC GLOB  D    0 .text
_Unwind_RaiseException
     [217]  0x000043a0 0x000000c4  FUNC GLOB  D    0 .text
_Unwind_ForcedUnwind
     [222]  0x00002ea0 0x0000000b  FUNC GLOB  D    0 .text
_Unwind_GetRegionStart
     [230]  0x00004470 0x000000d7  FUNC GLOB  D    0 .text
_Unwind_Resume
     [233]  0x000045a0 0x0000000b  FUNC GLOB  D    0 .text
_Unwind_GetIP

However, the libgcc_s.so.1 from the gcc-3.3 build seems to contain the
right symbol:

$ elfdump -s /playground/build/gccbuild/gcc/libgcc_s.so.1|fgrep "Unwind"
       [3]  0x00004710 0x0000000b  FUNC GLOB  D    2 .text
_Unwind_GetIP
      [13]  0x00004360 0x000000bd  FUNC GLOB  D    2 .text
_Unwind_ForcedUnwind
      [14]  0x00004720 0x0000000e  FUNC GLOB  D    2 .text
_Unwind_SetIP
      [17]  0x00002db0 0x0000000b  FUNC GLOB  D    2 .text
_Unwind_GetRegionStart
      [19]  0x00004620 0x000000b9  FUNC GLOB  D    3 .text
_Unwind_Backtrace
      [21]  0x000045f0 0x0000002f  FUNC GLOB  D    2 .text
_Unwind_DeleteException
      [22]  0x00004420 0x000000dc  FUNC GLOB  D    2 .text
_Unwind_Resume
      [27]  0x00002e10 0x0000000b  FUNC GLOB  D    2 .text
_Unwind_GetTextRelBase
      [29]  0x000058f0 0x0000017c  FUNC GLOB  D    2 .text
_Unwind_Find_FDE
      [31]  0x00002e00 0x0000000b  FUNC GLOB  D    2 .text
_Unwind_GetDataRelBase
      [32]  0x00002da0 0x0000000b  FUNC GLOB  D    2 .text
_Unwind_GetLanguageSpecificData
      [33]  0x000046e0 0x00000010  FUNC GLOB  D    2 .text
_Unwind_GetGR
      [44]  0x00004500 0x000000e8  FUNC GLOB  D    3 .text
_Unwind_Resume_or_Rethrow
      [47]  0x000040c0 0x00000179  FUNC GLOB  D    2 .text
_Unwind_RaiseException
      [52]  0x000046f0 0x00000013  FUNC GLOB  D    2 .text
_Unwind_SetGR
      [68]  0x00002d90 0x0000000b  FUNC GLOB  D    3 .text
_Unwind_GetCFA
      [81]  0x00002dc0 0x0000003c  FUNC GLOB  D    3 .text
_Unwind_FindEnclosingFunction
     [123]  0x00003ff0 0x000000cd  FUNC LOCL  D    0 .text
_Unwind_RaiseException_Phase2
     [124]  0x00004240 0x00000120  FUNC LOCL  D    0 .text
_Unwind_ForcedUnwind_Phase2
     [158]  0x00004710 0x0000000b  FUNC GLOB  D    0 .text
_Unwind_GetIP
     [168]  0x00004360 0x000000bd  FUNC GLOB  D    0 .text
_Unwind_ForcedUnwind
     [169]  0x00004720 0x0000000e  FUNC GLOB  D    0 .text
_Unwind_SetIP
     [172]  0x00002db0 0x0000000b  FUNC GLOB  D    0 .text
_Unwind_GetRegionStart
     [174]  0x00004620 0x000000b9  FUNC GLOB  D    0 .text
_Unwind_Backtrace
     [176]  0x000045f0 0x0000002f  FUNC GLOB  D    0 .text
_Unwind_DeleteException
     [177]  0x00004420 0x000000dc  FUNC GLOB  D    0 .text
_Unwind_Resume
     [182]  0x00002e10 0x0000000b  FUNC GLOB  D    0 .text
_Unwind_GetTextRelBase
     [184]  0x000058f0 0x0000017c  FUNC GLOB  D    0 .text
_Unwind_Find_FDE
     [186]  0x00002e00 0x0000000b  FUNC GLOB  D    0 .text
_Unwind_GetDataRelBase
     [187]  0x00002da0 0x0000000b  FUNC GLOB  D    0 .text
_Unwind_GetLanguageSpecificData
     [188]  0x000046e0 0x00000010  FUNC GLOB  D    0 .text
_Unwind_GetGR
     [199]  0x00004500 0x000000e8  FUNC GLOB  D    0 .text
_Unwind_Resume_or_Rethrow
     [202]  0x000040c0 0x00000179  FUNC GLOB  D    0 .text
_Unwind_RaiseException
     [207]  0x000046f0 0x00000013  FUNC GLOB  D    0 .text
_Unwind_SetGR
     [223]  0x00002d90 0x0000000b  FUNC GLOB  D    0 .text
_Unwind_GetCFA
     [236]  0x00002dc0 0x0000003c  FUNC GLOB  D    0 .text
_Unwind_FindEnclosingFunction

It would thus seem to be a build issue in that if the proper library had
been referred to, the symbol should have been located. However, the
command that abnormally terminates does seem to contain all the proper
library and search patch directives:

/playground/build/gccbuild/gcc/gcj -v
-B/playground/build/gccbuild/i686-sun-solaris2.9/libjava/
 -B/playground/build/gccbuild/gcc/ -ffloat-store -g -O2 -o .libs/jv-convert
--main=gnu.gcj.convert.Convert
 -shared-libgcc  -L/playground/build/gccbuild/i686-sun-solaris2.9/libjava
-L/playground/build/gccbuild/i686-sun-solaris2.9/libjava/.libs
 ./.libs/libgcj.so
-L/playground/build/gccbuild/i686-sun-solaris2.9/libstdc++-v3/src
-L/playground/build/gccbuild/i686-sun-solaris2.9/libstdc++-v3/src/.libs
 -L/usr/local/lib -lpthread -lrt -ldl -lz -L/playground/build/gccbuild/gcc
-L/usr/ccs/bin -L/usr/ccs/lib -lgcc_s -lgcc_s -Wl,-R -Wl,/opt/gcc-3.3/lib/.
-Wl,-R -Wl,/usr/local/lib

Any ideas? Anything else we can try?

Regards,
Bernard Michaud




------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 fails on undefined symbol in libgcj.so

[dhazeghi@yahoo.com]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10802



------- Additional Comments From dhazeghi@yahoo.com  2003-05-27 06:43 -------
Hello,

is it possible for you to move your existing gcc installation (or at least libgcc*) to another location, 
where the current build won't pick it up, and check whether the build finishes? Not that this is, I 
admit a good fix. But at least we'll know where the issue is.

Dara



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 fails on undefined symbol in libgcj.so

[michaud.bernard@ic.gc.ca]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10802



------- Additional Comments From michaud.bernard@ic.gc.ca  2003-05-27 11:22 -------
Subject: RE:  Bootstrap of gcc-3.3 on Solaris 9 x86 f
	ails on undefined symbol in libgcj.so

Thank you Dara... This worked!  The build was somehow picking up the
gcc-3.2.3 libgcc* libraries, which lack the required
_Unwind_FindEnclosingFunction symbol.  It now builds properly.  I concur
that it is not a perfect fix, but at least it does work.

Thank you again for the help... it is much appreciated.

Bernard Michaud




------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 fails on undefined symbol in libgcj.so

[pinskia@physics.uc.edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10802


pinskia@physics.uc.edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2003-05-27 15:15:13
               date|                            |


------- Additional Comments From pinskia@physics.uc.edu  2003-05-27 15:15 -------
See submitter's answer (comment 8) to Dara's question (comment 7): the problem is that 
some how the build of gcc is picking up libgcc from a pervious installed version.



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 fails on undefined symbol in libgcj.so

[michaud.bernard@ic.gc.ca]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10802



------- Additional Comments From michaud.bernard@ic.gc.ca  2003-06-03 19:42 -------
Subject: RE:  Bootstrap of gcc-3.3 on Solaris 9 x86 f
	ails on undefined symbol in libgcj.so

Sorry for the delay in answering Dara, things here have been quite hectic...

The gcc 3.2.3 compiler (the version doing the compiling) is installed in
/opt/gcc-3.2.3, and I was trying to build gcc 3.3 to /opt/gcc-3.3.

However, I may have found the cause of the original problem. I found
symbolic links pointing from /usr/local/lib to an older version of gcc-3.2,
as such:

libgcc_s.so -> libgcc_s.so.1
libgcc_s.so.1 -> /opt/gcc-3.2.1/lib/libgcc_s.so.1

This is possibly what is responsible for the build picking up the wrong
version of libgcc, as /usr/local/lib was in my search path.

Regards,
Bernard Michaud




------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 fails on undefined symbol in libgcj.so

[dhazeghi@yahoo.com]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10802


dhazeghi@yahoo.com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|major                       |minor
  GCC build triplet|                            |i686-pc-solaris2.9
   GCC host triplet|                            |i686-pc-solaris2.9
 GCC target triplet|                            |i686-pc-solaris2.9
           Priority|P3                          |P4


------- Additional Comments From dhazeghi@yahoo.com  2003-06-08 03:51 -------
Thanks for the response. I thinks it's hectic for everyone right now :-) I think this a gcc bug, but 
it'd be good to see if someone else can confirm that, and I suspect it will take a while to solve...

[Bug bootstrap/10802] Bootstrap of gcc-3.3 on Solaris 9 x86 fails on undefined symbol in libgcj.so

[neroden at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10802



------- Additional Comments From neroden at gcc dot gnu dot org  2003-07-12 03:39 -------
>From the response, this is almost certainly an environment bug.  But the new version of GCC should not go looking around for old copies of 
libgcc.so.1, but should instead be bulletproofed to only find the new copy.  So this is a bug, but it might be hard to reproduce.