This bug can be reproduce on Solaris 7/8 machines.

To reproduce bug:
1. cd /net/sqesvr/vsn/GammaBase/Bugs/<Bug ID>
2. edit script if needed
3. ksh doit.ksh

$ ksh doit.ksh
VM option '+CompileTheWorld'
VM option 'MinInliningThreshold=0'
VM option 'CompileTheWorldStartAt=3600'
VM option 'CompileTheWorldStopAt=3799'
CompileTheWorld : Compiling all classes in /net/sqesvr/vsn/testbase/tessrc/jit/compiletheworld/jars/a-d/compstudio.jar

CompileTheWorld (3600) : sunw/hotjava/misc/RequestProcessor$Request
CompileTheWorld (3601) : sunw/hotjava/misc/RequestProcessor$RequestNode
CompileTheWorld (3602) : sunw/hotjava/misc/RequestProcessor
CompileTheWorld (3603) : sunw/hotjava/misc/RfcDateParser
...
...
...
CompileTheWorld (3751) : textuality/lark/Namer
CompileTheWorld (3752) : textuality/lark/Segment
CompileTheWorld (3753) : textuality/lark/Text
CompileTheWorld (3754) : textuality/lark/XmlInputStream
#
# HotSpot Virtual Machine Error, assertion failure
# Please report this error at
# http://java.sun.com/cgi-bin/bugreport.cgi
#
# assert(my_base == my_inputs_base || phase->hash_find(my_base) == phase->hash_find(my_inputs_base), "base pointers must match")
#
# Error ID: /net/jano/export/disk05/hotspot/auto/ws_jano/fpt/src/share/vm/opto/addnode.cpp, 394 [ Patched ]
#
# Problematic Thread: prio=5 tid=0x15fa258 nid=0xb runnable
#
assertion failure
assert(my_base == my_inputs_base || phase->hash_find(my_base) == phase->hash_find(my_inputs_base), "base pointers must match")

Do you want to debug the problem?