[Lldb-commits] [PATCH] D16186: Unconditionally accept symbol sizes from .dynsym

[Tamas Berghammer via lldb-commits]

tberghammer created this revision.
tberghammer added a reviewer: clayborg.
tberghammer added a subscriber: lldb-commits.

Unconditionally accept symbol sizes from .dynsym

If an elf object file have no .symtab section then we create our symtab
representation based on the .dynsym section. The symtab creation have a
heuristic what updates the size of symbols with 0 size to last until the
beginning of the next symbol what is valid when we have full symbol table
but it isn't valid whit .dynsym because of the missing private symbols.
As a result if we have symbols with 0 size in the .dynsym we will expand
them until the next symbol and these artificially created symbol sizes will
confuse the stack unwinder.

This change disables the artificial symbol size calculation for the scenario
when we are using .dynsym.

http://reviews.llvm.org/D16186

Files:
  include/lldb/Symbol/Symbol.h
  source/Plugins/ObjectFile/ELF/ObjectFileELF.cpp
  source/Plugins/ObjectFile/ELF/ObjectFileELF.h
  source/Symbol/Symbol.cpp
  source/Symbol/Symtab.cpp

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D16186.44872.patch
Type: text/x-patch
Size: 8232 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/lldb-commits/attachments/20160114/1247bb34/attachment-0001.bin>