Re: [Evolution-hackers] symbol ugliness ...

[michael meeks <michael meeks novell com>]

On Tue, 2006-01-10 at 12:39 +0200, Tor Lillqvist wrote:
> Hmm, either you have a time machine Michael, or there has been some
> slight delay in the mailing list ;-)

	;-)

> Odd that your script didn't notice some duplicated functions from code
> generated the ORBit2 IDL compiler. There should have been such in, lemme
> recall, the libevolution-mail and libeshell DSOs, for instance. Or maybe
> those cases had been introduced only in HEAD? (Anyway, they should be
> fixed now.) Or isn't this finterpose script about just finding duplicate
> symbols, but something more complex?

	Ah ;-) well - you have to ignore data relocations - because of copy
relocs into binaries. But yes - I should do some more work - I was just
comparing the intersection of all DSOs on the system's symbols. You find
some amazing things:

$ objdump -T /usr/lib/jvm/jre-1.4.2/lib/i386/libjcov.so | grep list
00010eb0 g    DF .text  00000040  Base        add_to_list
00010ef0 g    DF .text  0000007e  Base        add_to_list_end
00010f70 g    DF .text  00000032  Base        list_size
00010e18 g    DF .text  00000098  Base        free_list
0001425c g    DO .data  00000004  Base        thread_list

	eg. ;-)

	Anyway - I need to expand it to binaries, no doubt that'll reveal yet
more interesting bugs :-)

	HTH,

		Michael.

-- 
 michael meeks novell com  <><, Pseudo Engineer, itinerant idiot