Danek Duvall wrote:
On Sun, Jun 28, 2009 at 01:10:56PM +0100, petede wrote:
http://cr.opensolaris.org/~petede/gate/
t_elf.py:
- line 137: this should be get_info() instead of get_dynamic()?
Absolutely! Argh. Fixed now.
The main code change seems fine to me, though I'm confused as to how the
bug is getting triggered. I see that liblist_foreach() is operating on
"deps" in getdynamic() but I don't see how deps could possibly be NULL when
it's passed into liblist_foreach(). It could only be NULL if
liblist_alloc() failed, but at that point we should have goneto fail.
The issue is that some binaries have a .dynamic section but no
.SUNW_version section. This means that during the getdynamic()
function the .dynamic section is found and the associated "deps"
list is populated. However as no .SUNW_version is found then no
corresponding "vers" list is populated.
This then means in the pythonify_2dliblist_cb() it attempts
to 'process' the empty vers list.
The binaries affected are those ones that have the ld called
on them to build the dynamic sections. For example:
/usr/ccs/bin/ld -r -dy -Nmisc/mac -Nmisc/net80211 -Ndrv/random -Ndrv/ip
-o debug32/wpi debug32/wpi.o
Does that help (I'll update the bug with this if it is clearer) ?
Ta
pete
What am I missing?
Danek
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
