Regression in libbind's resolver initialization

Andris Kalnozols andris at hpl.hp.com
Fri Jul 6 02:54:01 UTC 2007


> > Andris Kalnozols napsal(a):
> > > BTW, this problem only shows up on the ia64; things work fine on PA-RISC.
> > >
> > > ------
> > > Andris Kalnozols
> > >   
> > Could be compiler problem, couldn't it?
> > 
> > Adam
> 
> 	One would really need to look at *statp.  At the point
> 	in the code where this is occuring res_init is theoretically
> 	cleaning up a old *statp.
> 
> 	Mark

In the 2.5 months since I last applied HP-UX patches to this system,
the following compiler/linker patches (among 203 others) got applied
in the latest go-around:

  PHSS_36336   11.23 linker + fdp cumulative patch
  PHSS_36343   11.23 aC++ Runtime
  PHSS_36344   11.23 Math Library Cumulative Patch
  PHSS_36345   11.23 Integrity Unwind Library
  PHSS_36347   11.23 assembler patch

After recompiling libbind from v9.5.0a5 with Mark's res_init()
patches, Vixie's dnscap program is finally running on the Itanium.
What a colossal timesink this kind of issue can be :-(

------
Andris


More information about the bind-workers mailing list