[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: LSB 3.1 status for etch

Jeff Licquia wrote:
On Tue, 2006-08-01 at 11:53 +0200, Andreas Barth wrote:
* Jeff Licquia (licquia@debian.org) [060731 22:18]:
These failures are common to all distributions using X.org 7.  Several
symbols have moved from one library to another.
Would that failure also appear with LSB 1.3?

We would most likely see the same behavior on 1.3 (though I haven't
tested for it).

I have been told, however, that this is a problem with our tests again,
and not with the actual code.  Apparently, X.org's libX11 pulls in
libXau, meaning that the problem with resolving symbols should not
affect real applications.  Our test does not read this information,
however, which is why the tests fail.

Yes, the current thinking is that while the specification requires these
interfaces "in libX11" the real requirement is that an application using
them, and linked against libX11, work correctly (interesting concept,
that :-).  This would be the case if they're in another library but
libX11 has a DT_NEEDED on that library.  This seems to be the case in
all the instances of this problem we have seen, where the interfaces moved to libXau but libX11 depends on libXau. The checker tool does
not currently follow dependencies this way, so the concern about just
granting a blanket waiver for this is that a system which failed to
provide the DT_NEEDED on libXau would not conform but until the tool is fixed we can't detect the difference between that and a system which does.

As usual in the necessarily somewhat pedantic world of standards, this
is only an opinion until a waiver is actually granted or some other
solution is in place (like a fix to the tool, my preferred approach)

Reply to: