[lsb-discuss] Interpreting lts_vsx-pcts test failures?

[Andrew Josey]

Hello John

I discussed with my engineers and our comments are inserted below


> Subject: [lsb-discuss] Interpreting lts_vsx-pcts test failures?
> Date: Thu, 07 Feb 2008 10:17:33 +0000
> 
> I have been using parts of the lts_vsx-pcts-3.1.1 test suite to test a
> Linux filesystem.  This has shown up a number of test failures.  For
> some it is fairly clear where the problem lies.  However, others are
> harder to interpret ...
> 
> I have unpacked the source and found the MAN directory but attempting
> to print them as follows:
> 
> $ groff -man -t T.chown > /tmp/xxx.ps
> T.chown:24: warning: `deTL' not defined (probable missing space after
> `de')
> T.chown:25: warning: numeric expression expected (got `\')
> T.chown:27: warning: `deTI' not defined (probable missing space after
> `de')
> T.chown:32: warning: numeric expression expected (got `\')
> T.chown:35: warning: `deUL' not defined (probable missing space after
> `de')
> T.chown:37: warning: numeric expression expected (got `\')
> T.chown:40: warning: `deUI' not defined (probable missing space after
> `de')
> $
> 
> produces rather mixed results.  Am I missing some macro definitions?
> Is there a better way to print these pages?

The macros in these files predate groff by many years, and they
assume the original nroff/troff semantics of maximum two character
macro names.

They can be printed with a "real" troff (e.g. on Solaris), or
something like this would probably work for most of the files:

sed 's/^\.de/.de /' T.chown | groff -man -t > /tmp/xxx.ps

> The manual pages refer to various assertions.  For example:
> 
>     Posix Ref: Component CHOWN Assertion 5.6.5.2-04(A)
> 
> Is there an online source detailing these assertions or would I need
> to obtain the relevent standard document?

The references are to P1003.3.1 draft 11.8, which was the current
draft at the time the work to cross-reference the VSX4 assertions
with P1003.3.1 was done.  I think in those days the drafts were
not even available to balloters electronically, it was all paper.

The standard was later published as P2003.1, but there are likely to
be some differences in the assertion numbers.  In any case I doubt if
P2003.1 is available any more, as it was not renewed or revised, so
is now defunct.

The only real purpose of having the "Posix Ref" data in the VSX4
manual pages was to enable a POSIX.3-conforming test report to be
produced (using the prpt utility) when the test suite was used in
POSIX mode (now called POSIX90 mode) or FIPS mode.  We may delete them
from the manual pages during the next development phase.

regards
Andrew